tag:blogger.com,1999:blog-78907649721664111052024-03-18T10:08:06.721+01:00Nick Brown's blogThe adventures of a self-appointed data police cadetNick Brownhttp://www.blogger.com/profile/00172030184497186082noreply@blogger.comBlogger96125tag:blogger.com,1999:blog-7890764972166411105.post-12649050904990553932023-11-10T17:22:00.009+01:002023-11-11T20:21:10.809+01:00Attack of the 50 foot research assistants: Lee et al. (2019), Study 3<div style="line-height: 24px;">
<p>This post is about some issues in Study 3 of the following article:</p>
<div style="margin-left: 36pt; text-indent: -48px;"><span style="text-indent: 0px;"> </span>Lee, J. J., Hardin, A. E., Parmar, B., & Gino, F. (2019). The
interpersonal costs of dishonesty: How dishonest behavior reduces
individuals' ability to read others' emotions. <i>Journal of Experimental
Psychology: General</i>, <i>148</i>(9), 1557–1574. https://doi.org/10.1037/xge0000639</div><div style="margin-left: 36pt; text-indent: -36pt;"><br /></div><div style="margin-left: 36pt; text-indent: -36pt;">On 2023-11-06 I was able to download the article from <a href="https://francescagino.com/s/Lee-et-al-2019-The-Interpersonal-Costs-of-Dishonesty.pdf" target="_blank">here</a>.<br /></div><div style="margin-left: 36pt; text-indent: -36pt;"><br /></div><div style="margin-left: 36pt; text-indent: -36pt;"></div>
<div>This article is <a href="https://manycoauthors.org/gino/29" target="_blank">paper #29</a> in the <a href="https://manycoauthors.org/about/" target="_blank">Many Co-Authors</a> project, where researchers who have co-authored papers with Professor Francesca Gino are reporting the provenance of the data in those papers, following the discovery of <a href="https://datacolada.org/109/" target="_blank">problems</a> with the data in four articles co-authored by Professor Gino.</div><div> </div><div>In the tables on <a href="https://web.archive.org/web/20231106155610/https://manycoauthors.org/gino/29" target="_blank">the Many Co-Authors page for paper #29</a>, two of the three co-authors of this article have so far (2023-11-10) provided
information about the provenance of the data for this article, with both indicating that Professor Gino was involved in the data collection for Study 3. This note from Julia Lee Cunningham, the lead author, provides further confirmation:</div><div><br /></div><div style="margin-left: 40px; text-align: left;">For Study 3, Gino’s research assistant ran the laboratory study at Harvard Business School Research Lab for the partial data on Gino’s Qualtrics account. The co-authors have access to the raw data and were able to reproduce the key published results for Study 3. <br /></div><div></div><div><br /></div><div>In this study, pairs of participants interacted by telling each other stories. In one condition ("dishonest"), one member of the pair (A) told a fake story and the other (B) told a true story. In the other condition ("honest"), both members of the pair told true stories. Then, B evaluated their emotions during the exercise, and A evaluated their perceptions of B's emotions. The dependent variable ("emotional accuracy") was the ability of A to accurately evaluate how B had been feeling during the exercise. The results showed that when A had been dishonest (by telling a fake story), they were less accurate in their evaluation of B's emotional state.<br /></div><div> </div><div>The dataset for Study 3 is available as part of the OSF repository for the whole article <a href="https://osf.io/bquy5/" target="_blank">here</a>. It consists of an SPSS data file (.SAV) and a "syntax" (code) file (.SPS). I do not currently have an SPSS licence, so I was unable to run the code, but it seems to be fairly straightforward, running the focal <i>t</i> test from the study followed by the ANCOVAs to test whether gender moderated the relationship between condition and emotional accuracy.</div><div> </div><div>I converted the dataset file to .CSV format in R and was then able to replicate the focal result of the study ("participants in the <i>dishonest</i> condition (<i>M</i> = 1.58, <i>SD</i> = 0.63) were significantly less<br />accurate at detecting others’ mental and affective states than those in the <i>honest</i> condition (<i>M</i>=1.39, <i>SD </i>= 0.54), <i>t</i>(209) = 2.37, <i>p</i> = .019", p. 1564, emphasis in original). My R code gave me this result:<br /></div><div><br /></div><div><pre aria-label="Console Output" class="GONVA-ECC2B" id="rstudio_console_output" role="document" style="-webkit-text-stroke-width: 0px; background-color: white; border: medium; color: black; font-family: "Lucida Console", monospace; font-size: 13.3333px; font-style: normal; font-variant-caps: normal; font-variant-ligatures: normal; font-weight: 400; letter-spacing: normal; line-height: 1.2; margin: 0px; orphans: 2; outline: none; text-align: -webkit-left; text-decoration-color: initial; text-decoration-style: initial; text-indent: 0px; text-transform: none; user-select: text; white-space: pre-wrap; widows: 2; word-break: break-all; word-spacing: 0px;" tabindex="0"><span class="GONVA-ECD2B" role="document" style="outline: none;" tabindex="-1"><span class="GONVA-ECJ2B ace_keyword" style="color: blue; user-select: text; white-space: pre;">> t.test(df.repl.dis$EmoAcc, df.repl.hon$EmoAcc, var.equal=TRUE)<br /><br /> Two Sample t-test<br /><br />data: df.repl.dis$EmoAcc and df.repl.hon$EmoAcc<br />t = 2.369, df = 209, p-value = 0.01875</span></span></pre></div></div><div style="line-height: 24px;"></div><div style="line-height: 24px;"><br /><div>However, this is not the whole story. Although the dataset contains records from 250 pairs of participants, the article states (p. 1564):</div><div> </div><div style="margin-left: 40px; text-align: left;">As determined by research assistants monitoring each session, pairs were excluded for the following reasons: the wrong partner told their story first; they asked so many questions during the session that it became apparent they were not actually reading their survey instructions or questions (e.g., “What story am I supposed to be telling?”); or they were actively on their phone during the storytelling portion of the session. Exclusions were due to the actions of either individual in the pair; thus, of the 500 individuals, 39 did not follow instructions. This resulted in 106 pairs in the dishonest condition and 105 pairs in the honest condition.</div><div style="text-align: left;"><br /></div><div>The final total of 211 pairs is confirmed by the 209 degrees of freedom of the above <i>t</i> test.</div><div><br /></div><div>Conveniently, the results for the 39 excluded pairs are available in the dataset. They are excluded from analysis based on a variable named "Exclude_LabNotes" (although sadly, despite this name, the OSF data repository does not contain any lab notes that might explain the basis on which each exclusion was made). It is thus possible to run the analyses on the full dataset of 250 pairs, with no exclusions. When I did that, I obtained this result:</div><div><br /></div><div><pre aria-label="Console Output" class="GONVA-ECC2B" id="rstudio_console_output" role="document" style="-webkit-text-stroke-width: 0px; background-color: white; border: medium; color: black; font-family: "Lucida Console", monospace; font-size: 13.3333px; font-style: normal; font-variant-caps: normal; font-variant-ligatures: normal; font-weight: 400; letter-spacing: normal; line-height: 1.2; margin: 0px; orphans: 2; outline: none; text-align: -webkit-left; text-decoration-color: initial; text-decoration-style: initial; text-indent: 0px; text-transform: none; user-select: text; white-space: pre-wrap; widows: 2; word-break: break-all; word-spacing: 0px;" tabindex="0"><span class="GONVA-ECD2B" role="document" style="outline: none;" tabindex="-1"><span class="GONVA-ECJ2B ace_keyword" style="color: blue; user-select: text; white-space: pre;">> t.test(df.full.dis$EmoAcc, df.full.hon$EmoAcc, var.equal=TRUE)<br /><br /> Two Sample t-test<br /><br />data: df.full.dis$EmoAcc and df.full.hon$EmoAcc<br />t = 0.20148, df = 246, p-value = 0.8405</span></span></pre></div><div><br /></div>(Alert readers may notice that the degrees of freedom for this independent <i>t</i> test are only 246 rather than the expected 248. On inspection of the dataset, it appears that one record has NA for experimental condition and another has NA for emotional accuracy. Both of these records were also manually excluded by the research assistants, but they could not have been used in any of the <i>t</i> tests anyway. Hence, it seems fairer to say that 37 out of 248 participant pairs, rather than 39 out of 250, were excluded based on notes made by the RAs.)<br /><div> </div><div><div>As you can see, there is quite a difference from the previous <i>t</i> test (<i>p</i> = 0.8405 versus <i>p</i> = 0.01875). Had these 37 participant pairs not been excluded, there would be no difference between the conditions; put another way, the exclusions drive the entire effect. I ran the same <i>t</i> test on (only) these excluded participants:</div><div><br /></div><div><pre aria-label="Console Output" class="GONVA-ECC2B" id="rstudio_console_output" role="document" style="-webkit-text-stroke-width: 0px; background-color: white; border: medium; color: black; font-family: "Lucida Console", monospace; font-size: 13.3333px; font-style: normal; font-variant-caps: normal; font-variant-ligatures: normal; font-weight: 400; letter-spacing: normal; line-height: 1.2; margin: 0px; orphans: 2; outline: none; text-align: -webkit-left; text-decoration-color: initial; text-decoration-style: initial; text-indent: 0px; text-transform: none; user-select: text; white-space: pre-wrap; widows: 2; word-break: break-all; word-spacing: 0px;" tabindex="0"><span class="GONVA-ECD2B" role="document" style="outline: none;" tabindex="-1"><span class="GONVA-ECJ2B ace_keyword" style="color: blue; user-select: text; white-space: pre;">> t.test(df.dis$EmoAcc, df.hon$EmoAcc, var.equal=TRUE)<br /><br /> Two Sample t-test<br /><br />data: df.exconly.dis$EmoAcc and df.</span></span><span class="GONVA-ECD2B" role="document" style="outline: none;" tabindex="-1"><span class="GONVA-ECJ2B ace_keyword" style="color: blue; user-select: text; white-space: pre;">exconly.</span></span><span class="GONVA-ECD2B" role="document" style="outline: none;" tabindex="-1"><span class="GONVA-ECJ2B ace_keyword" style="color: blue; user-select: text; white-space: pre;">hon$EmoAcc<br />t = -4.1645, df = 35, p-value = 0.0001935<br /></span></span></pre></div><div><br /></div><div>Cohen's <i>d</i> for this test is 1.412, which is a very large effect indeed among people who are not paying attention.<br /></div><div> </div><div>I think it is worth illustrating these results graphically. First, a summary of the three <i>t</i> tests: <br /></div><div><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEg22VEGIeisrMmYHpmUc28r2lJNW991Q3tWmQ3ibHysP9Bti7HW9lAsDUPk_47w6W313PJLPDjEtGfWp_5x57aHUujsgHt4DjivthI5vrsQTHz865NvckfxiHadoktPigUB0Ct0Vp9nbDqFhy7kHa-rKAQx2WwqxLd-Hd8cdeY_hnyV3srNhuHOxAG6DR6F/s810/fig1.svg.png" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="540" data-original-width="810" height="426" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEg22VEGIeisrMmYHpmUc28r2lJNW991Q3tWmQ3ibHysP9Bti7HW9lAsDUPk_47w6W313PJLPDjEtGfWp_5x57aHUujsgHt4DjivthI5vrsQTHz865NvckfxiHadoktPigUB0Ct0Vp9nbDqFhy7kHa-rKAQx2WwqxLd-Hd8cdeY_hnyV3srNhuHOxAG6DR6F/w640-h426/fig1.svg.png" width="640" /></a></div></div><div>Second, an illustration of where each observation was dropped from its respective per-condition sample:</div><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhumNBjgF7jfpGGrcxNB6LBhbT1msPNbhcjzAFRlTQIwaFIkfL1CkZlPaMP_aLs_YemweFDP3sVpUg-ZB1hblZPryL-f7eLQ39MXlgg01KOuK6CWjlOpE6N8fe7AAFk0WnwclAJvMJXAENOQbOAOCaQMsKludNc6LNqUpAqGwNleDb99lmiyQXdOIKI26Do/s810/fig2.svg.png" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="630" data-original-width="810" height="498" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhumNBjgF7jfpGGrcxNB6LBhbT1msPNbhcjzAFRlTQIwaFIkfL1CkZlPaMP_aLs_YemweFDP3sVpUg-ZB1hblZPryL-f7eLQ39MXlgg01KOuK6CWjlOpE6N8fe7AAFk0WnwclAJvMJXAENOQbOAOCaQMsKludNc6LNqUpAqGwNleDb99lmiyQXdOIKI26Do/w640-h498/fig2.svg.png" width="640" /></a></div><div></div><div>[[Edit 2023-11-11 19:05 UTC: I updated the second figure above. The previous version reported "<i>t</i>(34) = 4.56", reflecting a <i>t</i> test with equal variances not assumed in which the calculated degrees of freedom were 34.466. This is actually the more correct way to calculate the <i>t</i> statistic, but I have been using "equal variances assumed" in all of the other analyses in this post for compatibility with the original article, which used analyses from SPSS in which the assumption of equal variances is the default. See also <a href="https://rips-irsp.com/articles/10.5334/irsp.82" target="_blank">this article</a>. ]] <br /></div><div> </div><div>This is quite remarkable. One might imagine that participants who were not paying attention to the instructions, or goofing off on their phones, would, overall, give responses that would show no effect, because their individual responses would have been noisy and/or because the set of excluded participants was approximately balanced across conditions (and there is no difference between the conditions for the full sample). Indeed, a legitimate reason to exclude these participants would be that their results are likely to be uninformative noise and so, if they were numerous enough, their inclusion might lead to a Type II error. But instead, it seems that these excluded participants showed a very strong effect <i>in the opposite direction to the hypothesis</i> (as shown by the negative <i>t</i> statistic). That is, if these results are to be believed, something about the fact that either A or B was not following the study instructions made A much better (or less bad) at determining B's emotions when telling a fake (versus true) story. There were 14 excluded participant pairs in the "dishonest" condition, with a mean emotional accuracy score (lower = more accurate) for A of 1.143, and 23 in the "honest" condition, with a mean emotional accuracy score of 2.076; for comparison, the mean score for the full sample across both conditions is 1.523.</div><div><br /></div><div>I hope the reader will forgive me for saying that this explanation does not seem very likely — and if it were true, it would presumably be the basis of intense interest among psychologists. Rather, there seem to be two other plausible explanations (but feel free to point out any others that you can think of in the comments). One is that the extreme results of the excluded participants arose by chance — and, hence, the apparent effect in favour of the authors' hypothesis caused by their exclusion was also the result of chance. The other, painful though it is to contemplate, is that the research assistants may have excluded participants in order to give a result in line with the hypotheses.</div><div> </div><div>I simulated how likely it would be for the removal of 37 random participant pairs from the sample of 248 complete records to give a statistically significant result. I ran 1,000,000 simulations and obtained only 12 <i>p</i> values less than 0.05 for the <i>t</i> test on the resulting sample of 211 pairs. The smallest <i>p</i> value that I obtained was 0.03387, which is higher than the one reported in the article. To put it another way, out of a million attempts I was unable to obtain even one result as extreme as the published one by chance.</div><br /></div>Now, this process can surely be subjected to some degree of criticism from a statistical
inference point of view. After all, if the report in the article is
correct, the excluded participants were not selected truly at random,
and they might differ in other ways from the rest of the sample, with those differences perhaps interacting with the experimental condition. There might be other, more formally correct ways to test the idea that the exclusions of participant pairs were independent of their scores. However,
as mentioned earlier, I do not think that it can be seriously argued that
there was some extremely powerful psychological process, contrary to the study hypothesis, taking place specifically with the excluded participants.<div><br /></div><div>So it seems to me that, by elimination, the most plausible remaining explanation is that the research assistants selected which participants to exclude based on their scores, in such a way as to produce results that favoured the authors' hypothesis. Exactly how they were able to do that, given that those scores were only available in Qualtrics, when their job was (presumably) to help participants sitting in the laboratory to understand the process and to check who was spending time on their phone, is unclear to me, but doubtless there is a coherent explanation. Indeed, Professor Gino has already suggested (see point 274 <a href="https://storage.courtlistener.com/recap/gov.uscourts.mad.259933/gov.uscourts.mad.259933.1.0.pdf" target="_blank">here</a>) that research assistants may have been responsible for perceived anomalies in other studies on which she was an author, although so far no details on how exactly this might have happened have been made public. I hope that she will be able to track down the RAs in this case and establish the truth of the matter with them.<br /></div><div><br /></div><div><h4 style="text-align: left;">Supporting files</h4>I have made my analysis code available <a href="http://nickbrown.fr/blog/lee-study3" target="_blank">here</a>. I encourage you to use the authors' original SPSS data file from the OSF link given above, and convert it to CSV format using the commented-out line at the top of the script. However, as a convenience, I have made a copy of that CSV file available along with my code.</div><div> </div><div></div></div><div style="line-height: 24px;"><h4 style="text-align: left;">Acknowledgements</h4><h4 style="text-align: left;"><span style="font-weight: normal;">I thank Daniël Lakens and two anonymous readers of earlier drafts of this post for their comments. One of those people also kindly provided the two charts in the post.</span></h4><h4 style="text-align: left;"><span style="font-weight: normal;"> </span></h4><h4 style="text-align: left;"><span style="font-weight: normal;">I wrote to Julia Lee Cunningham to give her a heads up about this post. With her permission, I will quote any reply that she might make here.<br /></span></h4><div></div><h4 style="text-align: left;"><br /></h4></div>Nick Brownhttp://www.blogger.com/profile/18266307287741345798noreply@blogger.com12tag:blogger.com,1999:blog-7890764972166411105.post-48898846342546582802023-07-09T15:49:00.367+02:002023-11-08T11:31:25.946+01:00Data errors in Mo et al.'s (2023) analysis of wind speed and voting patterns<div style="line-height: 24px;">
<p>This post is about some issues in the following article, and most notably its dataset:</p>
<div style="margin-left: 36pt; text-indent: -48px;"><span style="text-indent: 0px;">Mo, C. H., Jachimowicz, J. M., Menges, J. I., & Galinsky, A. D. (2023).</span> <span style="text-indent: -36pt;">The impact of incidental environmental factors on vote c</span><span style="text-indent: -36pt;">hoice: Wind speed is related to more prevention‑focused voting. <i>Political Behavior</i>. Advance online publication. </span>https://doi.org/10.1007/s11109-023-09865-y</div><div style="margin-left: 36pt; text-indent: -36pt;"><br /></div>
You can download the article from <a href="https://link.springer.com/article/10.1007/s11109-023-09865-y" target="_blank">here</a>, the Supplementary Information from <a href="https://static-content.springer.com/esm/art%3A10.1007%2Fs11109-023-09865-y/MediaObjects/11109_2023_9865_MOESM1_ESM.docx" target="_blank">here</a> [.docx], and the dataset from <a href="https://osf.io/9y4sn/" target="_blank">here</a>. Credit is due to the authors for making their data available so that others can check their work.</div><div style="line-height: 24px;"><br /></div><div style="line-height: 24px;"><h4 style="line-height: 24px; text-align: left;">Introduction</h4><div style="line-height: 24px;">The premise of this article, which was brought to my attention in a direct message by a Twitter user, is that the wind speed observed on the day of an "election" (although in fact, all the cases studied by the authors were referendums) affects the behaviour of voters, but only if the question on the ballot represents a choice between prevention- and promotion-focused options, in the sense of <a href="https://psycnet.apa.org/record/1997-43865-002" target="_blank">regulatory focus theory</a>. The authors stated in their abstract that "we find that individuals exposed to higher wind speeds become more prevention-focused and more likely to support prevention-focused electoral options".</div><div style="line-height: 24px;"><br /></div><div style="line-height: 24px;">This article (specifically the part that focused on the UK's referendum on leaving the European Union ("Brexit") has already been critiqued by Erik Gahner <a href="https://erikgahner.dk/2023/the-many-causes-of-brexit-2/" target="_blank">here</a>.</div></div><div style="line-height: 24px;"><br /></div><div style="line-height: 24px;">I should state from the outset that I was skeptical about this article when I read the abstract, and things did not get better when I found a couple of basic factual errors in the descriptions of the Brexit referendum:</div><div style="line-height: 24px;"><ol style="text-align: left;"><li>On p. 9 the authors claim that "The referendum for UK to leave the European Union (EU) was advanced by the Conservative Party, one of the three largest parties in the UK", and again, on p. 12, they state "In the case of the Brexit vote, the Conservative Party advanced the campaign for the UK to leave the EU". However, this is completely incorrect. The Conservative Party was split over how to vote, but the majority of its members of parliament, including David Cameron, the party leader and Prime Minister, campaigned for a Remain vote (<a href="In the latter case, we considered the percentage of council seats held by the Scottish National Party (SNP), as the SNP advanced the campaign for Scotland independence." target="_blank">source</a>).</li><li>At several points, the authors claim that the question posed in the Brexit referendum required a "Yes"/"No" answer. On p. 7 we read "For Brexit, the “No” option advanced by the Stronger In campaign was seen as clearly prevention-oriented ... whereas the “Yes” option put forward by the Vote Leave campaign was viewed as promotion-focused". The reports of result coding on p. 8, and the note to Table 1 on p. 10, repeat this claim. But this is again entirely incorrect. The options given to voters were to "Remain" (in the EU) or "Leave" (the EU). As the authors themselves note, the official campaign against EU membership was named "Vote Leave" (and there was also an unofficial campaign named "Leave.EU"). Indeed, this choice was adopted, rather than "Yes" or "No" responses to the question "Should the United Kingdom remain a member of the European Union?", precisely to avoid any perception of "positivity bias" in favour of a "Yes" vote (<a href="https://www.theguardian.com/politics/2015/sep/01/eu-referendum-cameron-urged-to-change-wording-of-preferred-question" target="_blank">source</a>). Note also here that, had this change not been made, the pro-EU vote would have been "Yes", and not the (prevention-focused) "No" claimed by the authors. (*)</li></ol></div><div style="line-height: 24px; text-align: left;">Nevertheless, the article's claims are substantial, with remarkable implications for politics if they were to be confirmed. So I downloaded the data and code and tried to reproduce the results. Most of the analysis was done in Stata, which I don't have access to, but I saw that there was an R script to generate Figure 2 of the study that analysed the Swiss referendum results, so I ran that.</div><div style="line-height: 24px; text-align: left;"><br /></div><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhCFz-44bgJZEnE4sY8ZrYO9krAR7U1ERa2fZTt-0GAAzpQPg1-Xlyh5CQepxUZO-vrD647URDiYyX_mXZRxNWXv29RhKvzzWrq34OELig_DDustXLH3s_-P5hEHDwIUGLDmil15_sLCi6hWQtnVNlm2WmPp5O-SNGzLIhuqpu48lAEj2rYTGQIrHVhXVmE/s1618/Fig2-full.png" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="1078" data-original-width="1618" height="426" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhCFz-44bgJZEnE4sY8ZrYO9krAR7U1ERa2fZTt-0GAAzpQPg1-Xlyh5CQepxUZO-vrD647URDiYyX_mXZRxNWXv29RhKvzzWrq34OELig_DDustXLH3s_-P5hEHDwIUGLDmil15_sLCi6hWQtnVNlm2WmPp5O-SNGzLIhuqpu48lAEj2rYTGQIrHVhXVmE/w640-h426/Fig2-full.png" width="640" /></a></div><div class="separator" style="clear: both; text-align: center;">My reproduction of the original Figure 2 from the article. The regression coefficient for the line in the "Regulatory Focus Difference" condition is B=0.545 (p=0.00006), suggesting that every 1km/h increase in wind speed produces an increase of more than half a percentage point in the vote for the prevention-oriented campaign.</div><div class="separator" style="clear: both; text-align: left;"><br /></div><h4 style="line-height: 24px; text-align: left;">Catastrophic data problems</h4><div style="line-height: 24px; text-align: left;">I had no problem in reproducing Figure 2 from the article. However, when I looked a little closer at the dataset (**) I noticed a big problem in the numbers. Take a look at the "DewPoint" and "Humidity" variables for "Election 50", which corresponds to Referendum 24 (***) in the Supplementary Information, and see if you can spot the problem.</div><div style="line-height: 24px; text-align: left;"><br /></div><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEigu0WHcm6ObqeZsTyQsDhsQuyb1W_XlCHiBlIWXillhQH1Lsh5esn0xpa5U53FRnb538SZzYNLqb4IdLqxzdkfQdiH0wRtD-ZQXAq8nIRtjV0jj0ol8D2XWNu8ryqOmbhaJR_x2Dd-FiOIbdvRN--mR5dxZSAMuh5UV0BTQ5ppawRThICLX_Zw5lFO2H_B/s685/Ref50.png" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="685" data-original-width="599" height="640" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEigu0WHcm6ObqeZsTyQsDhsQuyb1W_XlCHiBlIWXillhQH1Lsh5esn0xpa5U53FRnb538SZzYNLqb4IdLqxzdkfQdiH0wRtD-ZQXAq8nIRtjV0jj0ol8D2XWNu8ryqOmbhaJR_x2Dd-FiOIbdvRN--mR5dxZSAMuh5UV0BTQ5ppawRThICLX_Zw5lFO2H_B/w560-h640/Ref50.png" width="560" /></a></div><br /><div style="line-height: 24px; text-align: left;">Neither of those variables can possibly be correct for "Election 50" (note that the same issues affect the records for every "State", i.e., Swiss canton):</div><div style="line-height: 24px; text-align: left;"><ul style="text-align: left;"><li style="line-height: 24px;">DewPoint, which would normally be a Fahrenheit temperature a few degrees below the actual air temperature, contains numbers between 0.401 and 0.626. The air temperature ranges from 45.3 to 66.7 degrees. For the dew point temperatures to be correct would require the relative humidity to be around 10% (<a href="https://www.calculator.net/dew-point-calculator.html" target="_blank">calculator</a>), which seems unlikely in Switzerland on a mild day in May. Perhaps these DewPoint values in fact correspond to the relative humidity?</li><li style="line-height: 24px;">Humidity (i.e., relative atmospheric humidity), which by definition should be a fraction between 0 and 1, is instead a number in the range from 1008.2 to 1015.7. I am not quite sure what might have caused this. These numbers look like they could represent some measure of atmospheric pressure, but they only correlate at 0.538 with the "Pressure" variable for "Election 50".</li></ul><div>To evaluate the impact of these strange numbers on the authors' model, I modified their R script, Swiss_Analysis.R, to remove the records for "Election 50" and obtained this result from the remaining 23 referendums:</div><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhixAqUNIw90lbLP1SOsGZYf25u77GFEfjRqqdU-hsh_ZQGK1macNl3aHvmQXEL4QgbLA4SCRFURGBw8fUBTGp4CnV-w9vWN0p3oiJQG8_CbrzpVwfRujl-pcUp_rjWL6Rv_tu1nfqEIoZki_UP1IH3f56GEH2F1xHoMvqATDGPvpujHPhbS8xErKvwIE2T/s1618/Fig2-no50.png" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="1076" data-original-width="1618" height="426" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhixAqUNIw90lbLP1SOsGZYf25u77GFEfjRqqdU-hsh_ZQGK1macNl3aHvmQXEL4QgbLA4SCRFURGBw8fUBTGp4CnV-w9vWN0p3oiJQG8_CbrzpVwfRujl-pcUp_rjWL6Rv_tu1nfqEIoZki_UP1IH3f56GEH2F1xHoMvqATDGPvpujHPhbS8xErKvwIE2T/w640-h426/Fig2-no50.png" width="640" /></a></div><div style="line-height: 24px; text-align: left;"><span style="text-align: center;"><div class="separator" style="clear: both; line-height: 18px;">Figure 2 with "Election 50" (aka Referendum 24) removed from the model.</div></span></div><div style="line-height: 24px; text-align: left;"><br /></div><div style="line-height: 24px; text-align: left;">The angle of the regression line on the right is considerably less jaunty in this version of Figure 2. The coefficient has gone from B=0.545 (SE=0.120, <i>p</i>=0.000006) to B=0.266 (SE=0.114, <i>p</i>=0.02), simply by removing the damaged data that were apparently causing havoc with the model.</div><div style="line-height: 24px; text-align: left;"><br /></div><h4 style="line-height: 24px; text-align: left;">How robust is the model now?</h4><div style="line-height: 24px; text-align: left;">A <i>p</i> value of 0.02 does not seem like an especially strong result. To test this, after removing the damaged data for "Election 50", I iterated over the dataset removing a further different single "Election" each time. In seven cases (removing "Election" 33, 36, 39, 40, 42, 46, or 47) the coefficient for the interaction in the resulting model had a <i>p</i> value above the conventional significance level of 0.05. In the most extreme case, removing "Election 40" (i.e., Referendum 14, "Mindestumwandlungsgesetz") caused the coefficient for the interaction to drop to 0.153 (SE=0.215, <i>p</i>=0.478), as shown in the next figure. It seems to me that if the statistical significance of an effect disappears with the omission of just one of the 23 (****) valid data points in 30% of the possible cases, this could indicate a lack of robustness in the effect.</div><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhr8S-Yjttt8F8SS2jQ-40tuxmDhh7ZPCOxLEYeHPSo1tLSyhBkkdZzEsac_SDqY5-a884I74OovHdK5E4TdIie-RIJduX__y0G8Mvo-U0rrSWBn7T_N_xHTjbwDqnNSLjws47wgW08AV1MDmfcTpsglt0kSINI7zUgcjM8J4Rjv_0rfHT-S-2VsCcB57uL/s1618/Fig2-no50-no40.png" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="1078" data-original-width="1618" height="426" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhr8S-Yjttt8F8SS2jQ-40tuxmDhh7ZPCOxLEYeHPSo1tLSyhBkkdZzEsac_SDqY5-a884I74OovHdK5E4TdIie-RIJduX__y0G8Mvo-U0rrSWBn7T_N_xHTjbwDqnNSLjws47wgW08AV1MDmfcTpsglt0kSINI7zUgcjM8J4Rjv_0rfHT-S-2VsCcB57uL/w640-h426/Fig2-no50-no40.png" width="640" /></a></div><div><div style="line-height: 24px;"><span style="text-align: center;"><div class="separator" style="clear: both; line-height: 18px;">Figure 2 with "Election 50" (aka Referendum 24) and "Election 40" (aka Referendum 14) removed from the model.</div></span></div></div></div><h4 style="line-height: 24px; text-align: left;">Other issues</h4><div style="line-height: 24px; text-align: left;"><u>Temperature precision</u></div><div style="line-height: 24px;">The ambient temperatures on the days of the referendums (variable "Temp") are reported with eight decimal places. It is not clear where this (apparently spurious) precision could have come from. Judging from their range the temperatures would appear to be in degrees Fahrenheit, whereas one would expect the original Swiss meteorological data to be expressed in degrees Celsius. However, the conversion between the two scales is simple (F = C * 1.8 + 32) and cannot introduce more than one extra decimal place. The authors state that "Weather data were collected from www.forecast.io/raw/", but unfortunately that link redirects to a <a href="https://support.apple.com/en-us/HT213526" target="_blank">page</a> that suggests that this source is no longer available.</div><div style="line-height: 24px;"><br /></div><div style="line-height: 24px; text-align: left;"><u>Cloud cover</u></div><div style="line-height: 24px;">The "CloudCover" variable takes only eight distinct values across the entire dataset, namely 2, 3, 5, 6, 8, 24, 34, and 38. It is not clear what these values represent, but it seems unlikely that they (all) correspond to a percentage or fraction of the sky covered by clouds. Yet, this variable is included in the regression models as a linear predictor. If the values represent some kind of ordinal or even nominal coding scheme, rather than being a parameter of some meteorological process, then including this variable could have arbitrary consequences for the regression (after all, 24, 34, and 38 might equally well have been coded ordinally as 9, 10, and 11, or perhaps nominally as -99, -45, and 756). If the intention is for these numbers to represent obscured eighths of the sky ("<a href="https://worldweather.wmo.int/oktas.htm" target="_blank">oktas</a>"), then there is clearly a problem with the values above 8, which constitute 218 of the 624 records in the dataset (34.9%).</div><div style="line-height: 24px; text-align: left;"><u><br /></u></div><div style="line-height: 24px; text-align: left;"><u>Income</u></div><div style="line-height: 24px;">It would also be interesting to know the source of the "Income" data for each Swiss canton, and what this variable represents (e.g., median salary, household income, gross regional product, etc). After extracting the income data and canton numbers, and converting the latter into names, I consulted several Swiss or Swiss-based colleagues, who expressed skepticism that the cantons of Schwyz, Glarus, and Jura would have the #1, #3, and #4 incomes by any measure. I am slightly concerned that there may have been an issue with the sorting of the cantons when the Income variable was populated. The Supplementary Information says "Voting and socioeconomic information was obtained from the Swiss Federal Office of Statistics (Bundesamt für Statistik 2015)", and that reference points to a web page entitled “Detaillierte Ergebnisse Der Eidgenössischen Volksabstimmungen” with URL http://www.bfs.admin.ch/bfs/portal/de/index/themen/17/03/blank/data/01.html, but that link is dead (and in any case, the title means "Detailed results of Federal referendums"; such a page would generally not be expected to contain socioeconomic data).</div><div style="line-height: 24px;"><br /></div><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEh5E_-vwKAMMweJn-Y1egldyzBjX9eBNYA6gZlTAo1IdEx6OHwvImmO8p0K2R81xJIHPf3XPR24OifyaPBAW7zqAg-k3uK3NU7RDeonX3VqQuYYsM60LPTjg25hVWnTdPIUMJG2G-TaXdc-vKHsObBORDcqIcKe-RW9ayp45v0-1O2crDVfC_IPKBd_YnKs/s499/income.png" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="499" data-original-width="440" height="640" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEh5E_-vwKAMMweJn-Y1egldyzBjX9eBNYA6gZlTAo1IdEx6OHwvImmO8p0K2R81xJIHPf3XPR24OifyaPBAW7zqAg-k3uK3NU7RDeonX3VqQuYYsM60LPTjg25hVWnTdPIUMJG2G-TaXdc-vKHsObBORDcqIcKe-RW9ayp45v0-1O2crDVfC_IPKBd_YnKs/w564-h640/income.png" width="564" /></a></div><div class="separator" style="clear: both; text-align: center;"><br /></div><div style="line-height: 24px;"><div class="separator" style="clear: both; text-align: center;"></div><div><div style="line-height: 24px;"><span style="text-align: center;"><div class="separator" style="clear: both; line-height: 18px;">Swiss cantons (using the "constitution order" mapping from numbers to names) and their associated "Income", presumably an annual figure in Swiss francs. Columns "Income(Mo)" and the corresponding rank order "IncRank" are from Mo et al.'s dataset; "Statista" and "StatRank" are from statista.com.</div><div class="separator" style="clear: both; line-height: 18px; text-align: left;"><br /></div><div class="separator" style="clear: both; line-height: 18px; text-align: left;"><div style="line-height: 24px;">I obtained some fairly recent Swiss canton-level household income data from <a href="https://de.statista.com/statistik/daten/studie/1215204/umfrage/durchschnittliches-jahreseinkommens-in-der-schweiz-nach-kantonen/" target="_blank">here</a> and compared it with the data from the article. The results are shown in the figure above. The Pearson correlation between the two sets of numbers was 0.311, with the rank-order correlation being 0.093. I think something may have gone quite badly wrong here.</div><div style="line-height: 24px;"><br /></div></div></span></div></div></div><div style="line-height: 24px; text-align: left;"><u>Turnout</u></div><div style="line-height: 24px;">The value of the "Turnout" variable is the same for all cantons. This suggests that the authors may have used some national measure of turnout here. I am not sure how much value such a variable can add. The authors note (footnote 12, p. 17) that "We found that, except for one instance, no other weather indicator was correlated with the number of prevention-focused votes without simultaneously also affecting turnout rates. Temperature was an exception, as increased temperature was weakly correlated with a decrease in prevention-focused vote and not correlated with turnout". It is not clear to me what the meaning would be of calculating a correlation between canton-level temperature and national-level turnout.</div><div style="line-height: 24px;"><br /></div><div style="line-height: 24px; text-align: left;"><u>Voting results do not always sum to 1</u></div><div style="line-height: 24px;">Another minor point about whatever cleaning has been performed on the dataset is that in 68 out of 624 cases (10.9%), the sum of "VotingResult1" and "VotingResult2" — representing the "Yes" and "No" votes — is 1.01 and not 1.00. Perhaps this is the result of the second number being generated by the first being subtracted from 1.00 when the first number was expressed as a percentage with one decimal place, with both numbers subsequently being rounded and something ambiguous happening with the last digit 5. In any case, it would seem important for these two numbers to sum to 1.00. This might not make an enormous amount of difference to the results, but it does suggest that the preparation of the data file may not have been done with excessive care.</div><div style="line-height: 24px;"><br /></div><div style="line-height: 24px;"><div style="line-height: 24px; text-align: left;"><u>Mean-centred variables</u></div><div style="line-height: 24px;">Two of the control variables, "Pressure" and "CloudCover", appear in the dataset in two versions, raw and mean-centred. There doesn't seem to be any reason to mean-centre these variables, but it is something that is commonly done when calculating interaction terms. I wonder whether at some point in the analyses the authors tested atmospheric pressure and cloud cover, rather than wind speed, as possible drivers of an effect on voting. Certainly there seems to be quite a lot of scope for the authors to have wandered around <a href="https://statmodeling.stat.columbia.edu/2021/03/16/the-garden-of-forking-paths-why-multiple-comparisons-can-be-a-problem-even-when-there-is-no-fishing-expedition-or-p-hacking-and-the-research-hypothesis-was-posited-ahead-of-time-2/" target="_blank">Andrew Gelman's "Garden of forking paths"</a> in these analyses, which do not appear to have been pre-registered.</div><div style="line-height: 24px;"><br /></div></div><div style="line-height: 24px;"><div style="text-align: left;"><u>No measure of population</u></div></div><div style="line-height: 24px;">Finally, a huge (to me, anyway) limitation of this study is that there is no measure of, or attempt to weight the results by, the population of the cantons. The most populous Swiss canton (Zürich) has a population about 90 times that of the least populous (Appenzell Innerrhoden), yet the cantons all have equal weight in the models. The authors barely mention this as a limitation; they only mention the word "population" once, in the context of determining the average wind speed in Study 1. Of course, the <a href="https://web.stanford.edu/class/ed260/freedman549.pdf" target="_blank">ecological fallacy</a> [.pdf] is always lurking whenever authors try to draw conclusions about the behaviour of individuals, whether or not the population density is taken into account, although this did not stop the authors from claiming in their abstract that "we find that <i>individuals</i> [emphasis added] exposed to higher wind speeds become more prevention-focused and more likely to support prevention-focused electoral options", or (on p. 4) stating that "We ... tested whether higher wind speed increased <i>individual’s</i> [punctuation <i>sic</i>; emphasis added] prevention focus".</div><div style="line-height: 24px;"><br /></div><h4 style="line-height: 24px; text-align: left;">Conclusion</h4><div style="line-height: 24px;">I wrote this post principally to draw attention to the obviously damaging errors in the records for "Election 50" in the Swiss data file. I have also written to the authors to report those issues, because these are clearly in need of urgent correction. Until that has happened, and perhaps until someone else (with access to Stata) has conducted a re-analysis of the results for both the "Swiss" and "Brexit/Scotland" studies, I think that caution should be exercised before citing this paper. The other issues that I have raised in this post are, of course, open to critique regarding their importance or relevance. For the avoidance of doubt, given the nature of some of the other posts that I have made on this blog, I am not suggesting that anything untoward has taken place here, other than perhaps a degree of carelessness.</div><div style="line-height: 24px;"><br /></div><h4 style="line-height: 24px;">Supporting files</h4><div style="line-height: 24px;">I have made my modified version of Mo et al.'s code to reproduce Figure 2 available <a href="http://nickbrown.fr/blog/mo-et-al" target="_blank">here</a>, in the file "(Nick) Swiss_Analysis.R". If you decide to run it, I encourage you to use the authors' original data file ("Swiss.dta") from the ZIP file that can be downloaded from the link at the top of this post. However, as a convenience, I have made a copy of this file available along with my code. In the same place you will also find a small Excel table ("Cantons.xls") containing data for my analysis of the canton-level income question.</div><div style="line-height: 24px;"><br /></div><h4 style="line-height: 24px;">Acknowledgements</h4><div style="line-height: 24px;">Thanks to Jean-Claude Fox for doing some further digging on the Swiss income numbers after this post was first published.</div><div style="line-height: 24px;"><br /></div><h4 style="line-height: 24px; text-align: left;">Footnotes</h4><div style="line-height: 24px;">(*) Interestingly, the title of Table 1 and, even more explicitly, the footnote on p. 10 ("Remain" with an uppercase initial letter) suggest that the authors may have been aware that the actual voting choices were "Remain" and "Leave". Perhaps these were simplified to "No" and "Yes", respectively, for consistency with the reports of the Scottish independence referendum; but if so, this should have been reported.</div><div style="line-height: 24px;">(**) I exported the dataset from Stata's .dta format to .csv format using rio::convert(). I also confirmed that the errors that I report in this post were present in the Stata file by inspecting the data structure after the Stata file had been read in to R.</div><div style="line-height: 24px;">(***) The authors coded the Swiss referendums, which are listed with numbers 1–24 in the Supplementary Information, as 27–50, by adding 26. They also coded the 26 cantons of Switzerland as 51–76, apparently by adding 50 to the constitutional order number (1 = Zürich, 26 = Jura; see <a href="https://en.wikipedia.org/wiki/Cantons_of_Switzerland#List" target="_blank">here</a>), perhaps to ensure that no small integer that might creep into the data would be seen as either a valid referendum or canton (a good practice in general). I was able to check that the numerical order of the "State" variable is indeed the same as the constitutional order by examining the provided latitude and longitude for each canton on Google Maps (e.g., "State" 67, corresponding to the canton of St Gallen with constitutional order 17, has reported coordinates of 47.424482, 9.376717, which are in the centre of the town of St Gallen).</div><div style="line-height: 24px;">(****) I am not sure whether a single referendum in 26 cantons represents 1 or 26 data points. The results from one canton to the next are clearly not independent. I suppose I could have written "4.3% of the data" here.</div><div style="line-height: 24px;"><br /></div><div style="line-height: 24px;"><br /></div>Nick Brownhttp://www.blogger.com/profile/07481236547943428014noreply@blogger.com2tag:blogger.com,1999:blog-7890764972166411105.post-48281905396059489462023-07-02T01:48:00.015+02:002023-07-03T18:16:39.522+02:00Strange numbers in the dataset of Zhang, Gino, & Norton (2016)<div style="line-height: 24px;">
<p>In this post I'm going to be discussing this article:</p>
<div style="margin-left: 36pt; text-indent: -36pt;"><span style="text-indent: 0px;">Zhang, T., Gino, F., & Norton, M. I. (2016). The surprising effectiveness of hostile mediators. <i>Management Science</i>, <i>63</i>(6)<i>, </i>1972–1992. https://doi.org/10.1287/mnsc.2016.2431</span></div><div style="margin-left: 36pt; text-indent: -36pt;"><span style="text-indent: 0px;"> </span></div><div style="margin-left: 36pt; text-indent: -36pt;"><span style="text-indent: 0px;">You can download the article from <a href="https://www.hbs.edu/ris/download.aspx?name=zhang%20gino%20norton.pdf" target="_blank">here</a> and the dataset from <a href="https://doi.org/10.1287/mnsc.2016.2431" target="_blank">here</a>.</span></div><div style="margin-left: 36pt; text-indent: -36pt;"><span style="text-indent: 0px;"><br /></span></div>[[ Begin update 2023-07-03 16:12 UTC ]]<div style="margin-left: 36pt; text-indent: -36pt;"></div><div style="text-align: left;"><span style="text-indent: 0px;">Following feedback from several sources, I now see how it is in fact possible that these data could have been the result of using a slider to report the amount of money being requested. I still think that this would be a terrible way to design a study (see my previous update, below), as it causes a loss of precision for no obvious good reason compared to having participants type a maximum of 6 digits, and indeed the input method is not reported in the article. However, if a slider was used, then with multiple platforms the observed variety of data could have arisen.</span></div><div style="text-align: left;"><span style="text-indent: 0px;"> </span></div><div style="text-align: left;"><span style="text-indent: 0px;">In the interests of transparency I will leave this post up, but with the caveat that readers should apply caution in interpreting it until we learn the truth from the various inquiries and resolution exercises that are ongoing in the Gino case. <br /></span></div><div style="margin-left: 36pt; text-indent: -36pt;"><span style="text-indent: 0px;">[[ End update 2023-07-03 16:12 UTC ]]<br /></span></div><p>The
focus of my interest here is Study 5. Participants (MTurk workers) were
asked to imagine that they were a carpenter who had been given a
contract to furnish a number of houses, but decided to use better
materials than had been specified and so had overspent the budget by
$300,000. The contractor did not want to reimburse them for this. The
participants were presented with interactions that represented a
mediation process (in the social sense of the word "mediation", not the statistical one) between the
carpenter and the contractor. The mediator's interactions were
portrayed as "Nice", "Bilateral hostile" (nasty to both parties), or
"Unilateral hostile" (nasty to the carpenter only). After this exercise,
the participants were asked to say how much of the $300,000 they would
ask from the contractor. This was the dependent variable to show how
effective the different forms of mediation were.</p><p>The authors reported (p. 1986):</p><p></p><p style="margin-left: 40px; text-align: left;">We
conducted a between-subjects ANOVA using participants’ demands from
their counterpart as the dependent variable. This analysis revealed a
significant effect for mediator’s level and directedness of hostility,
F(2, 134) = 6.86, p < 0.001, partial eta² = 0.09. Post hoc tests
using LSD corrections indicated that participants in the bilateral
hostile mediator condition demanded less from their counterpart (M = $149,457, SD = 65,642) compared with participants in the unilateral
hostile mediator condition (M = $208,807, SD = 74,379, p < 0.001)
and the nice mediator condition (M = $183,567, SD = 85,616, p =
0.04). The difference between the latter two conditions was not
significant (p = 0.11).<br /></p>Now, imagine that you are a participant
in this study. You are being paid $1.50 to pretend to be someone who
feels that they are owed $300,000. How much are you going to ask for?
I'm guessing you might ask for all $300,000; or perhaps you are prepared
to compromise and ask for $200,000; or you might split the difference
and ask for $150,000; or you might be in a hurry and think that 0 is the
quickest number to type.</div><div style="line-height: 24px;"><br /></div><div style="line-height: 24px;">Let's look at what the participants actually entered. In this table, each cell is one participant; I have arranged them in columns, with each column being a condition, and sorted the values in ascending order.</div><br /><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/a/AVvXsEhxr8uX_fbvYN0NWH8K2x9YHKFYf6xCjTW6hQYpuSfF7Id3WpOQsGjAe1ZqWoecLwa2JkcesNh6l6hNKVdsAuMQQ6LM_MMGnSTEGTAm38_e9LVNvtegM7icG8gT3g4VKoXYcEY4aklzyI1bbLndQY0UZSLF4Wa0kopCQsDyTFeHd_w9_jwYapmd274Y_oUK" style="margin-left: 1em; margin-right: 1em;"><img alt="" data-original-height="924" data-original-width="483" src="https://blogger.googleusercontent.com/img/a/AVvXsEhxr8uX_fbvYN0NWH8K2x9YHKFYf6xCjTW6hQYpuSfF7Id3WpOQsGjAe1ZqWoecLwa2JkcesNh6l6hNKVdsAuMQQ6LM_MMGnSTEGTAm38_e9LVNvtegM7icG8gT3g4VKoXYcEY4aklzyI1bbLndQY0UZSLF4Wa0kopCQsDyTFeHd_w9_jwYapmd274Y_oUK=s16000" /></a></div><div><div style="line-height: 24px;"></div><div style="line-height: 24px;"></div><div style="line-height: 24px;"></div><div style="line-height: 24px;"><h4 style="text-align: left;"><span style="font-weight: normal;"> </span></h4><h4 style="text-align: left;"><span style="font-weight: normal;">This
makes absolutely no sense. Not only did 95 out of 139 participants
choose a number that wasn't a multiple of $1,000, but also, they chose
remarkably similar non-round numbers. Twelve participants chose to ask for exactly
$150,323 (and four others asked for $10,323, $170,323, or $250,323).
Sixteen participants asked for exactly $150,324. Ten asked for $149,676,
which interestingly is equal to $300,000 minus the aforementioned
$150,324. There are several other six-digit, non-round numbers that
occur multiple times in the data. Remember, every number in this table represents the response of an independent MTurk worker, taking the survey in different locations across the United States.<br /></span></h4><div><span style="font-weight: normal;"><br /></span></div><h4 style="text-align: left;"><span style="font-weight: normal;">To coin a phrase, it is not clear how these numbers could have arisen as a result of a natural process. If the authors can explain it, that would be great.</span></h4><div><span style="font-weight: normal;"><br /></span></div><div>[[ Begin update 2023-07-02 12:38 UTC ]]</div><div><br /></div><div>Several people have asked, on Twitter and in the comments here, whether these numbers could be explained by a slider having been used, instead of a numerical input field. I don't think so, for several reasons:</div><div><ol><li>It makes no sense to use a slider to ask people to indicate a dollar amount. It's a number. The authors report the mean amount to the nearest dollar. They are, ostensibly at least, interested in capturing the precise dollar amount.</li><li>Had the authors used a slider, they would presumably have done so for a very specific reason, which one would imagine they would have reported.</li><li>Some of the values reported are 153,023, 153,024, 150,647, 150,972, 151,592, and 151,620. The differences between these values are 1, 323, 325, 620, and 28. In another sequence, we see 180,130, 180,388, and 180,778, separated by 258 and 390; and in another, 200,216, 200,431, 200,864, and 201,290, separated by 215, 433, and 426. Even if we assume that the difference of 1 is a rounding error, in order for a slider to have the granularity to be able to indicate all of those numbers while also covering the range from 0 to 300,000, it would have to be many thousands of pixels wide. Real-world on-screen sliders typically run from 0 to 100 or 0 to 1000, with each of 400 or 500 pixels representing perhaps 0.20% or 0.25% of the available range.</li></ol></div><div><div>Of course, all of this could be checked if someone had access to the original Qualtrics account. Perhaps Harvard will investigate this paper too...</div><div><br /></div><div>[[ End update 2023-07-02 12:38 UTC ]]</div><div><br /></div></div><h4 style="text-align: left;"><span style="font-weight: normal;"> </span></h4><h4 style="text-align: left;">Acknowledgements</h4><p>Thanks to James Heathers for useful discussions, and to Economist 1268 at econjobrumors.com for suggesting that this article might be worth looking at.<br /></p><p><br /></p><p></p><p><br /></p>
</div></div>Nick Brownhttp://www.blogger.com/profile/18266307287741345798noreply@blogger.com14tag:blogger.com,1999:blog-7890764972166411105.post-47565802942734030742023-06-28T14:57:00.006+02:002023-06-28T14:57:46.407+02:00A coda to the Wansink story<div style="line-height: 24px;">
<p>The investigation of scientific misconduct by Ivy League universities is <a href="https://www.nytimes.com/2023/06/24/business/economy/francesca-gino-harvard-dishonesty.html" target="_blank">once again in the news</a> at the moment, which prompts me to write up something that I should have written up quite a while ago. (The time I spend <a href="https://mriyareport.org/our-staff/" target="_blank">thinking about, and trying to help people understand, the Russian invasion of Ukraine</a> has made as big a dent in my productivity as Covid-19.)</p><p>On October 31, 2018, I sent <a href="https://docs.google.com/document/u/2/d/1HSisU6VgYwxq_DvueBq_goU4f6YLQ-dYcDZmwvHn9RY/" target="_blank">an open letter</a>, signed by me and 50 colleagues, to Cornell. In it, I asked that they release the report the full text of the report of their inquiry into the misconduct of Professor Brian Wansink. On November 5, 2018, I received <a href="https://www.documentcloud.org/documents/5028990-BrownandFellowSignatories-11-05-18.html" target="_blank">a reply</a> from Michael Kotlikoff, the Provost of Cornell. He explained why the full text of the report was not being released (an explanation that <a href="https://retractionwatch.com/2018/11/06/its-time-to-end-the-code-of-silence-at-universities/" target="_blank">did not impress Ivan Oransky at Retraction Watch</a>), and added the following:</p><blockquote style="border: none; margin: 0px 0px 0px 40px; padding: 0px;"><div style="line-height: 24px;"><p style="text-align: left;">Cornell is now conducting a Phase II investigation to determine the degree to which any acts of research misconduct may have affected federally (NIH and USDA) funded research projects. ... As part of Phase II of the university’s investigation, Cornell has required Professor Wansink to collect and submit research data and records for all of his publications since 2005, when he came to the university, so that those records may be examined. <i>We will provide a summary of this Phase II investigation at its conclusion</i>. [emphasis added]</p></div></blockquote><p>The Wansink story faded into the background after that, but a few months ago a small lightbulb fizzled into life above my head and I decided to find out what happened to that Phase II report. So I wrote to Provost Kotlikoff. He has kindly given me permission to quote his response verbatim:<br /></p></div><blockquote style="border: none; margin: 0px 0px 0px 40px; padding: 0px; text-align: left;"><div style="line-height: 24px;"><p>Following my November 5 letter we indeed conducted a comprehensive Phase II analysis, but this was restricted to those scientific papers from Professor Wansink’s group that identified, or could be linked to, support from federal funds. This analysis, which was conducted on a subset of papers and followed federal guidelines, was reported to the NIH and to the USDA (the relevant funding organizations), and accepted by them. I should point out that this Phase II analysis did not include many of the papers identified by you and others as failing to meet scientific norms, as those were not associated with federal support, and therefore was not a comprehensive summary of the scientific issues surrounding Professor Wansink’s work.</p></div><div style="line-height: 24px;"><p>I am sorry to say that Cornell does not release scientific misconduct reports provided to the NIH and the USDA. However, I believe that Cornell has appropriately addressed the scientific concerns that were identified by you and others (for which I thank you), and considers this matter closed.</p></div></blockquote><div style="line-height: 24px;"><p>So that seems to be it. We are apparently not going to see a summary of the Phase II investigation. Perhaps it was Cornell's initial intention to release this, but they were unable to do so for legal reasons. In any case, it's a little disappointing.</p><p><br /></p>
Nick Brownhttp://www.blogger.com/profile/07481236547943428014noreply@blogger.com2tag:blogger.com,1999:blog-7890764972166411105.post-70214402781080771332023-03-10T00:59:00.024+01:002023-07-11T21:07:49.156+02:00Some interesting discoveries in a shared dataset: Néma et al. (2022).<div style="line-height: 24px;">
<p>In this post I'm going to be discussing this article, but mostly its dataset:</p>
<div style="margin-left: 36pt; text-indent: -36pt;"><span style="text-indent: 0px;">Néma, J., Zdara, J., Lašák, P., Bavlovič, J., Bureš, M., Pejchal, J., & Schvach, H. (2023). Impact of cold exposure on life satisfaction and physical composition of soldiers. <i>BMJ Military Health</i>. Advance online publication. https://doi.org/10.1136/military-2022-002237</span></div><p>The article itself doesn't need much commentary from me, since it has already been covered by Stuart Ritchie on Twitter <span><a href="https://twitter.com/StuartJRitchie/status/1632360272710062080" style="background-color: white;" target="_blank">here</a></span> and in his iNews column <span><a href="https://inews.co.uk/news/new-ice-bath-lose-weight-science-2191656" style="background-color: white;" target="_blank">here</a></span>, as well as by Gideon Meyerowitz-Katz on Twitter <a href="https://twitter.com/GidMK/status/1633318873352241152" target="_blank">here</a>. So I will just cite or paraphrase some sentences from the Abstract:</p><blockquote style="border: none; margin: 0px 0px 0px 40px; padding: 0px;"><p style="text-align: left;">[T]he aim of this study was to examine the effect of regular cold exposure on the psychological status and physical composition of healthy young soldiers in the Czech Army. A total of 49 (male and female) soldiers aged 19–30 years were randomly assigned to one of the two groups (intervention and control). The participants regularly underwent cold exposure for 8 weeks, in outdoor and indoor environments. Questionnaires were used to evaluate life satisfaction and anxiety, and an "InBody 770" device was used to measure body composition. Among other results, systematic exposure to cold significantly lowered perceived anxiety (p=0.032). Cold water exposure can be recommended as an addition to routine military training regimens and is likely to reduce anxiety among soldiers.</p></blockquote><p>The article PDF file contains a <a href="https://www.vyzkumodolnosti.cz/en/datasets" target="_blank">link to a repository</a> in which the authors originally placed an Excel file of their main dataset named "Dataset_ColdExposure_sorted_InBody.xlsx" (behind the link entitled "InBody - Body Composition"). I downloaded this file and explored it, and found some interesting things that complement the investigations of the article itself; these discoveries form the main part of this blog post.</p><p>Recently, however—probably in reaction to the authors being warned by Gideon or someone else that their dataset contained personally identifying information (PII)—this file has been replaced with one named "Datasets_InBody+WC_ColdExposure.csv". I will discuss the new file near the end of this post, but for now, the good news is that the file containing PII is no longer publicly available.</p><p>[[ Update 2023-03-11 23:23 UTC: Added new information here about the LSQ dataset, and—further along in this post—a paragraph about the analysis of these data. ]]</p><div>The repository also contains a data file called "Dataset_ColdExposure_LSQ.csv", which represents the participants' responses at two timepoints to the Life Satisfaction Questionnaire. I downloaded this file and attempted to match the participant data across the two datasets.</div><div><br /></div><h4 style="text-align: left;">The structure of the main dataset file</h4><p>The Excel file that I downloaded contains six worksheets. Four of these contain the data for the two conditions that were reported in the article (Cold and Control), one each at baseline and at the end of the treatment period. Within those worksheets, participants are split into male and female, and within each gender a sequence number starting at 1 identifies each participant. A fifth worksheet named "InBody Začátek" contains the baseline data for each participant, and a sixth, named "InBody1", appears to contain every data record for each participant, as well as some columns which, while mostly empty, appear designed to hold contact information for each person.</p><h4 style="text-align: left;">Every participant's name and date of birth is in the file (!)</h4><p>The first and most important problem in the file as it was uploaded, and was in place until a couple of days ago, is that a lot of PII was left in there. Specifically, the file contained <i>the first and last names and date of birth of every participant</i>. This study was carried out in the Czech Republic, and I am not familiar with the details of research ethics in that country, but it seems to me to be pretty clear that it is not acceptable to <a href="https://www.wma.net/policies-post/wma-declaration-of-helsinki-ethical-principles-for-medical-research-involving-human-subjects/" target="_blank">conduct before-and-after physiological measurements on people and then publish those numbers</a> along with information that in most cases probably identifies them uniquely among the population of their country.</p><p>I have modified the dataset file to remove this PII before I share it. Specifically, I did this:</p><p></p><ol style="text-align: left;"><li>Replaced the names of participants with random fake names assembled from lists of popular English-language first and last names. I use these names below where I need to identify a particular participant's data.</li><li>Replaced the date of birth with a fake date consisting of the same year, but a random month and day. As a result of this, the "Age" column, which appears to have been each participant's age at their last birthday before they gave their baseline data, may no longer match the reported (fake) date of birth.</li></ol><p></p><h4 style="text-align: left;">What actually happened in the study?</h4><p>The Abstract states (see above) that 49 participants were in two conditions: exposure to Cold (<i>Chlad</i>, in Czech) and a no-treatment Control group (<i>Kontrolní</i>). But in the dataset there are 99 baseline measurement records, and the participants are recorded as being in <i>four</i> conditions. As well as Cold and Control, there is a condition called Mindfulness (the English word is used), and another called <i>Spánek</i>, which means Sleep in Czech.</p><p>This is concerning because these additional participants and conditions are not mentioned in the article. The Method section states that "A total of 49 soldiers (15 women and 34 men) participated in the study and were randomly divided into two groups (control and intervention) before the start of the experiment". If the extra participants and conditions were part of the same study, this should have been reported; the above sentence, as written, seems to be stretching the idea of innocent omission quite a bit. Omitting conditions and participants is a powerful "researcher degree of freedom" in sense of Simmons et al.'s <a href="https://doi.org/10.1177/0956797611417632" target="_blank">classic paper</a> entitled "False-Positive Psychology". If these participants and conditions were <i>not</i> part of the same study then something very strange is happening, as it would imply that there were at least two studies being conducted with the same participant ID sequence number assignment and reported in the same data file.</p><p>Data seem to have been collected in two principal waves, January (<i>leden</i>) and March (<i>březen</i>) 2022. It is not clear why this was done. A few tests seem to have been performed in late 2021 or in February, April, or May 2022, but whatever the date, all participants were assigned to one of the two month groups. For reasons that are not clear, one participant whose baseline data were collected in December 2021 ("Martin Byrne") was assigned to the March group, although his data did not end up in the final group worksheets that formed the basis of the published article. Meanwhile, six participants ("George Fletcher", "Harold Gregory", "Harvey Barton", "Nicole Armstrong", "Christopher Bishop", and "Graham Foster") were assigned to the January group even though their data were collected in March 2022 or later; five of these (all except for "Nicole Armstrong") did end up in the final group worksheets. It seems that the grouping into "January" and "March" did not affect the final analyses, but it does make me wonder what the authors had in mind in creating these groups and then assigning people to them without apparently respecting the exact dates on which the data were collected. Again, it seems that plenty of researcher degrees of freedom were available.</p><h4 style="text-align: left;">How were participants filtered out?</h4><p>There are 99 records in the baseline worksheet. These are in conditions as follows: <i>Chlad</i> (Cold), 28 (18 “<i>leden</i>/January”, 10 “<i>březen</i>/March”); <i>Kontrolní</i> (Control), 41 (33 “<i>leden</i>/January”, 8 “<i>březen</i>/March”); Mindfulness, 11 (all “<i>leden</i>/January”); <i>Spánek</i> (Sleep), 19 (11 “<i>leden</i>/January”, 8 “<i>březen</i>/March”).</p><p>Of the 28 participants in the Cold condition, three do not appear in the final Cold group worksheets that were use for the final analyses. Of the 41 in the Control condition, 17 do not appear in the final Control group worksheets. It is not clear what criteria were used to exclude these 3 (of 28) or 17 (of 41) people. The three in the Cold condition were all aged over 30, which corresponds to the reported cutoff age from the article, but does rather suggest that this cutoff might have been decided post hoc. Of the 17 people in the Control condition who did not make it to the final analyses, 11 were aged over 30, but six were not, so it is even more unclear why they were excluded. </p><p>Despite the claim by the authors that participants were aged 19–30, four people in the final Cold condition worksheets ("Anthony Day", "Hunter Dunn", "Eric Collins", and "Harvey Barton") are aged between 31 and 35.</p><h4 style="text-align: left;">Were the authors participants themselves?</h4><p>At least two, and likely five, of the participants appear to be authors of the article. I base this observation on the fact that in five cases, a participant has the same last name and initial as an author. In two of those cases, an e-mail address is reported that appears to correspond to the institution of that author. For the other three, I contacted a Czech friend, who used <a href="http://www.nasejmena.cz/nj/cetnost.php" target="_blank">this website</a> to look up the frequencies of the names in question; he told me that the last names (with any initials) only correspond to 55, 10, and 3 people in the entire Czech Republic, out of a population of 10.5 million.</p><p>Now, perhaps all of these people—one of whom ended up in the final Control group—are also active-duty military personnel, but it still does not seem appropriate for a participant in a psychological study that involves self-reported measures of one's attitudes before and after an intervention to also be an author on the associated article and hence at least implicitly involved with the design of the study. This also calls into question the randomisation and allocation process, as it is unlikely a randomised trial could have been conducted appropriately if investigators were also participants. (The article itself gives no detail about the randomisation process.)</p><h4 style="text-align: left;">Some of the participants in the final sample are duplicates (!)</h4><div><br /></div><div>The authors claimed that their sample (which I will refer to as the "final sample", given the uncertainty over the number of people who actually participated in the study) consisted of 49 people, which the reader might reasonably assume means 49 unique individuals. Yet, there are some obvious duplicates in the worksheets that describe the Cold and Control groups:</div><div><ol style="text-align: left;"><li>The participant to whom I have assigned the fake name "Stella Arnold" appears both in the Cold group with record ID #7 and in the Control group with record ID #6, both with Gender=F (there are separate sequences of ID numbers for male and female participants within each worksheet, with both sequences starting at 1, so the gender is needed to distinguish between them). The corresponding baseline measurements are to be found in rows 9 and 96 of the "InBody Začátek" (baseline measurements) worksheet.</li><li>The participant to whom I have assigned the fake name "Harold Gregory" appears both in the Cold group with record ID #15 and in the Control group with record ID #12, both with Gender=M. The corresponding baseline measurements are to be found in rows 38 and 41 of the "InBody Začátek" worksheet.</li><li>The participant to whom I have assigned the fake name "Stephanie Bird" appears twice in the Control groups with record IDs #4 and #7, both with Gender=F. The corresponding baseline measurements are to be found in rows 73 and 99 of the "InBody Začátek" worksheet.</li><li>The participant to whom I have assigned the fake name "Joe Gill" appears twice in the Control groups with record IDs #4 and #7, both with Gender=M. The corresponding baseline measurements are to be found in rows 57 and 62 of the "InBody Začátek" worksheet.</li></ol></div><div><div><div>In view of this, it seems difficult to be certain about the actual sample size of the final (two-condition) study, as reported in the article.</div><div><br /></div><div></div></div><div></div></div><h4 style="text-align: left;">Many other participants were assigned to more than one of the four conditions</h4><p>36 of the 99 records in the baseline worksheet have duplicated names. Put another way, 18 people appear to have been enrolled in the overall (four-condition) study in two different conditions. Of these, five were in the Control condition in both time periods ("<i>leden</i>/January" and "<i>březen</i>/March"); four were in the Control condition once and a non-Control condition once; and nine were recorded as being in two non-Control conditions. In 17 cases the two conditions were labelled with different time periods, but in one case ("Harold Gregory"), both conditions (Cold/Control) were labelled "<i>leden</i>/J<span style="background-color: white;">anuary". <span>This participant was one of the two who appeared in both final conditions (see previous paragraph)</span>; he is also one of the six participants assigned to a "January" group with data that were actually collected </span>in March 2022.</p><p>Continuing on this point, two records in the worksheet ("InBody1") that contains the record of all tests, both baseline and subsequent, appear to refer to the same person, as the dates of the birth are the same (although the participant ID numbers are different) and the original Czech names differ only in the addition/omission of one character; for example, if the names were English, this might be "John Davis" and "John Davies". The fake names of these two records in the dataset that I am sharing are "Anthony Day" and "Arthur Burton", with "Anthony Day" appearing in the final worksheets and being 31 years old, as mentioned above. The height and other physiological data for these two records, dated two months apart, are similar but not identical.</p><h4>Inconsistencies across the datasets</h4><p>The LSQ data contains records for 49 people, with 25 in the Cold group and 24 in the Control group, which matches the main dataset. However, there are some serious inconsistencies between the two datasets.</p><p>First, the gender split of the Control group is not the same between the datasets. In the main dataset, and in the article, there were 17 men and 7 women in this group. However, in the LSQ dataset, there are 14 men and 10 women in the Control group.</p><p>Second, the ages that are reported for the participants in the LSQ data do not match the ages in the main dataset. There is not enough information in the LSQ dataset—which has its own participant ID numbering scheme—to reliably connect individual participants across the two, but both datasets report the age of the participants and so as a minimum it should be possible to find correspondences at that level. However, this is not the case. Leaving aside the three participants who differ on gender (which I chose to do by assuming that the main database was correct, since its gender split matches the article), there are 11 other entries in the LSQ dataset where I was unable to find a corresponding match on age in the main dataset. Of those 11, three differ by just one year, which could perhaps just be explained by the participant having a birthday between two data collection timepoints, but for the other eight, the difference is at least 3 years, no matter how one arranges the records.</p><p>In summary, the LSQ dataset is inconsistent with the main dataset on 14 out of 49 (28%) of its records.</p><h4 style="text-align: left;">Other curiosities</h4><div><br /></div><div>Several participants, including two in the final Control condition, have decimal commas instead of decimal points for their non-integer values. There are also several instances in the original datafile where cells in analysed columns have numbers recorded as text. It is not clear how these mixed formats could have been either generated or (conveniently) analysed by software.</div><div><br /></div><div>Participants have a “Date of Registration”. It is not clear what this means. In 57 out of 99 cases, this date is the same as the date of the tests, which might suggest that this is the date when the participant joined the study, but some of the dates go back as far as 2009.</div><div><br /></div><div>Data were sometimes collected on more than two occasions per participant (or on more than four occasions for some participants who were, somehow, assigned to two conditions). For example, data were collected four times from "Orlando Goodwin" between 2022-03-07 and 2022-05-21, while he was in the Cold condition (to add to the two times when data were collected from him between 2022-01-13 and 2022-02-15, when he was in the Sleep condition). However, the observed values in the record for this participant in the "Cold_Group_After" worksheet suggest that the last of these four measurements to be used was the third, on 2022-05-05. The purpose of the second and fourth measurements of this person in the Cold condition is thus unclear, but again it seems that this practice could lead to abundant researcher degrees of freedom.</div><div><br /></div><div>There are three different formats for the participant ID field in the worksheet that contains all the measurement records. In the majority of the cases, the ID seems to be the date of registration in YYMMDD format, followed by a one- or two-digit sequence number, for example "220115-4" for the fourth participant registered on January 15, 2022. In some cases the ID is the letters "lb" followed by what appears to be a timestamp in YYMMDDHHMMSS format, such as "lb151210070505". Finally, one participant (fake name "Keith Gordon") has the ID "d14". This degree of inconsistency does not convey an atmosphere of rigour.</div><div><br /></div><h4 style="text-align: left;">The new data file</h4><p>I downloaded the original (XLSX format) data file on 2023-03-05 (March 5) at 20:52 UTC. That file (or at least, the link to it) was <a href="https://web.archive.org/web/20230307115947/https://www.vyzkumodolnosti.cz/en/datasets" target="_blank">still there on 2023-03-07 (March 7) at 11:59 UTC</a>. When I checked on 2023-03-08 (March 8) at 17:26 UTC the link was dead, implying that the file with PII had been removed at some intermediate point in the previous 30 hours. At some point after that a new dataset file was uploaded to the same location, which I downloaded on 2023-03-09 (March 9) at 14:27 UTC. This new file, in CSV format, is greatly simplified compare to the original. Specifically:</p><p></p><ol style="text-align: left;"><li>The data for the two conditions (Cold/Control) and the two timepoints (baseline/end) have been combined into one sheet in place of four.</li><li>The worksheets with the PII and other experimental conditions have been removed.</li><li>Most of the data fields have been removed; I assume that the remaining fields are sufficient to reproduce the analyses from the paper, but I haven't checked that as it isn't my purpose here.</li><li>Two data fields have been added. One of these, named "SMM(%)", appears to be calculated as the fraction of the participant's weight that is accounted for by their skeletal muscle mass, both of which were present in the initial dataset. However, the other, named "WC (waist circumference)", appears to be new, as I cannot find it anywhere in the initial dataset. This might make one wonder what other variables were collected but not reported.</li></ol><div>Apart from these changes, however, the data concerning the final conditions (49 participants, Cold and Control) are identical to the first dataset file. That is, the four duplicate participants described above are still in there; it's just harder to spot them now without the baseline record worksheet to tie the conditions together.</div><p></p><h4 style="text-align: left;">Data availability</h4><div style="text-align: left;"><br /></div><div style="text-align: left;">I have made two censored versions of the dataset available <span style="font-weight: 400;"><a href="http://nickbrown.fr/blog/cold-exposure" style="background-color: white;" target="_blank">here</a></span><span style="font-weight: 400;">. One of these ("Simply_anonymized_dataset.xlsx") has been made very quickly from the original dataset by simply deleting the participants' names, dates of birth, and (where present) e-mail addresses. The other ("Public_analysis_dataset.xls") has been cleaned up from the original in several ways, and includes the fake names and dates of birth discussed above. This file is probably easier to follow if you want to reproduce my analyses. I believe that in both cases I have taken sufficient steps to make it impractical to identify any of the participants from the remaining information.</span></div><div><span style="font-weight: 400;"><div><span>At the same location I have also placed </span>another file ("Compare_datasets.xls") in which I compare the data from the initial and new dataset files, and demonstrate that where the same fields are present, their values are identical.</div><div><br /></div><div>If anyone wants to check my work against the original, untouched dataset file, which includes the PII of the participants, then please contact me and we can discuss it. There's no obvious reason why I should be entitled to see this PII and another suitably qualified researcher should not, but of course it would not be a good idea to share it for all to see.</div><div><br /></div></span></div><h4 style="text-align: left;">Acknowledgements</h4><p>My thanks go to:</p><p></p><ul style="text-align: left;"><li>Gideon Meyerowitz-Katz (@GidMK) for interesting discussions and contributing a couple of the points in this post, including making the all-important discovery of the PII (and writing to the authors to get them to take it down).</li><li>@matejcik for looking up the frequencies of Czech names.</li><li>Stuart Ritchie for tweeting skeptically about the hyping of the results of the study.</li></ul><div><br /></div><br /><p></p><p><br /></p>
</div>Nick Brownhttp://www.blogger.com/profile/07481236547943428014noreply@blogger.com4tag:blogger.com,1999:blog-7890764972166411105.post-38260697519049723952022-09-22T22:01:00.007+02:002023-09-20T19:27:25.314+02:00Further apparent (self-)plagiarism in the work of Dr Paul McCrory<div style="line-height: 24px;">
<p>In <a href="https://steamtraen.blogspot.com/2022/03/some-examples-of-apparent-plagiarism.html" target="_blank">an earlier post</a> I reported on a number of apparently plagiarised or self-plagiarised articles by Dr Paul McCrory. Since then there have been a number of developments in this story, which has attracted attention from media worldwide, especially in Australia, led by Melissa Davey of the Guardian, who has written about the present blog post <a href="https://www.theguardian.com/sport/2022/sep/23/new-plagiarism-claims-against-sport-concussion-guru-paul-mccrory" target="_blank">here</a>.</p><p>In this post I present 10 more examples of apparent text recycling by the same author. These are admittedly quite similar in style and content to the first set, but I felt that having done the work to identify these supplementary issues it was worthwhile documenting them. I feel that they demonstrate the scale of the problem: Dr McCrory has been churning out very similar stories (mostly about concussion in various sports) for 20 years, while—as far as I have been able to establish—performing very little original empirical or other research in that time.</p><p>Note that Exhibits 8, 9, and 10 include the apparent recycling of text from work that does not have Dr McCrory listed as an author (i.e., apparent plagiarism-proper rather than self-plagiarism). I omitted the case of a chapter in a book (B) that contained many paragraphs of recycled text from a chapter by Dr McCrory in an earlier book (A), because in book B he is only listed as a contributor at the start of the book (i.e., his name does not appear directly on the chapter in question).</p><p>There are probably still a few more cases to be uncovered, but it can be laborious work, especially when the only source of a document is the Google Books preview.</p><p>Finally, as a "bonus", and to show just how big a problem plagiarism is in science and academia more generally, I've included a case where Dr McCrory's work was apparently plagiarised by other authors.</p><p><br /></p><h4>Exhibit 1</h4><p>McCrory, P. (2005). Does second impact syndrome exist? <i>Clinical Journal of Sports Medicine</i>, <i>11</i>(3), 144–149. https://doi.org/10.1097/00042752-200107000-00004 </p><p>About 60% of this article appears to have been copied, verbatim and without appropriate attribution, from:</p><p></p><ul style="text-align: left;"><li>McCrory, P. R., & Berkovic, S. F. (1998). Second impact syndrome. <i>Neurology</i>, <i>50</i>(3), 677–683. https://doi.org/10.1212/WNL.50.3.677</li></ul><p></p><p>The copied text is highlighted in blue here:</p><p></p><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/a/AVvXsEj9aSJEXrQeho23X8W3kf0PZu70gdum_bfINaz_wDGH3UKfVVwXdZtyxXABETOEKP1cS7fQR_L9maggJIHTL1XRE2iXXGsQ8-HACDRGv1T9FRZjmuzrbi9XfUUwGA7UinEKS4fWuZY4vVSreeCj1_-MBSv2sdQwarjrs1xItc3_385jw0vqxs9zapeCPg" style="margin-left: 1em; margin-right: 1em;"><img alt="" data-original-height="978" data-original-width="1040" height="602" src="https://blogger.googleusercontent.com/img/a/AVvXsEj9aSJEXrQeho23X8W3kf0PZu70gdum_bfINaz_wDGH3UKfVVwXdZtyxXABETOEKP1cS7fQR_L9maggJIHTL1XRE2iXXGsQ8-HACDRGv1T9FRZjmuzrbi9XfUUwGA7UinEKS4fWuZY4vVSreeCj1_-MBSv2sdQwarjrs1xItc3_385jw0vqxs9zapeCPg=w640-h602" width="640" /></a></div><p></p><h4>Exhibit 2</h4><p>McCrory, P., le Roux, P. D., Turner, M., Kirkeby, I. R., & Johnston, K. M. (2012). Rehabilitation of acute head and facial injuries. In R. Bahr (Ed.), <i>The IOC manual of sports injuries</i> (pp. 95–100). Wiley. </p><p>About 70% of this chapter appears to have been copied, verbatim and without appropriate attribution, from the following sources:</p><p></p><ul style="text-align: left;"><li>Yellow: McCrory, P., & Rise, I. R. (2004). Head and face. In R. Bahr & S. Maehlum (Eds.), <i>Clinical guide to sports injuries</i> (pp. 55–90). Human Kinetics. https://books.google.com/books?id=mmRnr0x0p4QC</li><li>Blue: McCrory, P. (2007). Who should retire after repeated concussions? In D. MacAuley & T. M. Best (Eds.), Evidence-based sports medicine (2nd ed., pp. 93–107). Blackwell.</li><li>Green: McCrory, P., Meeuwisse, W., Johnston, K., Dvorak, J., Aubry, M., Molloy, M., & Cantu, R. (2009). Consensus statement on concussion in sport: The 3rd international conference on concussion in sport held in Zurich, November 2008. <i>Journal of Athletic Training</i>, <i>44</i>(4), 434–448. https://doi.org/10.4085/1062-6050-44.4.434</li></ul><div><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/a/AVvXsEgnkGX8juJX6j9um03hWUdGA3dY6qDouSdpOX1S_EcQ0OOJwBhXkVqooYIwB_O-KTNxeW6FDiWW2pd5FqfR7g99G-ZqC6DQsKFLE6zkvx33KKPj40WWkoqbsIWYjPNLRY5VedSmznYt3m5IV32ZSTAGkZFDpsc6qqFKxv0elne2GEKqUoRqgq4p7s8MDQ" style="margin-left: 1em; margin-right: 1em;"><img alt="" data-original-height="852" data-original-width="968" height="563" src="https://blogger.googleusercontent.com/img/a/AVvXsEgnkGX8juJX6j9um03hWUdGA3dY6qDouSdpOX1S_EcQ0OOJwBhXkVqooYIwB_O-KTNxeW6FDiWW2pd5FqfR7g99G-ZqC6DQsKFLE6zkvx33KKPj40WWkoqbsIWYjPNLRY5VedSmznYt3m5IV32ZSTAGkZFDpsc6qqFKxv0elne2GEKqUoRqgq4p7s8MDQ=w640-h563" width="640" /></a></div></div><h4>Exhibit 3</h4><p>McCrory, P. (2007). Who should retire after repeated concussions? In D. MacAuley & T. M. Best (Eds.), <i>Evidence-based sports medicine</i> (2nd ed., pp. 93–107). Blackwell. </p><p>About 35% of this chapter (which appeared as a source text in the previous exhibit) appears to have been copied, verbatim and without appropriate attribution, from the following sources:</p><p></p><ul><li>Yellow: McCrory, P. (2002). Treatment of recurrent concussion. <i>Current Sports Medicine Reports</i>, <i>1</i>(1), 28–32. https://doi.org/10.1249/00149619-200202000-00006</li><li>Blue: McCrory, P. (2001). When to retire after concussion? <i>British Journal of Sports Medicine</i>, <i>35</i>(6), 380–382.</li><li>Pink: McCrory, P. (2002). Boxing and the brain. <i>British Journal of Sports Medicine</i>, <i>36</i>(1), 2.</li><li>Green: McCrory, P., Johnston, K., Meeuwisse, W., Aubry, M., Cantu, R., Dvorak, J., T Graf-Baumann, T., Kelly, J., Lovell, M., & Schamasch, P. (2005). Summary and agreement statement of the 2nd International Conference on Concussion in Sport, Prague 2004. <i>British Journal of Sports Medicine</i>, <i>39</i>(4), 196–204. https://doi.org/10.1136/bjsm.2005.018614</li></ul><div><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/a/AVvXsEhCKk4j2c2ZYSYJ2FF5j_PFU0YW89kiKkbBlYTeDEehWwD42QvopX7GivXOTJU-EueBc7htFwHL8DwlDwHWmN47r7_ijJkanJXNh1HVmRdXkYgQ-HHQx2IrdiwXQlKLtxheIxZO3PlDRsjEKsI8erCa2vurisbbLjuT6imccqoydH-IqZME6ehSHGAwVw" style="margin-left: 1em; margin-right: 1em;"><img alt="" data-original-height="778" data-original-width="1226" height="406" src="https://blogger.googleusercontent.com/img/a/AVvXsEhCKk4j2c2ZYSYJ2FF5j_PFU0YW89kiKkbBlYTeDEehWwD42QvopX7GivXOTJU-EueBc7htFwHL8DwlDwHWmN47r7_ijJkanJXNh1HVmRdXkYgQ-HHQx2IrdiwXQlKLtxheIxZO3PlDRsjEKsI8erCa2vurisbbLjuT6imccqoydH-IqZME6ehSHGAwVw=w640-h406" width="640" /></a></div></div><h4>Exhibit 4</h4><p>McCrory, P. (2002). Treatment of recurrent concussion. <i>Current Sports Medicine Reports</i>, <i>1</i>(1), 28–32. https://doi.org/10.1249/00149619-200202000-00006 </p><p>About 40% of this article (which appeared as a source text in the previous exhibit) appears to have been copied, verbatim and without appropriate attribution, from the following sources (which also appeared as source texts in the previous exhibit):</p><p></p><ul><li>Yellow: McCrory, P. (2001). When to retire after concussion? <i>British Journal of Sports Medicine</i>, <i>35</i>(6), 380–382.</li><li>Blue: McCrory, P. (2002). Boxing and the brain. <i>British Journal of Sports Medicine</i>, <i>36</i>(1), 2.</li><li>Pink: McCrory, P., Johnston, K. M., Mohtadi, N. G., & Meeuwisse, W. (2001). Evidence-based review of sport-related concussion: Basic science. <i>Clinical Journal of Sport Medicine</i>, <i>11</i>(3), 160–165. https://doi.org/10.1097/00042752-200107000-00006</li></ul><div><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/a/AVvXsEjRJ26id2NK3PtXckxkXt7p3o6qa-b7YKnfqx1AEePTOiJV3YJKgwRSQF9PJBEt_x2yts_Ns1hykCQZWWyUOeidvyC5rfCcHoxNDFrjDdI00FCOJ7wEK-VHW0c8ugcxCwJBFD-LmCINefQ73GQowpkca_Id4LOUH8rezAhuYTKKlDhUjOAUmO1E5P1J3g" style="margin-left: 1em; margin-right: 1em;"><img alt="" data-original-height="1002" data-original-width="796" height="640" src="https://blogger.googleusercontent.com/img/a/AVvXsEjRJ26id2NK3PtXckxkXt7p3o6qa-b7YKnfqx1AEePTOiJV3YJKgwRSQF9PJBEt_x2yts_Ns1hykCQZWWyUOeidvyC5rfCcHoxNDFrjDdI00FCOJ7wEK-VHW0c8ugcxCwJBFD-LmCINefQ73GQowpkca_Id4LOUH8rezAhuYTKKlDhUjOAUmO1E5P1J3g=w509-h640" width="509" /></a></div></div><h4>Exhibit 5</h4><p>McCrory, P. (2001). The “piriformis syndrome”—myth or reality? <i>British Journal of Sports Medicine</i>, <i>35</i>(4), 209–211. https://doi.org/10.1136/bjsm.35.4.209-a</p><p>About 90% of this editorial appears to have been copied, verbatim and without appropriate attribution, from:</p><p></p><ul><li>McCrory, P., & Bell, S. (1999). Nerve entrapment syndromes as a cause of pain in the hip, groin and buttock. <i>Sports Medicine</i>, <i>27</i>(4), 261–274. https://doi.org/10.2165/00007256-199927040-00005</li></ul><div><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/a/AVvXsEj7lH8C9VvQqisZunURD674IdyaK_b5342WSSrtmveSViFKu1vpz3izuS6DYXaKX-ZnQ2riexa3bmIBurWabLB-t7F-D7YiR4_D_wpX0Vj9RTlHe8ZYLWsk1GuATEwA8pQuqd5J7OGEoUarWFtDtP0c2hl1yks_KsPdFKH9QVst1cp-r_WYOaqV4P7I-g" style="margin-left: 1em; margin-right: 1em;"><img alt="" data-original-height="738" data-original-width="692" height="640" src="https://blogger.googleusercontent.com/img/a/AVvXsEj7lH8C9VvQqisZunURD674IdyaK_b5342WSSrtmveSViFKu1vpz3izuS6DYXaKX-ZnQ2riexa3bmIBurWabLB-t7F-D7YiR4_D_wpX0Vj9RTlHe8ZYLWsk1GuATEwA8pQuqd5J7OGEoUarWFtDtP0c2hl1yks_KsPdFKH9QVst1cp-r_WYOaqV4P7I-g=w600-h640" width="600" /></a></div></div><div><h4>Exhibit 6</h4><p>McCrory, P. (2002). What advice should we give to athletes postconcussion? <i>British Journal of Sports Medicine</i>, <i>36</i>(5), 316–318. https://doi.org/10.1136/bjsm.36.5.316</p><p>About 50% of this article appears to have been copied, verbatim and without appropriate attribution, from:</p><p></p><ul><li>McCrory, P. (1997). Were you knocked out? A team physician's approach to initial concussion management. <i>Medicine & Science in Sports & Exercise</i>, <i>29</i>(7 suppl.), S207–212. https://doi.org/10.1097/00005768-199707001-00002</li></ul><div><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/a/AVvXsEjyGM4xrdtSMe67-g6q0UHBLB6FuSfr3_VEqRcQXntZGXkFSqo_kzeyNJr4du2recF34EZJSWz5MSRJU7lWrAgRo71TsP3AqqcOYwCywe1u2fAfxoFo5qVhNIb5AmVBvURja4pjS8DmoUvp6xtgwSDRxfgjGcYB1g8NBwAjDtZYYLRtO21_AQLNEjOSIQ" style="margin-left: 1em; margin-right: 1em;"><img alt="" data-original-height="758" data-original-width="1440" height="336" src="https://blogger.googleusercontent.com/img/a/AVvXsEjyGM4xrdtSMe67-g6q0UHBLB6FuSfr3_VEqRcQXntZGXkFSqo_kzeyNJr4du2recF34EZJSWz5MSRJU7lWrAgRo71TsP3AqqcOYwCywe1u2fAfxoFo5qVhNIb5AmVBvURja4pjS8DmoUvp6xtgwSDRxfgjGcYB1g8NBwAjDtZYYLRtO21_AQLNEjOSIQ=w640-h336" width="640" /></a></div></div><div><h4>Exhibit 7</h4><p>McCrory, P. (2005). Head injuries in sport. In G. Whyte, M. Harries, & C. Williams (Eds.), <i>S</i><i>ABC of Sports and Exercise Medicine</i> (3rd ed., pp. 8–15). Blackwell.</p><p>About 30% of the main text of this chapter (which is not the same as the 2015 chapter "Head injuries in sport<b>s</b>", which was discussed in my earlier blog post) appears to have been copied, verbatim and without appropriate attribution, from the following sources:</p><p></p><ul><li>Yellow: McCrory, P. (1997). Were you knocked out? A team physician's approach to initial concussion management. <i>Medicine & Science in Sports & Exercise</i>, <i>29</i>(7 suppl.), S207–212. https://doi.org/10.1097/00005768-199707001-00002</li><li>Pink: Aubry, M., Cantu, R., Dvorak, J., T Graf-Baumann, T., Johnston, K., Kelly, J., Lovell, M., McCrory, P., Meeuwisse, W., & Schamasch, P. (2005). Summary and agreement statement of the first International Conference on Concussion in Sport, Vienna 2001. <i>British Journal of Sports Medicine</i>, <i>36</i>(1), 6–10. https://doi.org/10.1136/bjsm.36.1.6</li></ul><div><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/a/AVvXsEhnPeJV0S_dyp3xidBGREF6tlMjnsUkpOfe3QlPOzy1EyOg8LfZU5lU5Be-fl4GVN-syAVKigFHQBHml6zRO_EK0iSXbwYbdkFsN-jtAAIts64LfxuFSY5SznFOOrz4XTm2lyKzFdo3FatzQ9Il5nOr2VPL9DguL0LsHxfwAsGckZ0lVTHaeM8sCmTuWA" style="margin-left: 1em; margin-right: 1em;"><img alt="" data-original-height="1018" data-original-width="1482" height="440" src="https://blogger.googleusercontent.com/img/a/AVvXsEhnPeJV0S_dyp3xidBGREF6tlMjnsUkpOfe3QlPOzy1EyOg8LfZU5lU5Be-fl4GVN-syAVKigFHQBHml6zRO_EK0iSXbwYbdkFsN-jtAAIts64LfxuFSY5SznFOOrz4XTm2lyKzFdo3FatzQ9Il5nOr2VPL9DguL0LsHxfwAsGckZ0lVTHaeM8sCmTuWA=w640-h440" width="640" /></a></div></div><h4>Exhibit 8</h4><p>McCrory, P., Davis, G., & Makdissi, M. (2012). Second impact syndrome or cerebral swelling after sporting head injury. <i>Current Sports Medicine Reports</i>, <i>11</i>(1), 21–23. https://doi.org/10.1249/JSR.0b013e3182423bfd</p><p>About 35% of this article appears to have been copied, verbatim and without appropriate attribution, from the following sources:</p><p></p><ul><li>Yellow: McCrory, P. (2005). Does second impact syndrome exist? <i>Clinical Journal of Sports Medicine</i>, <i>11</i>(3), 144–149. https://doi.org/10.1097/00042752-200107000-00004</li><li>Blue: Randolph, C., & Kirkwood, M. W. (2009). What are the real risks of sport-related concussion, and are they modifiable? <i>Journal of the International Neuropsychological Society</i>, <i>15</i>(4), 512–520. https://doi.org/10.1017/S135561770909064X</li><li>Green: Davis, G. A. (2012). Neurological outcomes. In M. W. Kirkwood & K. O. Yeates (Eds.), <i>Mild traumatic brain injury in children and adolescents: From basic science to clinical management</i> (pp. 99–122). Guilford Press.</li></ul><div><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/a/AVvXsEgx7AIWlzhYa92N146GX9Bp0zMj55y3t0NyS0Rg-ZTqYeXbu0YmeDmwn0Gqh2qrYcOgvJnPDeDFXGe_q03Yn92bye86ZH_47wklNlPzaesMup2iVVbCirKTkk_vTNpHltR0UHOb1XANtRh4ArVCf-Hh8948oQguNChYzmkkka_kL77PjZxHGv08Ra4_EQ" style="margin-left: 1em; margin-right: 1em;"><img alt="" data-original-height="571" data-original-width="1102" height="332" src="https://blogger.googleusercontent.com/img/a/AVvXsEgx7AIWlzhYa92N146GX9Bp0zMj55y3t0NyS0Rg-ZTqYeXbu0YmeDmwn0Gqh2qrYcOgvJnPDeDFXGe_q03Yn92bye86ZH_47wklNlPzaesMup2iVVbCirKTkk_vTNpHltR0UHOb1XANtRh4ArVCf-Hh8948oQguNChYzmkkka_kL77PjZxHGv08Ra4_EQ=w640-h332" width="640" /></a></div></div><div><h4>Exhibit 9</h4><p>McCrory, P. (2018). Concussion revisited: A historical perspective. In I. Gagnon & A. Ptito (Eds.), <i>Sports concussions: </i><i>A complete guide to recovery and management</i> (pp. 9–24). CRC Press.</p><p>About 15% of this chapter appears to have been copied, verbatim and without appropriate attribution, from the following sources:</p><p></p><ul><li>Yellow: McCrory, P., Feddermann-Demont, N., Dvořák, J., Cassidy, J. D., McIntosh, A., Vos, P. E., Echemendia, R. J., Meeuwisse, W., & Tarnutzer, A. A. (2017). What is the definition of sports-related concussion: A systematic review. <i>British Journal of Sports Medicine</i>, <i>51</i>(11):877–887. https://doi.org/10.1136/bjsports-2016-097393</li><li>Blue: Zezima, K. (2014, May 29). How Teddy Roosevelt helped save football. <i>The Washington Post</i>. https://www.washingtonpost.com/news/the-fix/wp/2014/05/29/teddy-roosevelt-helped-save-football-with-a-white-house-meeting-in-1905/</li><li>Pink: Johnston, K. M., McCrory, P., Mohtadi, N. G., & Meeuwisse, W. (2001). Evidence-based review of sport-related concussion: Clinical science.<span> </span><i>Clinical Journal of Sports Medicine</i>, <i>11</i>(3), 150–159. https://doi.org/10.1097/00042752-200107000-00005</li></ul><div><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/a/AVvXsEizyMd7lPtcnSIYk8msi2RJckAYDBE18ORQy8eBzLN3TRqUJV1064PlAGMsQie3FTxvKlnbG9lbMiG_Sm1J9oQpo5Yyexv3SlA-zAg2LwkVKMHy4XwPbJYtQStQVSRVxuU6dO92nenb2kpoarvNrMXXkotPEec5xp9rZeTbXxZVHrI-7khhc4iDkrNCAg" style="margin-left: 1em; margin-right: 1em;"><img alt="" data-original-height="656" data-original-width="1336" height="314" src="https://blogger.googleusercontent.com/img/a/AVvXsEizyMd7lPtcnSIYk8msi2RJckAYDBE18ORQy8eBzLN3TRqUJV1064PlAGMsQie3FTxvKlnbG9lbMiG_Sm1J9oQpo5Yyexv3SlA-zAg2LwkVKMHy4XwPbJYtQStQVSRVxuU6dO92nenb2kpoarvNrMXXkotPEec5xp9rZeTbXxZVHrI-7khhc4iDkrNCAg=w640-h314" width="640" /></a></div></div><div><div><h4>Exhibit 10</h4><div><p>McCrory, P., Bell, S., & Bradshaw, C. (2002). Nerve entrapments of the lower leg, ankle and foot in sport. <i>Sports Medicine</i>, <i>32</i>(6), 371–391. https://doi.org/10.2165/00007256-200232060-00003</p><p>About 20% of the text of this article appears to have been copied, verbatim and without appropriate attribution, from the following sources:</p><p></p><ul><li>Yellow: McCrory, P. (2000). Exercise-related leg pain: Neurological perspective. <i>Medicine & Science in Sports & Exercise</i>, <i>32</i>(3 suppl.), S11–14. https://doi.org/10.1097/00005768-200003001-00003</li><li>Green: Pecina, M. M., Markiewitz, A. D., & Krmpotic-Nemanic, J. (2001). <i>Tunnel syndromes</i> (3rd ed.). See Chapter 44, p. 229.</li></ul><div><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/a/AVvXsEhoOdGBaTqjP8TyJ8BjNJyoQKdcfDT8tauvXqQBqF_EZBnl3y4yEvbI9PveS060YhloNi4pHSoMinjuALiRU_T_z1mbYrEkeYExZJtZ9STHZ_nrUx292kEXkKxyL2GkBpBbTUU1que8k6QAK7byFphYrH-u2_Yf9Jq44WowBZg7GqYi-93GKI8X1m1BxA" style="margin-left: 1em; margin-right: 1em;"><img alt="" data-original-height="1040" data-original-width="1477" height="450" src="https://blogger.googleusercontent.com/img/a/AVvXsEhoOdGBaTqjP8TyJ8BjNJyoQKdcfDT8tauvXqQBqF_EZBnl3y4yEvbI9PveS060YhloNi4pHSoMinjuALiRU_T_z1mbYrEkeYExZJtZ9STHZ_nrUx292kEXkKxyL2GkBpBbTUU1que8k6QAK7byFphYrH-u2_Yf9Jq44WowBZg7GqYi-93GKI8X1m1BxA=w640-h450" width="640" /></a></div></div></div></div><div><h4>Bonus Exhibit: The biter bit?</h4><div><p>Espinosa, N., Jr. & Klammer, G. (2018). Peripheral nerve entrapment around the foot and ankle. In M. D. Miller & S. R. Thompson (Eds.), <i>DeLee & Drez's orthopaedic sports medicine</i> (5th ed., pp. 1402–1420). Elsevier Health Sciences.</p><p>The highlighted sentences of this chapter appear to have been copied, verbatim and without appropriate attribution, from Exhibit 10:</p><p></p><ul><li>McCrory, P., Bell, S., & Bradshaw, C. (2002). Nerve entrapments of the lower leg, ankle and foot in sport. <i>Sports Medicine</i>, <i>32</i>(6), 371–391. https://doi.org/10.2165/00007256-200232060-00003</li></ul><div><div class="separator" style="clear: both; text-align: center;"><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/a/AVvXsEhNQg_LLtuFPo3MTGk2tZMZ7l8oa25wf8tKxd8LtwTddlBd8q9UbiupRclbp0zenMOoU57HtrLxjofbhDOhx3qpUZm-q5LEW35-BQNYvh7sOcVKOhAWgSYlteOl0rghH8YTaw1TI78oYjD5EB3uq8zhR-0iHcKBtar98Jg8urv6nik30uryXbqKphHWNw" style="margin-left: 1em; margin-right: 1em;"><img alt="" data-original-height="650" data-original-width="1486" height="280" src="https://blogger.googleusercontent.com/img/a/AVvXsEhNQg_LLtuFPo3MTGk2tZMZ7l8oa25wf8tKxd8LtwTddlBd8q9UbiupRclbp0zenMOoU57HtrLxjofbhDOhx3qpUZm-q5LEW35-BQNYvh7sOcVKOhAWgSYlteOl0rghH8YTaw1TI78oYjD5EB3uq8zhR-0iHcKBtar98Jg8urv6nik30uryXbqKphHWNw=w640-h280" width="640" /></a></div><div style="text-align: left;"><br /></div></div>(This book was first published in 2003 and as far as I can have been able to establish, the chapter by Espinosa and Klammer was first added in the 4th edition in 2014. If by some chance I have got the order of recycling wrong then I humbly apologise to Drs. Espinosa and Klammer and will issue a correction.)</div></div></div><div><br /></div><div><h4 style="text-align: left;">Data availability</h4><div>All of the supporting files for this post can be found <a href="http://nickbrown.fr/blog/mccrory2" target="_blank">here</a>. I imagine that this involves quite a few copyright violations of my own, in that many of the source documents are not open access. I hope that the publishers will forgive me for this, but if I receive a legal request to take down any specific file I will, of course, comply with that.</div></div><div><br /></div><div>(The preceding paragraph has been copied verbatim from my first blog post on the McCrory matter. Ironic, I know.)</div><div><br /></div><div><h4>Acknowledgements</h4></div><div>Big thanks to Sean Rife and James Heathers for letting me use their TAPAS tool to compare documents.</div><p><br /></p></div></div></div></div></div>Nick Brownhttp://www.blogger.com/profile/07481236547943428014noreply@blogger.com3tag:blogger.com,1999:blog-7890764972166411105.post-45387855062069913952022-08-09T22:59:00.004+02:002023-02-14T12:33:55.504+01:00An interesting lack of randomness in a published dataset: Scott and Dixson (2016)<div style="line-height: 24px;">
<p>Martin Enserink has just published <a href="https://www.science.org/content/article/star-marine-ecologist-committed-misconduct-university-says" target="_blank">the third instalment</a> in an ongoing story of strange results and possible → likely → confirmed misconduct in the field of marine biology, and more specifically the purported effects of climate change on the behaviour of fish. The first two instalments are <a href="https://www.science.org/content/article/analysis-challenges-slew-studies-claiming-ocean-acidification-alter-fish-behavior" target="_blank">here</a> (2020) and <a href="https://www.science.org/content/article/does-ocean-acidification-alter-fish-behavior-fraud-allegations-create-sea-doubt" target="_blank">here</a> (2021).</p><p>After Martin's 2021 article, I wrote <a href="https://steamtraen.blogspot.com/2021/05/todays-topic-is-report-xxlink-in.html" target="_blank">this blog post</a> describing a few analyses that I had contributed to this investigation. Today I want to look at a recently-corrected article from the same lab, mentioned by Martin in his latest piece (see the section entitled "A corrected paper"), and in particular at the data file that was released as part of the correction, as I think that it illustrates an interesting point about the forensic investigation of data.</p><p>Here is the article:</p><div style="line-height: 24px; margin-left: 36pt; text-indent: -36pt;"><span style="mso-no-proof: yes;">Scott, A., & Dixson, D. L. (2016). Reef fishes can recognize bleached habitat during settlement: Sea anemone bleaching alters anemonefish host selection. <i>Proceedings of the Royal Society B</i>, <i>283</i>, 20152694. </span>https://doi.org/10.1098/rspb.2015.2694</div><div style="line-height: 24px;"><br /></div><div style="line-height: 24px;">A correction notice was issued for this article on July 8, 2022, and that correction was accompanied by a data file, which can be downloaded from <a href="https://doi.org/10.5281/zenodo.6565204" target="_blank">here</a>. </div><div style="line-height: 24px;"><br /></div><div style="line-height: 24px;">I suggest that you read Martin's articles to get an idea of the types of experiments being conducted here, as the Scott and Dixson article is typical of many coming from the same lab. Basically, 20 (usually) fish were each tested 24 times to see if they "preferred" (i.e., chose to swim in) one of two streams of water (known as "flumes"), A and B, and then this set of 24 trials was repeated a second time, so each A/B pair of flumes was tested by 20 × 24 × 2 = 960 trials. In some cases the fish would be expected to have no preference between the flumes, and in others they should have shown a preference for water type A over B (for example, if B contained the odour of a predator, or some other chemical suggesting an unfavourable environment). The fact that in many cases the fish preferred water B, either when they were expected to have no preference or (even worse) when they were expected to prefer water A, was taken by the authors as an indication that something had gone wrong in the fish's ability to make adaptive choices in their environment.</div><p>Here are the two main issues that I see with this (claimed) dataset.</p><p><br /></p><h4 style="text-align: left;">This isn't what a dataset looks like</h4><p>As I noted in my earlier post, <i>this just isn't what a dataset looks like</i>. You don't collect data in multiple 2-D panels and lay those out in a further 2-D arrangement of 18 x 5 panels, because that's just making life difficult for yourself when you come to do the analyses. If for some reason you actually collected your data like that, you would need to write some quite sophisticated software—probably a couple of hundred lines of Python or R code—to reliably read the data in a form that is ready to generate the figures and/or tables that you would need for your article. That code would have to be able to cope with the inevitable errors that sneak into files when you are collecting data, such as occasional offsets, or a different number of fish in each chunk (the chunks on lines 78 through 94 only have 17 fish rather than 20; incidentally, the article says that each experiment was run on 18 to 20 fish), or an impossible value such as we see at cell DU46. (The code that I wrote to read the dataset from the Dixson et al. article that was the subject of my earlier blog post is around 300 lines long, including reasonable spacing.) </p><p>So there would seem to be two possibilities. Either the authors have some code that reads this file and reliably extracts the data in a form suitable for running the analyses; or, they have another data file which is more suited to reading directly into SPSS or R without having to strip away all of the formatting that makes the Excel sheet relatively visually appealing. Either way, they can surely provide one or other of those to us so that we can see how they dealt with the problems that I listed above. (I will leave it up to the reader to decide if there are any other possibilities.)</p><p><br /></p><h4 style="text-align: left;">There is too little variation... in the <i>unremarkable</i> results</h4><p>In my earlier blog post on this topic I analysed another dataset from the same lab (Dixson et al., 2014) in which there were numerous duplications, whereby the sequence of the numbers of choices of one or other flume for the 20 fish in one experiment were often very similar to those in another experiment, when there was no reason for that to be the case.</p><p>In the current dataset there are a few sets of repeated numbers of this kind (see image), but I don't think that they are necessarily a problem by themselves, for a couple of reasons.</p><p></p><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/a/AVvXsEgJIn7Q0sg7sCTBHLUM5h40JXFPyXAMUMtGZbRkRgvjAWdZFeq-9yTlGfzCrmQKyCZzwglqPx_yLV1ZFXHeXahMBI1LiebqkZN_e85QHo0bxA8r_m6-20ske1aGK2h-xsU9dRM_WWE7eU8nYsFZ_Uh4obkVrWiOrjW2QJVyogXMQdhrWZX_2CULznH84A" style="margin-left: 1em; margin-right: 1em;"><img alt="" data-original-height="897" data-original-width="838" height="640" src="https://blogger.googleusercontent.com/img/a/AVvXsEgJIn7Q0sg7sCTBHLUM5h40JXFPyXAMUMtGZbRkRgvjAWdZFeq-9yTlGfzCrmQKyCZzwglqPx_yLV1ZFXHeXahMBI1LiebqkZN_e85QHo0bxA8r_m6-20ske1aGK2h-xsU9dRM_WWE7eU8nYsFZ_Uh4obkVrWiOrjW2QJVyogXMQdhrWZX_2CULznH84A=w598-h640" width="598" /></a></div><br /><div style="text-align: center;">Were these lines (in green) copied, or are the similarities caused by the limited range of the data? My hunch is that it's the latter, but it isn't really important.</div><p></p><p><br /></p><p>First, these lines only represent sequences of a few identical numbers at a time, whereas in the 2014 dataset there were often entire duplicated groups of 20 fish.</p><p>Second, for most of these duplications, the range of the numbers is severely restricted because they (at least ostensibly) correspond to a large experimental effect. The Scott and Dixson article reports that in many cases, the fish chose flume A over flume B almost all of the time. The means that the numbers of observations of each fish in flume A, out of 24 opportunities, must almost always be 22 or 23 or 24, in order for the means to correspond to the figures in the article. There are only so many ways that such a small number of different predicted values can be distributed, and given that the person examining the dataset is free to look for matches across 180 20-fish (or 17-fish) columns of data, a number of duplicates of a certain length will very likely arise by chance.</p><p>However, the dataset also contains a number of cases where the fish appeared to have no preference between the two. The mean number of times out of 24 trials that they were recorded as having chosen flume A (or flume B) in these cases was close to 12. And it turns out that in almost all of these cases, there is a different lack of variation, not in the sequence of the observations (i.e., the numbers observed from top to bottom of the 20 fish across experiments), but in the variability of the numbers within each group of fish.</p><p>If the fish genuinely don't have a preference between the two flumes, then each observation of a fish is basically a Bernoulli trial with a probability of success equal to 0.5—which is a fancy way of saying "a coin toss"—and so the 24 trials for each fish represent 24 coin tosses. Now, when you toss a coin 24 times, the most likely result is 12 heads and 12 tails, corresponding to the fish being in flume A and B 12 times each. However, although this result is the most likely, it's not especially likely; in fact, the exact 12–12 split will only occur about 16% of the time, as you can see at <a href="https://planetcalc.com/7044/" target="_blank">this site</a> (put 24 into "Number of Bernoulli trials", click Calculate, and the probability of each result will be in the table under the figure with the curve). If you repeat those 24 trials 100 times, you would expect to get 8 As and 16 Bs (or 8 Bs and 16 As) either 4 or 5 times.</p><p>Now let's look at the dataset. I identified 32 columns of data with 20 (or, in a few cases, 17) fish and a mean of around 12. I also included 3 other columns which had one or more values of 12; as I hope will become clear, this inclusion works in the authors' favour. I then calculated the standard deviation (SD) of the 20 (or 17) scores that are composed of 24 trials for each of these 35 columns of data.</p><p>Next, I generated one million random samples of 24 trials for 20 simulated fish and calculated the SD of each sample. For each of the 35 SDs taken from the dataset, I calculated the fraction of those million simulated SDs that were smaller than the dataset value. In other words, I calculated how likely it was that one would observe an SD as small as the one that appears in the dataset if the values in the dataset were indeed taken from 24 trials of 20 fish that had no preference between the flumes. Statistically-minded readers may recognise this as the <i>p</i> value for the null hypothesis that these data arose as the result of a natural process, as described by the authors of the Scott and Dixson paper.</p><p>The results are not very good for the authors. For only nine of the samples, including the three that contain a small number of scores of 12 but otherwise have a substantially different mean, the <i>p</i> values are greater than 0.05. Seven of the <i>p</i> values are zero, meaning that an SD as low as the one corresponding to the data reported by the authors did not occur at all in one million simulated samples (see image below for an example). A further six <i>p</i> values are less than 0.0001 and four are less than 0.001. The overall chances of obtaining these results from a natural process are hard to calculate accurately (for example, one would need to make a small adjustment for the fact that the results come in pairs of 20-fish samples, as each fish took part in 2 sets of 24 trials and those two sets are not independent), but in any case I think it can safely be described as homeopathic, if only from the seven cases of zero matches out of one million. </p><div><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/a/AVvXsEiSACI3UBYnBvuTm-UfJwOMbYe0Dee5v4w5zIhvexeCURv0aWJOqARWabhNp4GWaTYCYZDHCTjDbaz2BVLqULkBZxGv-a8lJ4XOU1xWUeUt_ybr7q0utJOTq1Dzh1G5G2ipq112PNSn5c_fihC1U3UfC2-Qv1f1vW2Gb5RXhcDHe6MAZLsnxBiqLTl37A" style="margin-left: 1em; margin-right: 1em;"><img alt="" data-original-height="500" data-original-width="385" height="640" src="https://blogger.googleusercontent.com/img/a/AVvXsEiSACI3UBYnBvuTm-UfJwOMbYe0Dee5v4w5zIhvexeCURv0aWJOqARWabhNp4GWaTYCYZDHCTjDbaz2BVLqULkBZxGv-a8lJ4XOU1xWUeUt_ybr7q0utJOTq1Dzh1G5G2ipq112PNSn5c_fihC1U3UfC2-Qv1f1vW2Gb5RXhcDHe6MAZLsnxBiqLTl37A=w493-h640" width="493" /></a></div><div style="text-align: center;"><br /></div><div style="text-align: center;">Remarkably consistent results. SD in yellow (0.7863), proportion of simulated data values that have a lower SD in green (0.000000).</div></div><p><u><br /></u></p><h4>Conclusion</h4><p>Lack of expected variability is a recurring theme in the investigation of bad science. Uri Simonsohn was one of the pioneers of this in his paper "<a href="https://journals.sagepub.com/doi/abs/10.1177/0956797613480366" target="_blank">Just Post It</a>", and more recently Kyle Sheldrick came up with <a href="http://kylesheldrick.blogspot.com/2021/08/data-from-cadegiani-et-al-contains.html" target="_blank">a novel method</a> of checking whether the sequence of values in a dataset is "too regular". I hope that my explanation of the issues that I see in the Scott and Dixson dataset is clear.</p><p>Martin Enserink's latest piece mentions that the University of Delaware is seeking the retraction of three papers with Danielle Dixson as an author. Apparently the Scott and Dixson (2016) article—which, remember, has already been corrected once—is among those three papers. If nobody identifies a catastrophic error in my analyses then I plan to write to the editors of the journal to bring this issue to their attention.</p><div><br /></div><h4 style="text-align: left;">Data availability</h4><p>I have made an annotated copy of the article PDF file available <a href="http://www.meta-systems.eu/nickbrown/blog/dixson/Scott-Dixson-2016/" target="_blank">here</a>, which I think constitutes fair use. As mentioned earlier, the dataset is available <a href="https://doi.org/10.5281/zenodo.6565204" target="_blank">here</a>.</p><p><br /></p><p><br /></p>
<div style="line-height: 24px; margin-left: 36pt; text-indent: -36pt;"><br /></div></div>
Nick Brownhttp://www.blogger.com/profile/07481236547943428014noreply@blogger.com4tag:blogger.com,1999:blog-7890764972166411105.post-38699774279243115302022-03-07T13:21:00.013+01:002022-04-06T23:45:14.491+02:00Some examples of apparent plagiarism and text recycling in the work of Dr Paul McCrory<div style="line-height: 24px;">
<p><a href="https://scholar.google.com/citations?user=4kVEw14AAAAJ" target="_blank">Dr Paul McCrory</a> of the <a href="https://florey.edu.au/science-research/scientist-directory/associate-professor-paul-mccrory" target="_blank">Florey Institute of Neuroscience and Mental Health</a> has been in the news in the past few days. This started with a single retraction of <a href="https://bjsm.bmj.com/content/39/11/785" target="_blank">an apparently plagiarised editorial piece</a> in the <i>British Journal of Sports Medicine</i> from 2005, but after I started digging further and more problems came to light, he has now resigned as chair of the influential Concussion in Sport Group (CISG), as reported by <a href="https://www.theguardian.com/sport/2022/mar/05/concussion-kingpin-resigns-global-post-over-plagiarism-scandal" target="_blank">The Guardian</a> and <a href="https://theathletic.com/news/concussion-expert-dr-paul-mccrory-tenders-resignation-from-cisg-following-plagiarism-claims/k9nhZ0gbADWn/" target="_blank">The Athletic</a>, among other outlets.</p><p>Since much of this story has only been covered in a series of separate threads on Twitter up to now, I thought I would take some time to document in one place the full extent of what I have found about Dr McCrory's extensive recycling of his own and others' writing.</p><p>The first five exhibits are already in the public domain, but I will include them here for completeness. If you have been following the story on Twitter up to now, you can skip straight to Exhibit 6.</p><p><br /></p><h4 style="text-align: left;">Exhibit 1</h4><p>McCrory, P. (2005). The time lords. <i>British Journal of Sports Medicine</i>, <i>39</i>(11), 785–786.</p><p>About 50% of this article has been copied, verbatim and without appropriate attribution, from <a href="https://physicsworld.com/a/physics-technology-and-the-olympics/" target="_blank">this 2000 article in Physics Today</a> by Steve Haake, who was the person who first discovered Dr McCrory's plagiarism and brought it to the attention of the current editor-in-chief of the <i>British Journal of Sports Medicine</i>. The copied text is highlighted in pink here:</p><p></p><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/a/AVvXsEg8IAg7AxsnDFCQlXXZpwdkG8vnvHvRC72EJ9D9XI6ATv3ouHwX5-7rWmSMMb4YuWMfFr-hAerN74XBOFj3n42I90_MnQPZ4TSOV9cgQIhWewLI-fF3qmjJjEnQ-NCIYZbLA9pAkbxRG87nsueEKgAbtagOBk02Lqk_gVDqZhaJKtDP3T_uHPfkDAJKEw=s945" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="945" data-original-width="517" height="640" src="https://blogger.googleusercontent.com/img/a/AVvXsEg8IAg7AxsnDFCQlXXZpwdkG8vnvHvRC72EJ9D9XI6ATv3ouHwX5-7rWmSMMb4YuWMfFr-hAerN74XBOFj3n42I90_MnQPZ4TSOV9cgQIhWewLI-fF3qmjJjEnQ-NCIYZbLA9pAkbxRG87nsueEKgAbtagOBk02Lqk_gVDqZhaJKtDP3T_uHPfkDAJKEw=w350-h640" width="350" /></a></div><div class="separator" style="clear: both; text-align: center;"><br /></div><p></p><p>The editorial has now been <a href="https://bjsm.bmj.com/lookup/doi/10.1136/bjsports-39-11-785ret" target="_blank">retracted</a>. This was <a href="https://retractionwatch.com/2022/02/28/this-is-frankly-insulting-an-author-plagiarized-by-a-journal-editor-speaks/" target="_blank">reported by Retraction Watch on February 28, 2022</a>. At that point I started looking into other articles by the same author.</p><h4 style="text-align: left;"><br /></h4><h4 style="text-align: left;">Exhibit 2</h4><p>McCrory, P. (2005). Definitions for the purist. <i>British Journal of Sports Medicine</i>, <i>39</i>(11), 786.</p><p>About 70% of this article has been copied, verbatim and without appropriate attribution, from <a href="http://www.weizmann.ac.il/sci-tea/physics/sciam.html#q8" target="_blank">this website</a>. A copy of that page, archived on May 22, 2003 (that is, two years before Dr McCrory's article was published) can be found <a href="https://web.archive.org/web/20030522033549/http://www.weizmann.ac.il/sci-tea/physics/sciam.html#q8" target="_blank">here</a>. The copied text is highlighted in yellow here:<a href="https://blogger.googleusercontent.com/img/a/AVvXsEiXVUCImR6_SbvZKVqIJj7Klr9dx4OI5KE7eamhi4cc03aZRKexm78xbDM-ruZxA5iztuVwcDVlI8VPTqrgunK87Zdj7QFEdNLcGjbG-uX93elLudEx4qBj-FE7fWCQXU3ImdDlgFfs3z6-hGaKn2PU4je4s0K0fwUo_Htb3eMJmXums1J4WUktEtjueg=s940" style="margin-left: 1em; margin-right: 1em; text-align: center;"><img border="0" data-original-height="635" data-original-width="940" height="432" src="https://blogger.googleusercontent.com/img/a/AVvXsEiXVUCImR6_SbvZKVqIJj7Klr9dx4OI5KE7eamhi4cc03aZRKexm78xbDM-ruZxA5iztuVwcDVlI8VPTqrgunK87Zdj7QFEdNLcGjbG-uX93elLudEx4qBj-FE7fWCQXU3ImdDlgFfs3z6-hGaKn2PU4je4s0K0fwUo_Htb3eMJmXums1J4WUktEtjueg=w640-h432" width="640" /></a></p><p>I <a href="https://twitter.com/sTeamTraen/status/1498792034680909829" target="_blank">tweeted</a> about this article on March 1, 2022. Retraction Watch <a href="https://retractionwatch.com/2022/03/02/was-leading-sports-medicine-researchers-plagiarism-an-isolated-and-unfortunate-incident/" target="_blank">picked up on that</a> and later <a href="https://retractionwatch.com/2022/03/04/sports-medicine-researcher-paul-mccrory-requests-another-retraction/" target="_blank">reported that the author had asked for the article to be retracted</a>, giving an explanation that <a href="https://twitter.com/sTeamTraen/status/1499730887839297542" target="_blank">I found less than impressive</a>.</p><p><br /></p><h4>Exhibit 3</h4><p>McCrory, P. (2006). Take nothing but pictures, leave nothing but footprints…? <i>British Journal of Sports Medicine</i>, <i>40</i>(7), 565. https://doi.org/10.1136/bjsm.2006.029231</p><p>Nearly 80% of the words in this article have been copied, verbatim and without appropriate attribution, from the following sources:</p><p></p><ul style="text-align: left;"><li>Yellow: <a href="at https://www.gdrc.org/uem/footprints/what-is-ef.html" target="_blank">This website</a>. A copy of that page, archived on December 7, 2003 (that is, more than two years before Dr McCrory's article was published) can be found <a href="https://web.archive.org/web/20031207041154/https://www.gdrc.org/uem/footprints/what-is-ef.htm" target="_blank">here</a>.</li><li>Pink: <a href="https://www.newscientist.com/article/dn7274-sports-events-leave-a-giant-ecological-footprint/" target="_blank">This article</a> from <i>New Scientist</i>, dated April 16, 2005.</li><li>Blue: <a href="https://www.globalurban.org/GUDMag06Vol2Iss1/Roper.htm" target="_blank">This website</a>, dated March 2006 (several months before Dr McCrory's article was published). An archived copy from May 2, 2006 can be found <a href="https://web.archive.org/web/20060502170728/https://www.globalurban.org/GUDMag06Vol2Iss1/Roper.htm" target="_blank">here</a>.</li><li>Green: <a href="https://indianwildlifeclub.com/ezine/view/details.aspx?m=11&y=2005" target="_blank">This website</a>, dated November 2005.</li><li>Grey: <a href="at https://www.gdrc.org/uem/footprints/" target="_blank">This website</a>. An archived copy from September 6, 2003 can be found <a href="https://web.archive.org/web/20030906153759/http://www.gdrc.org/uem/footprints/" target="_blank">here</a>.</li></ul><p></p><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/a/AVvXsEhMnR0klGGvIaLYv3umjX77ffPc0A4jFh77C4N-WWd7aTw4FK_S1gR82ruJbcKrKg7i-cNnr3IpAcrw3LYgKPLTGS9yCYUcE32LjljBYJTjfNeNEz-_aF5w4V4myCPRzLogBYxbQDAEkv_4_Fo9RIPj9TfsRtZM3KXH-mcH26Lb6iGPUz-WNgIJ0iUUJw=s1017" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="1017" data-original-width="739" height="640" src="https://blogger.googleusercontent.com/img/a/AVvXsEhMnR0klGGvIaLYv3umjX77ffPc0A4jFh77C4N-WWd7aTw4FK_S1gR82ruJbcKrKg7i-cNnr3IpAcrw3LYgKPLTGS9yCYUcE32LjljBYJTjfNeNEz-_aF5w4V4myCPRzLogBYxbQDAEkv_4_Fo9RIPj9TfsRtZM3KXH-mcH26Lb6iGPUz-WNgIJ0iUUJw=w466-h640" width="466" /></a></div><div style="line-height: 24px;"><br /></div>As with Exhibit 2, I <a href="https://twitter.com/sTeamTraen/status/1498792034680909829" target="_blank">tweeted</a> about this on March 1, 2022. The author came up with <a href="https://retractionwatch.com/2022/03/04/sports-medicine-researcher-paul-mccrory-requests-another-retraction/" target="_blank">a quite remarkable story for Retraction Watch</a> about why this article only merited a correction. I <a href="https://twitter.com/sTeamTraen/status/1499741979072208897" target="_blank">found that even less impressive</a> than his excuses in the previous case.</div><div style="line-height: 24px;"><p><br /></p><h4>Exhibit 4</h4><p>McCrory, P. (2002). Commotio cordis. <i>British Journal of Sports Medicine</i>, <i>36</i>(4), 236–237.</p><p>About 90% of the words in this article have been copied, verbatim and without appropriate attribution, from the following sources:</p><p></p><p></p><p></p><p></p><ul style="text-align: left;"><li>Yellow:<span> </span>Curfman, G. D. (1998). Fatal impact — Concussion of the heart. <i>New England Journal of Medicine</i>, <i>338</i>(25), 1841-1843. https://doi.org/10.1056/NEJM199806183382511</li><li>Blue: Nesbitt, A. D., Cooper, P. J., & Kohl, P. (2001). Rediscovering commotio cordis. <i>The Lancet</i>, <i>357</i>(9263):1195–1197. https://doi.org/10.1016/S0140-6736(00)04338-5<br />.</li></ul><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/a/AVvXsEjArr_2WIVr4lyROD_b-3gYSuKDbNIj83OVmGKGillhDJo29y6PdQDR3RFqGbP95u_h5GhMCy12rYbou8c_C1Dsxw_XqjcX8pbIzOGFBfvX2jKVhvw-yTIEsx0fC38B1c-b_ry6461qGx9p3eb0nfpEb4YJfiuyMI_hY7TWNu29DrAewigGWfBpRluX0Q=s996" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="996" data-original-width="690" height="640" src="https://blogger.googleusercontent.com/img/a/AVvXsEjArr_2WIVr4lyROD_b-3gYSuKDbNIj83OVmGKGillhDJo29y6PdQDR3RFqGbP95u_h5GhMCy12rYbou8c_C1Dsxw_XqjcX8pbIzOGFBfvX2jKVhvw-yTIEsx0fC38B1c-b_ry6461qGx9p3eb0nfpEb4YJfiuyMI_hY7TWNu29DrAewigGWfBpRluX0Q=w444-h640" width="444" /></a></div><div style="line-height: 24px;"><br /></div>James Heathers <a href="https://twitter.com/jamesheathers/status/1499249931047129088" target="_blank">discovered a couple of these overlaps</a> on March 3, 2022 and I <a href="https://twitter.com/sTeamTraen/status/1499747065307582470" target="_blank">tweeted</a> the full picture on March 4, 2022.</div><div style="line-height: 24px;"><br /><h4>Exhibit 5</h4><p>McCrory, P. (2005). A cause for concern? <i>British Journal of Sports Medicine</i>, <i>39</i>(5), 249.</p><p>Almost half of the words in this article have been copied, verbatim and without appropriate attribution, from the following source:</p><p></p><ul style="text-align: left;"><li>Piazza, O., Anna-Leena Sirén, A.-L., & Ehrenreich, H. (2004). Soccer, neurotrauma and amyotrophic lateral sclerosis: Is there a connection? <i>Current Medical Research and Opinion</i>, <i>20</i>(4), 505–508. https://doi.org/10.1185/030079904125003296</li></ul><p></p><p> The copied text is highlighted in pink here:</p><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/a/AVvXsEj3VIHxLlBrs4zPadNks0Jcv-vqXkyIlcwnog7bEW4ypnUO5LPytTkp1KJLL7Ha3M-Hs_EzyIjhoPgiLhTs_WmQzQoRx8Tm3qZ1I8zVGhB4-clDEGuyU3VtX0M_oUHy800wRXkrvlOz1lK1NxsZjQE6PAyCP7ybYFFan7enTyjvvLmXUv1iWeoi-aqJuw=s982" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="982" data-original-width="708" height="640" src="https://blogger.googleusercontent.com/img/a/AVvXsEj3VIHxLlBrs4zPadNks0Jcv-vqXkyIlcwnog7bEW4ypnUO5LPytTkp1KJLL7Ha3M-Hs_EzyIjhoPgiLhTs_WmQzQoRx8Tm3qZ1I8zVGhB4-clDEGuyU3VtX0M_oUHy800wRXkrvlOz1lK1NxsZjQE6PAyCP7ybYFFan7enTyjvvLmXUv1iWeoi-aqJuw=w462-h640" width="462" /></a></div>I <a href="https://twitter.com/sTeamTraen/status/1499749087108861964" target="_blank">tweeted</a> about this on March 4, 2022.</div><div style="line-height: 24px;"><br /><h4>Exhibit 6</h4><p>McCrory, P. (2002). Should we treat concussion pharmacologically? <i>British Journal of Sports Medicine</i>, <i>36</i>(1), 3–5.</p><p>Almost 100% of the text has been copied, verbatim and without appropriate attribution, from:</p><p></p><ul style="text-align: left;"><li>McCrory, P. (2001). New treatments for concussion: The next millennium beckons. <i>Clinical Journal of Sport Medicine</i>, <i>11</i>(3), 190–193.</li></ul><p></p><p>That copied text is highlighted blue (light or dark) in the image below. The text in dark blue also overlaps with <a href="https://www.medlink.com/articles/neurotrophic-factors-for-treatment-of-neurotrauma" target="_blank">this MedLink article</a>. Thus, either Dr McCrory plagiarised three paragraphs from MedLink in two separate articles, or MedLink plagiarised him. The MedLink article was initially published in 1997, but it has been updated since, so the direction of copying cannot be established with certainty unless I can find an archived copy from 2001. It may, however, be interesting that the "phase II safety and efficacity trial" mentioned (Dr McCrory's reference 22) has a date of 1997.</p><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/a/AVvXsEiNq2FAoK9tLednxxr2Uu__Ey_20VhCaF6MGDT0-RVyUKUmjgK4VWJJLRDFxVdJ1haz_if4JBSvWvts6Ftnw3nnMPHuLC4cqPBA8C82FcwJ046X8aVng51F61nYNbAS4_ZTEejBoja_2YZWLPSGT5tCGNRwTDBwHTGqcl7eCCrMi6ePeT2f5XpGdg6IyA=s1079" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="926" data-original-width="1079" height="550" src="https://blogger.googleusercontent.com/img/a/AVvXsEiNq2FAoK9tLednxxr2Uu__Ey_20VhCaF6MGDT0-RVyUKUmjgK4VWJJLRDFxVdJ1haz_if4JBSvWvts6Ftnw3nnMPHuLC4cqPBA8C82FcwJ046X8aVng51F61nYNbAS4_ZTEejBoja_2YZWLPSGT5tCGNRwTDBwHTGqcl7eCCrMi6ePeT2f5XpGdg6IyA=w640-h550" width="640" /></a></div><br /><p>James Heathers <a href="https://twitter.com/jamesheathers/status/1499249926496264192" target="_blank">discovered one of the overlaps in this text</a> on March 3, 2022, but it took another couple of hours work at my end to uncover the full extent of the text recycling and possible plagiarism in this article.</p><div><br /></div><div><h4>Exhibit 7</h4><p>McCrory, P. (2006). How should we teach sports medicine? <i>British Journal of Sports Medicine</i>, <i>40</i>(5), 377.</p></div><div><p>About 60% of the words in this article have been copied, verbatim and without appropriate attribution, from the following sources:</p><p></p><p></p><p></p><p></p><ul><li>Pink: Fallon, K. E., & Trevitt, A. C. (2006). Optimising a curriculum for clinical haematology and biochemistry in sports medicine: A Delphi approach. <i>British Journal of Sports Medicine</i>, <i>40</i>(2), 139–144. https://doi.org/10.1136/bjsm.2005.020602</li><li>Blue: Long, G., & Gibbon, W. W. (2000). Postgraduate medical education: Methodology. <i>British Journal of Sports Medicine</i>, <i>34</i>(4), 235–245.</li></ul></div><div>Note that the Fallon & Trevitt article was published in the same journal just three months before it was plagiarised.</div><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/a/AVvXsEi96g2zFO78dEweMquY4PBfa38kK-YZtrP7zDfWyhxzuNiW8nYnmVQCwJK_NMYOY9w3mj_We5QJ4Zu_4Vq_XoVeas1EzlNnUqhxvq4Zodp2awTIVLFKpbTKEtbAb_bWKMn_uwwQp3roVUNDjuYWVkii8Y-m-y5PtOaFu_ISNVoQP_2vKNb4tByRrDfyTQ=s813" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="813" data-original-width="683" height="640" src="https://blogger.googleusercontent.com/img/a/AVvXsEi96g2zFO78dEweMquY4PBfa38kK-YZtrP7zDfWyhxzuNiW8nYnmVQCwJK_NMYOY9w3mj_We5QJ4Zu_4Vq_XoVeas1EzlNnUqhxvq4Zodp2awTIVLFKpbTKEtbAb_bWKMn_uwwQp3roVUNDjuYWVkii8Y-m-y5PtOaFu_ISNVoQP_2vKNb4tByRrDfyTQ=w538-h640" width="538" /></a></div><div><h4><br /></h4><h4>Exhibit 8</h4><p>McCrory, P. (2008). Neurologic problems in sport. In M. Schwellnus (Ed.), <i>Olympic textbook of medicine in sport</i> (pp. 412–428). Wiley.</p></div><div></div><div>About 25% of the words in this book chapter have been copied, verbatim and without appropriate attribution, from other sources. Of that 25%, about two-thirds is recycled from other publications by the same author, and the remainder is plagiarised from other authors, as follows:</div><div><div><ul style="text-align: left;"><li>Orange: McCrory, P. (2000). Headaches and exercise. <i>Sports Medicine</i>, <i>30</i>(3), 221–229. https://doi.org/10.2165/00007256-200030030-00006</li><li>Green: McCrory, P. (2001). Headache in sport. <i>British Journal of Sports Medicine</i>, <i>35</i>(5), 286–287.</li><li>Blue: McCrory, P. (2005). A cause for concern? <i>British Journal of Sports Medicine</i>, <i>39</i>(5), 249. (See also Exhibit 5.)</li><li>Yellow: Showalter, W., Esekogwu, V., Newton, K. I., & Henderson, S. O. (1997). Vertebral artery dissection. <i>Academic Emergency Medicine</i>, <i>4</i>(10), 991–995. https://doi.org/10.1111/j.1553-2712.1997.tb03666.x</li><li>Pink: <a href="https://www.medlink.com/articles/stroke-associated-with-drug-abuse" target="_blank">This MedLink article</a>, which was initially published in 1996, but has been updated since, so the direction of copying cannot be established with certainty unless I can find an archived copy from 2008. It may, however, be interesting that the citations in the pink text (Kaku & Lowenstein 1990; Brust & Richter 1977) both (a) predate the MedLink article and (b) are not — or no longer — referenced at the equivalent points in the MedLink text. It would seem unlikely that MedLink would (a) plagiarise Dr McCrory's article from 2008 at some point after that date and (b) remove these rather old citations (without replacing them with new ones).</li></ul></div></div><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/a/AVvXsEj5G7SNmgc26YoVPxTEDa3LYdepqsHF0MTdShPCEb22HPRKn3wGaJrJo9MCvv408gBDqRRuZuZEvd32wpc1q2L936Q9gvoH6eaKgPRBq1Biscdu1JEmX4wXMsANdDrSSe9SSKoJFox6UsOzwgmHKCivJFevxh8hljbSk2n1KVqvqj7KSm4UYND_yLb2pA=s1448" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="1042" data-original-width="1448" height="460" src="https://blogger.googleusercontent.com/img/a/AVvXsEj5G7SNmgc26YoVPxTEDa3LYdepqsHF0MTdShPCEb22HPRKn3wGaJrJo9MCvv408gBDqRRuZuZEvd32wpc1q2L936Q9gvoH6eaKgPRBq1Biscdu1JEmX4wXMsANdDrSSe9SSKoJFox6UsOzwgmHKCivJFevxh8hljbSk2n1KVqvqj7KSm4UYND_yLb2pA=w640-h460" width="640" /></a></div><div>(Don't bother squinting too hard at the page - the annotated PDF is available for you to inspect. See link at the end of this post.)</div><div><br /></div><div><h4>Exhibit 9</h4><p>McCrory, P., & Turner, M. (2015). Concussion – Onfield and sideline evaluation. In D. McDonagh & D. Zideman (Eds.), <i>The IOC manual of emergency sports medicine</i> (pp. 93–105). Wiley.</p><p>About 50% of the words in this book chapter have been copied, verbatim and without appropriate attribution, from other sources, as follows:</p><p></p><ul style="text-align: left;"><li>Blue: McCrory, P., le Roux, P. D., Turner, M., Kirkeby, I. R., & Johnston, K. M. (2012). Head injuries. In R. Bahr (Ed.), <i>The IOC manual of sports injuries</i> (pp. 58–94). Wiley.</li><li>Yellow: McCrory, P., le Roux, P. D., Turner, M., Kirkeby, I. R., & Johnston, K. M. (2012). Rehabilitation of acute head and facial injuries. In R. Bahr (Ed.), <i>The IOC manual of sports injuries</i> (pp. 95–100). Wiley.</li><li>Green: Aubry, M., Cantu, R., Dvorak, J., Graf-Baumann, T., Johnston, K., Kelly, J., Lovell, M., McCrory, P., Meeuwisse, W., & Schamasch, P. (2001). Summary and agreement statement of the first International Conference on Concussion in Sport, Vienna 2001. <i>British Journal of Sports Medicine</i>, <i>36</i>(1), 6–10. https://doi.org/10.1136/bjsm.36.1.6</li><li>Pink: McCrory, P. (2015). Head injuries in sports. In M. N. Doral & J. Karlsson (Eds.), <i>Sports injuries</i> (pp. 2935–2951). Springer.</li></ul><p></p><p>The pink text also appears in Exhibit 9, which was published in the same year, so it's not clear which is the original and which is the copy. I tweeted about some of the similarities between Exhibits 9 and 10 <a href="https://twitter.com/sTeamTraen/status/1498798541103370240" target="_blank">here</a>, although I hadn't found everything at that point.</p><p>The green text in the final paragraph on page 105 appears to have been copied and pasted twice (it appears in two paragraphs on page 104), which might cause the reader to wonder exactly how much care and attention went into this copy-and-paste job.</p><p>Readers who are interested in the activities of the CISG might be interested to note that the 2001 Vienna conference (the "green" text reference above) was where the name of this group was first coined.</p><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/a/AVvXsEh4aahQURU1RwFT2RTJDu8DbwZqUoMXhmpZqW9HEMLx0YTGn0eWRJe9VwwRyWGkbKpWscvDaAFD3BpJQEOyo5MKk41T0fZt5k-D3UZTQ6qMLOpJEg9VafTpCe0OGZWiLHlUYMbuu6Vzm8bhE92LuioCZ5hihy5CC_prv9qyuSlEIMmTA-hDf5sW-FZcew=s1466" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="1032" data-original-width="1466" height="450" src="https://blogger.googleusercontent.com/img/a/AVvXsEh4aahQURU1RwFT2RTJDu8DbwZqUoMXhmpZqW9HEMLx0YTGn0eWRJe9VwwRyWGkbKpWscvDaAFD3BpJQEOyo5MKk41T0fZt5k-D3UZTQ6qMLOpJEg9VafTpCe0OGZWiLHlUYMbuu6Vzm8bhE92LuioCZ5hihy5CC_prv9qyuSlEIMmTA-hDf5sW-FZcew=w640-h450" width="640" /></a></div>(Note that five pages, corresponding to the photographic reproduction of the Sport Concussion Assessment Tool and the Pocket Concussion Recognition Tool, have been omitted from this image.)</div><div><br /><h4>Exhibit 10</h4><p>McCrory, P. (2015). Head injuries in sports. In M. N. Doral & J. Karlsson (Eds.), <i>Sports injuries</i> (pp. 2935–2951). Springer.</p><p>About 90% of the words in this book chapter have been copied, verbatim and without appropriate attribution, from other sources, as follows:</p></div><div><div><ul><li>Blue (light and dark): McCrory, P. le Roux, P. D., Turner, M., Kirkeby, I. R., & Johnston, K. M. (2012). Head injuries. In R. Bahr (Ed.), <i>The IOC manual of sports injuries</i> (pp. 58–94). Wiley.</li><li>Yellow: McCrory, P. le Roux, P. D., Turner, M., Kirkeby, I. R., & Johnston, K. M. (2012). Rehabilitation of acute head and facial injuries. In R. Bahr (Ed.), <i>The IOC manual of sports injuries</i> (pp. 95–100). Wiley.</li><li>Pink: McCrory, P., & Turner, M. (2015). Concussion – Onfield and sideline evaluation. In D. McDonagh & D. Zideman (Eds.), <i>The IOC manual of emergency sports medicine</i> (pp. 93–105). Wiley.</li></ul></div><div><p>The pink text also appears in Exhibit 9, which was published in the same year, so it's not clear which is the original and which is the copy.</p><p>The text in dark blue has been copied twice from the same source; again, it seems as if this chapter was not assembled with any great amount of care.</p></div><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/a/AVvXsEgzMvuvKAQ1-rvqnWo7V8ADFHvLrAti13ViyGwjxLLguIqtqvO7mianRPQJ3OWL-ZZOwCR2hh4pQfsmY0Fo9WhjvpsD2FfrJ6hGRFe0OsHQqEklaAg50D_5tw0zl8AnsrNM3eIifb_S2bJwPFeUtU5ZmVARsZYbszbBbZxIO7pR4NCZpTOPYHjwDaZyZA=s1398" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="1020" data-original-width="1398" height="466" src="https://blogger.googleusercontent.com/img/a/AVvXsEgzMvuvKAQ1-rvqnWo7V8ADFHvLrAti13ViyGwjxLLguIqtqvO7mianRPQJ3OWL-ZZOwCR2hh4pQfsmY0Fo9WhjvpsD2FfrJ6hGRFe0OsHQqEklaAg50D_5tw0zl8AnsrNM3eIifb_S2bJwPFeUtU5ZmVARsZYbszbBbZxIO7pR4NCZpTOPYHjwDaZyZA=w640-h466" width="640" /></a></div>(Note that six pages, corresponding to the photographic reproduction of the Sport Concussion Assessment Tool and the Pocket Concussion Recognition Tool, have been omitted from this image.)<br /><div><br /></div><div><br /></div><h4 style="text-align: left;">Conclusion</h4></div><p>The exhibits above present evidence of extensive plagiarism and self-plagiarism in seven editorial pieces in the <i>British Journal of Sports Medicine</i> from 2002 through 2006, and three book chapters from 2008 through 2015. As well as the violations of publication ethics and other elementary academic norms, most of these cases would also seem to raise questions about copyright violations.</p><p>This is not an exhaustive collection; I have evidence of these transgressions on a smaller scale in a number of other articles and book chapters from the same author, but a combination of time, weariness (of me as investigator and, presumably, of the reader too), and lack of access to source materials (for example, I was only able to find one extensively recycled book chapter on Google Books, which is not very practical for marking up) has led me to stop at 10 exhibits here.</p><p>I have no background or experience in the field of head trauma or sports medicine, and I had never heard of Dr McCrory or the CISG until last week. Hence, I am unable to comment about what all of this might mean for the CISG or its influence on the rules and practices of sport. However, although I try not to editorialise too much in this blog, I must say that, based on what I have found here, Dr McCrory does not strike me as an especially outstanding example of scientific integrity, and it does make me wonder what other aspects of his life as a scientist and influencer of public policy might not stand up to close scrutiny.</p><p><br /></p><h4 style="text-align: left;">Data availability</h4><p>All of the supporting files for this post can be found <a href="http://nickbrown.fr/blog/mccrory" target="_blank">here</a>. I imagine that this involves quite a few copyright violations of my own, in that many of the source documents are not open access. I hope that the publishers will forgive me for this, but if I receive a legal request to take down any specific file I will, of course, comply with that.</p><p><br /></p></div>Nick Brownhttp://www.blogger.com/profile/07481236547943428014noreply@blogger.com5tag:blogger.com,1999:blog-7890764972166411105.post-71956128615005849592021-10-31T15:54:00.009+01:002021-11-01T01:36:20.655+01:00 A bug and a dilemma<div style="line-height: 24px;">
<p>A few months ago, I discovered that the SAS statistical software package, which is used worldwide by universities and other large organisations to analyse their data, contained—until quite recently—a bug that could result in information that the user thought they had successfully deleted (and was no longer visible from within the application itself) still being present in the saved data file. This could lead to <a href="https://www.dol.gov/general/ppii" target="_blank">personal identifiable information (PII)</a> about study participants being revealed, alongside whatever other data might have been collected from these participants, which—depending on the study—could potentially be extremely sensitive. I found this entirely by chance when looking at an SAS data file to try and work out why some numbers weren't coming out as expected, for which it would have been useful to know if numbers are stored in ASCII or binary. (It turned out that they are stored in binary.)</p><p>Here's how this bug works: Suppose that as a researcher you have run a study on 80 named participants, and you now have a dataset containing their names, study ID numbers (for example, if the study code within your organisation is XYZ this code might be XYZ100, XYZ101, etc, up to XYZ179), and other relevant variables from the study. One day you decide to make a version of the dataset that can be shared without the participants being identifiable, either because you have to deposit this in an archive when you submit the study to a journal, or because somebody has read the article and asked for your data. You could share this in <span style="font-family: courier;">.CSV</span> file format, and indeed that would normally be considered best practice for interoperability; but there may be good reasons to share it in SAS's native binary data file format with a <span style="font-family: courier;">.sas7bdat</span> extension, which can in any case be opened in R (using a package named "<a href="https://www.rdocumentation.org/packages/sas7bdat/versions/0.5/topics/read.sas7bdat" target="_blank">sas7bdat</a>", among others) or in <a href="https://stats.idre.ucla.edu/other/mult-pkg/faq/how-do-i-use-a-sas-data-file-in-spss/" target="_blank">SPSS</a>. </p><p>So you open your file called <span style="font-family: courier;">participants-final.sas7bdat</span> in the SAS data editor and delete the column with the participants' names (and any other PII, such as IP addresses, or perhaps dates of birth if those are not needed to establish the participants' ages, etc), then save it as <span style="font-family: courier;">deidentified-participants-final.sas7bdat</span>, and share the latter file. But what you don't know is that, because of this bug, in some unknown percentage of cases the text of most of the names can sometimes still be sitting in the <span style="font-family: courier;">sas7bdat</span> binary data file, close to the alphanumeric participant IDs. That is, if the bug has struck, someone who opens the "deidentified" file in a plain text editor (which could be as simple as Notepad on Windows) might see the names and IDs among the binary gloop, as shown in this image.</p><div class="separator" style="clear: both; text-align: center;"><div class="separator" style="clear: both;"><div class="separator" style="clear: both; text-align: center;"><div class="separator" style="clear: both; text-align: center;"><div class="separator" style="clear: both; text-align: center;"><a href="https://lh3.googleusercontent.com/--Twnct13bQU/YX6kRpCeU6I/AAAAAAAAGY4/VHYD63BPE6U8HJKnLrS7DlOjWmKGCk5WACLcBGAsYHQ/image.png" style="margin-left: 1em; margin-right: 1em;"><img alt="" data-original-height="255" data-original-width="621" src="https://lh3.googleusercontent.com/--Twnct13bQU/YX6kRpCeU6I/AAAAAAAAGY4/VHYD63BPE6U8HJKnLrS7DlOjWmKGCk5WACLcBGAsYHQ/s16000/image.png" /></a></div></div></div></div><div style="text-align: center;"><span style="text-align: left;">I am pretty sure these two people did not take part in this study.</span></div><div style="text-align: left;"><br /></div><div style="text-align: left;">This screenshot shows an actual extract from a data file that I found, with only the names and the study ID codes replaced with those of others selected from the phone book. The full names of about two-thirds of the participants in this study were readable. Of course, you can't read the binary data and it would take a lot of work to do so, but given the participant IDs (PRZ045 for Trump, PRZ046 for Biden) you can simply open the "anonymised" data file in SAS and find out all you want about those two people from within the application.</div><div style="text-align: left;"><br /></div><div style="text-align: left;">Even worse, though, is the fact that unless the participant's name is extremely common, when combined with knowledge of approximately where and when the study was conducted it might very well let someone identify them with a high degree of confidence for relatively little effort. And by opening the file in SAS—for example, with the free service <a href="https://www.sas.com/en_us/software/on-demand-for-academics.html" target="_blank">SAS OnDemand for Academics</a>, or in SPSS or R as previously mentioned—and looking at the data that was intended to be shared, we will be able to see that our newly-identified participant is 1.73 metres tall, or takes warfarin, or is HIV-positive.</div><div style="text-align: left;"><br /></div><div style="text-align: left;">(A number of Microsoft products, including Word and Excel, <a href="https://archive.nytimes.com/query.nytimes.com/gst/fullpage-990CE6D8163DF933A25753C1A963958260.html" target="_blank">used to have a bug like this</a>, many versions ago. When you chose "Save" rather than "Save As", it typically <a href="https://catless.ncl.ac.uk/Risks/17.78.html#subj8" target="_blank">would not physically overwrite on the disk any text that you had deleted</a>, perhaps because the code had originally been written to minimise writing operations with diskettes, which are slow.)</div></div><p>I have been told by SAS support (see screenshot below) that this bug was fixed in version 9.4M4 of the software, which was <a href="https://blogs.sas.com/content/iml/2013/08/02/how-old-is-your-version-of-sas-release-dates-for-sas-software.html" target="_blank">released</a> on 16 November 2016. The support agent told me that the problem was known to be present in version 9.4M3, which was released on 14 July 2015; however, I do not know whether the problem also existed in previous versions. I think it would be prudent to assume that any file in <span style="font-family: courier;">.sas7bdat</span> format created by a version of SAS prior to 9.4M4 may have this issue. Neither the existence of the problem, nor the fact that it had been fixed, were documented by SAS in <a href="https://support.sas.com/techsup/reports/maintSAS94/SAS94_TS1M4_issues_addressed.html" target="_blank">the release notes for version 9.4M4</a>; equally, however, the support representative did not tell me that the problem is regarded as top secret or subject to any sort of embargo.</p><p></p><div class="separator" style="clear: both; text-align: center;"><a href="https://lh3.googleusercontent.com/-2YDOAy0wKIQ/YX2zB_UUTpI/AAAAAAAAGYU/Cq0CVPPIOaUJ0fIPphBU1jXUK3M-WDLXQCLcBGAsYHQ/image.png" style="margin-left: 1em; margin-right: 1em;"><img alt="" data-original-height="381" data-original-width="1019" height="240" src="https://lh3.googleusercontent.com/-2YDOAy0wKIQ/YX2zB_UUTpI/AAAAAAAAGYU/Cq0CVPPIOaUJ0fIPphBU1jXUK3M-WDLXQCLcBGAsYHQ/w640-h240/image.png" width="640" /></a></div><div style="text-align: center;">(The identity of the organisation that shared the files in which I found the bug has been redacted here.)</div><br /><p></p><p>SAS is a complex software package and it will generally take a while for large organisations to migrate to a new version. Probably by now most versions have been upgraded to 9.4M4 or later, but quite a few sites might have been using the previous version containing this bug until quite recently, and as I already mentioned, it's not clear how old the bug is (i.e., at what point it was introduced to the software). So it could have been around for many years prior to being discovered, and it could well have still been around for two or three years after that date at many sites.</p><p>Now, this discovery caused me a dilemma. I worried that, if I were to go public with this bug, this might start a race between people who have already shared their datasets that were made with a version prior to 9.4M4 trying to replace or recall their files, and Bad People™ trying to find material online to exploit. That is, to reveal the existence of the problem might increase the risk of data leaking out. On the other hand, it's also possible that the bad people are <i>already</i> aware of the problem and are actively looking for that material, in which case every day that passes without the problem becoming public knowledge increases the risk, and going public would be the start of the solution.</p><p>Note that this is different from the typical "white hat"/"bug bounty" scenario, in which the Good People™ who find a vulnerability tell the software company about the bug and get paid to remain silent until a reasonable amount of time has passed to patch the systems, after which they are free to reveal the existence of the problem. In those cases, patching the software fixes the problem immediately, because the extent of the vulnerability is limited to the software itself. But here, the vulnerability is <i>in the data files</i> that were not anonymised as intended. There is no way to patch anything to stop those files from being read, because that only needs a text editor. The only remedy is for the files to be deleted from, or replaced in, repositories as their authors or guardians become aware of the issue.</p><p>In the original case where I discovered this issue, I reported it to the owner of the dataset and he arranged for the offending file to be recalled from the repository where he had placed it, namely the Open Science Framework. (I also gave a heads-up to the Executive Director of the Center for Open Science, Brian Nosek, at that time.) The dataset owner also reported the problem to their management, as they thought (and I completely agree) that dealing with this sort of issue is beyond the pay grade of any individual principal investigator. I do not know what has happened since, nor do I think it's really my business. I would argu<span style="background-color: white;">e that SAS ought to have done something more about this than just sneaking out a fix without telling anybody; but perhaps they, too, looked at the trade-off described above and decided to keep quiet on that basis, rather than merely avoiding embarrassment.</span></p><p>I have spent several months wondering what to do about this knowledge. In the end, I decided that (a) there probably aren't too many corrupt files out there, and (b) there probably aren't too many Bad People™ who are likely to go hunting for sensitive data this way, because it just doesn't seem like a very productive way of being a Bad Person. So I am going public today, in the hope that the practical consequences of revealing the existence of this problem are unlikely to be major, and that giving people the chance to correct any SAS data files that they might have made public will be, on balance, a net win for the Good People. (For what it's worth, I asked two professors of ethics about this, one of them a specialist in data-related issues, and they both said "Ouch. Tough call. I don't know. Do what you think is best".)</p><p>Now, what does this discovery mean? Well, if you use SAS and have made your data available using the <span style="font-family: courier;">.sas7bdat</span> file format, you might want to have a look in the data files with a text editor and check that there is nothing in there that you wouldn't expect. But even if you don't use SAS, there may still be a couple of lessons for you from this incident, because (a) the fact that this particular software bug is fixed doesn't mean there aren't others, and (b) everyone makes mistakes.</p><p>First, consider always using <span style="font-family: courier;">.CSV</span> files to share your data, if there is no compelling reason not to do so. The other day I had to download a two-year-old <span style="font-family: courier;">.RData</span> file from OSF and it contained data structures that were already partly obsolete when read by newer versions of the package that had create them; I had to hunt around online for the solution, and that might not work at all at some future point. When I had sorted that out I saved the resulting data in a <span style="font-family: courier;">.CSV</span> file, which turned out to be nearly 20% smaller than the <span style="font-family: courier;">.RData</span> file anyway.</p><p>Second, try to keep all PII out of the dataset altogether. Build a separate file or files that connects each participant's study ID number to their name and any other information that is not going to be an analysed variable. If your study requires you to generate a personalised report for the participants that includes their name then this might represent a little extra effort, but generally this approach will greatly reduce the chances of a leak of PII. (I suspect that for every participant whose PII is revealed by bugs, several more are the victims of either data theft or simply failure on the part of the researchers to delete the PII before sharing their data.)</p><p>(Thanks to Marcus Munafò and Brian Nosek for valuable discussions about an earlier draft of this post.)</p><p><br /></p>
</div>Nick Brownhttp://www.blogger.com/profile/07481236547943428014noreply@blogger.com7tag:blogger.com,1999:blog-7890764972166411105.post-82789936757484016222021-10-28T13:38:00.016+02:002022-12-24T01:25:18.899+01:00A catastrophic failure of peer review in obstetrics and gynaecology<div style="line-height: 24px;">
<p><span style="background-color: white; color: #222222; font-family: inherit;">In this post I will discuss a set of 46 articles from the same institution that appear to show severe problems in many journals in the field of </span>obstetrics and gynaecology. These are not entirely new discoveries; worrying overlaps among 35 of these articles have already been investigated in <a href="https://www.ejog.org/article/S0301-2115(20)30184-6/fulltext" style="font-family: inherit;" target="_blank">a commentary article</a><span style="background-color: white; color: #222222; font-family: inherit;"> from 2020 by Esmée Bordewijk and colleagues that critiqued 24 articles on which Dr Ahmed Badawy was lead author (19) or a co-author (5), plus 11 articles lead-authored by Dr Hatem Abu Hashim, who is Dr Badawy's colleague in the Department of Obstetrics and Gynaecology at Mansoura University in Egypt.</span></p><p><span style="background-color: white; color: #222222; font-family: inherit;">Bordewijk et al. reported that they had detected a large number of apparent duplications in the summary statistics </span><span style="background-color: white; color: #222222;">across</span><span style="background-color: white; color: #222222;"> </span><span style="background-color: white; color: #222222; font-family: inherit;">those articles, which mostly describe randomized controlled trials carried out in the Mansoura ObGyn department. Nine of these articles appear as chapters in </span><a href="https://dspace.library.uu.nl/handle/1874/31693" style="font-family: inherit;" target="_blank">Dr Badawy's PhD thesis</a> <span style="background-color: white; color: #222222; font-family: inherit;">, which he defended in December 2008 at the University of Utrecht in the Netherlands.</span></p><p><span style="background-color: white; color: #222222; font-family: inherit;">I think it is fair to say that Dr Badawy was not especially impressed by the arguments in Bordewijk et al.'s commentary; indeed, he wrote a <a href="https://www.ejog.org/article/S0301-2115(20)30370-5/fulltext" target="_blank">reply</a> with the uncompromising title of "Data integrity of randomized controlled trials: A hate speech or scientific work?" in which he questioned, among other things, the simulation techniques that Bordewijk et al. had used to demonstrate how unlikely it was that the patterns that they had observed across the 35 articles that they examined had arisen by chance.</span></p><p><span style="background-color: white; color: #222222; font-family: inherit;">The senior author on </span><span style="background-color: white; color: #222222;">the Bordewijk et al. commentary was </span><span style="background-color: white; color: #222222; font-family: inherit;">Dr Ben Mol of Monash University in Melbourne, Australia. Since </span>their commentary was published, Dr Mol and his colleagues have been attempting to get the journal editors who published the 35 articles in question to take some form of action on them. To date, five articles have been retracted and another 10 have expressions of concern. The story, including its potential legal fallout, has been covered in considerable detail at Retraction Watch in <a href="https://retractionwatch.com/2020/10/12/researchers-face-disciplinary-action-as-dozens-of-their-studies-fall-under-scrutiny/" target="_blank">December 2020</a> and again in <a href="https://retractionwatch.com/2021/08/10/critics-face-legal-threats-as-journal-takes-more-than-three-years-to-act/" target="_blank">August 2021</a>.</p><p>However, to some extent, Dr <span style="background-color: white; color: #222222;">Badawy</span><span style="background-color: white; color: #222222;"> may have a point: T</span>he evidence presented in the commentary is circumstantial and depends on a number of probabilistic assumptions, which editors may not be inclined to completely trust (although personally I find Bordewijk et al.'s analysis thoroughly convincing). And, as an editor, even if you believe that two or more articles are based on recycled data or summary statistics, how are you to know that the one in <i>your</i> journal is not the original ("good") one?</p><p>Fortunately (at least from the point of view of the error correction process) there is a much simpler approach to the problem at hand. It can be shown that almost all of the articles that were analysed by Bordewijk et al.—plus a few more that did not make it into their commentary—have very substantial statistical flaws at the level of each <i>individual</i> article. In my opinion, in most cases these errors would justify a rapid retraction based solely on the evidence that is to be found in each article's PDF file. There is no need for simulations or probability calculations; in the majority of cases, the numbers sitting there in the tables of results are demonstrably incorrect.</p><p><br /></p><h4 style="text-align: left;">General description of the articles</h4><p>As mentioned above, for this blog post I examined 46 articles from the<span style="background-color: white; color: #222222;"> </span><span style="background-color: white; color: #222222;">Department of Obstetrics and Gynaecology at </span>Mansoura University. Of these, 35 had already been analysed by Bordewijk et al., and the rest were included either at the suggestion of Ben Mol or after I searched for any other empirical studies that I could find in the Google Scholar profiles of <a href="https://scholar.google.com/citations?hl=en&user=ezrPsaEAAAAJ" target="_blank">Dr Badawy</a> and <a href="https://scholar.google.com/citations?hl=en&user=1cyzL-MAAAAJ" target="_blank">Dr Abu Hashim</a>. Seven of the 46 articles had neither Dr Badawy nor Dr Abu Hashim as co-authors, but for all seven of those <a href="https://www.researchgate.net/profile/Tarek-Shokeir" target="_blank">Dr Tarek Shokeir</a> was listed as a co-author (or, in one case, sole author).</p><p>These articles mostly describe RCTs of various interventions for conditions such as infertility, heavy menstrual bleeding, polycystic ovary syndrome, preterm labour, or endometriosis. Several of them have more than 100 citations according to Google Scholar. The studies seem to be well-powered, many with more than 100 participants in each group (e.g., for <a href="https://doi.org/10.1007/s00404-007-0527-x" target="_blank">this one</a> the authors claimed to have recruited 996 infertile women), and it is not hard to imagine that their findings may be affecting clinical practice around the world.</p><p>The typical article is relatively short, and contains a baseline table comparing the groups of patients (usually two), followed by one or sometimes more tables comparing the outcomes across those groups. These are usually expressed as simple unpaired comparisons of parameters (e.g., height, with mean and standard deviation reported for each group), or as tests of proportions (e.g., in the treatment group X% of N1 participants became pregnant, versus Y% of N2 participants in the control group). The statistics are therefore for the most part very simple; for example, there are no logistic regressions with covariates. This means that we can readily check most of the statistics from the tables themselves.</p><p><br /></p><h4 style="text-align: left;"><span style="background-color: white; color: #222222; font-family: inherit;">The <i>t</i> statistics</span></h4><p><span style="color: #222222;"><span style="background-color: white;">First up, I note that in about half of these articles, no <i>t</i> statistics at all are reported for the comparisons of continuous variables across groups. Sometimes we get just a <i>p</i> value. In other cases we are only told that individual comparisons (or all of the comparisons in a table, via a note at the end) were statistically significant or not; typically we are left to infer that that means <i>p</i> < 0.05. (</span></span><span style="background-color: white; color: #222222;">In a few articles the authors reported using the Mann-Whitney U test when data were not normally distributed, but they do not generally indicate which variables are concerned by this in each case.)</span></p><p><span style="color: #222222;"><span style="background-color: white;">In quite a few cases the errors in the implicit <i>t</i> statistics are visible from space, as in this example from <a href="https://doi.org/10.1111/j.1447-0756.2010.01383.x" target="_blank">10.1111/j.1447-0756.2010.01383.x</a>:</span></span></p><p><span style="color: #222222;"><span style="background-color: white;"></span></span></p><div class="separator" style="clear: both; text-align: center;"><span style="color: #222222;"><span style="background-color: white;"><a href="https://lh3.googleusercontent.com/-23ZktNbQWVs/YXnBv3Di5wI/AAAAAAAAGXo/SOp2uxskq_Y2Zf5ExQ1JEDOXuxLx962wgCLcBGAsYHQ/image.png" style="margin-left: 1em; margin-right: 1em;"><img alt="" data-original-height="387" data-original-width="728" height="340" src="https://lh3.googleusercontent.com/-23ZktNbQWVs/YXnBv3Di5wI/AAAAAAAAGXo/SOp2uxskq_Y2Zf5ExQ1JEDOXuxLx962wgCLcBGAsYHQ/w640-h340/image.png" width="640" /></a>(<span style="font-family: inherit;">This table has been truncated on the right in order to fit on the page.)</span></span></span></div><p></p><span style="background-color: white; color: #222222; font-family: inherit;"><div style="line-height: 24px; text-align: center;"><span style="background-color: white; color: #222222; font-family: inherit;"><br /></span></div><div style="line-height: 24px;"><span style="background-color: white; color: #222222; font-family: inherit;"><br /></span></div>Have a look at the "Fasting glucose" numbers (fourth line from the bottom). The difference between the means is 5.4 (which means a minimum of 5.3 even after allowing for rounding) and just by approximating a weighted mean you can see that the pooled SD is going to be about 1.6, so this is a Cohen's <i>d</i> of around 3.3, which is never going to be non-significant at the 0.05 level. You don't have to carry the formula for the pooled standard error around in your head to know that with <i>df</i> = 136 the <i>t</i> statistic here is going to be huge, and indeed it is: the minimum <i>t</i> value is 17.31, the midpoint is 18.17, the maximum is 19.09, <i>p</i> = homeopathic. (Aside: p</span><span style="background-color: white; color: #222222; text-align: center;">hysiologists might wonder about the homogeneity of the testosterone levels in the ovulation group, with an SD of just 0.01.)</span></div><div style="line-height: 24px;"><span style="background-color: white; color: #222222; font-family: inherit;"><br /></span></div><div style="line-height: 24px;"><span style="background-color: white; color: #222222; font-family: inherit;">When we do get to see <i>t</i> statistics, the majority are incorrect given the mean, SD, and sample size, and <a href="https://steamtraen.blogspot.com/2018/07/this-researcher-compared-two-identical.html" target="_blank">even after allowing for rounding</a> (but see also <a href="https://psyarxiv.com/ctu9z/" target="_blank">our RIVETS preprint</a> for what happens if test statistics are derived with <i>too much</i> allowance for rounding). See this table from <a href="https://doi.org/10.1016/j.fertnstert.2008.06.013" target="_blank">10.1016/j.fertnstert.2008.06.013</a>:</span></div><div style="line-height: 24px;"><span style="background-color: white; color: #222222; font-family: inherit;"><br /></span></div><div style="line-height: 24px;"><span style="background-color: white; color: #222222; font-family: inherit;"><div class="separator" style="clear: both; text-align: center;"><a href="https://lh3.googleusercontent.com/-AoITGgtaQ-U/YXboxo7Et4I/AAAAAAAAGWc/ZCuCB5uya38rPUkiMsBcrbLyvdA4SLVbQCLcBGAsYHQ/image.png" style="margin-left: 1em; margin-right: 1em;"><img alt="" data-original-height="390" data-original-width="892" height="280" src="https://lh3.googleusercontent.com/-AoITGgtaQ-U/YXboxo7Et4I/AAAAAAAAGWc/ZCuCB5uya38rPUkiMsBcrbLyvdA4SLVbQCLcBGAsYHQ/w640-h280/image.png" width="640" /></a></div><br />Several of the effect sizes here are huge; for example, it is close to <i>d</i> = 4 on the first line, for which the correct <i>t</i> statistic is not 4.4 but instead somewhere between 38.71 and 40.26. If you are more familiar with ANOVA, you can square those numbers to get the corresponding <i>F</i> statistic. (Spoiler: <i>F</i> = 1,500 is <i>a lot</i>.)</span></div><div style="line-height: 24px;"><span style="background-color: white; color: #222222; font-family: inherit;"><br /></span></div><div style="line-height: 24px;"><span style="background-color: white; color: #222222; font-family: inherit;">The overall results for the set of 46 articles are disastrous. In most of the articles the majority of the <i>p</i> values are incorrect, sometimes by a wide margin. Even in those articles where the authors did not explicitly report which test they used, leaving open the possibility that they might have used the Mann-Whitney U test, the discrepancies between the reported <i>p</i> values and those that I obtained from a <i>t</i> test were very often such that even the discovery that the Mann-Whitney test had been used would not be sufficient to explain them.</span></div><div style="line-height: 24px;"><span style="background-color: white; color: #222222; font-family: inherit;"><br /></span></div><div style="line-height: 24px;"><span style="background-color: white; color: #222222; font-family: inherit;">[Begin update 2021-10-28 19:57 UTC]</span></div><div style="line-height: 24px;"><span style="color: #222222;"><span style="background-color: white;">Brenton Wiernik <a href="https://twitter.com/bmwiernik/status/1453790285402677254?s=20" target="_blank">asked</a> whether it was possible that the authors had accidentally reported the variability of their means as the standard error of the mean (SEM) rather than the standard deviation. I must confess that when doing the analyses I did not consider this possibility, not least because all of the SDs for things like age (around 3) or height (around 5–6cm) seemed quite reasonable</span></span><span style="background-color: white; color: #222222;">, although perhaps a bit on the low side of what one might expect.</span><span style="background-color: white; color: #222222;"> </span></div><div style="line-height: 24px;"><span style="color: #222222;"><span style="background-color: white;"><br /></span></span></div><div style="line-height: 24px;"><span style="color: #222222;"><span style="background-color: white;">However, I did identify seven articles (</span><a href="https://doi.org/10.3109/01443615.2010.497873" target="_blank">10.3109/01443615.2010.497873</a></span><span style="background-color: white; color: #222222;">, </span><span style="color: #222222;"><a href="https://doi.org/10.1016/j.fertnstert.2008.04.065" target="_blank">10.1016/j.fertnstert.2008.04.065</a>, </span><a href="https://doi.org/10.1016/j.fertnstert.2008.06.013" target="_blank">10.1016/j.fertnstert.2008.06.013</a>, <a href="https://doi.org/10.1016/j.fertnstert.2007.08.034" target="_blank">10.1016/j.fertnstert.2007.08.034</a>, <a href="https://doi.org/10.1016/s1472-6483(10)60148-4" target="_blank">10.1016/s1472-6483(10)60148-4</a>, <a href="https://doi.org/10.1016/j.fertnstert.2007.05.010" target="_blank">10.1016/j.fertnstert.2007.05.010</a>, and <a href="https://doi.org/10.1016/j.fertnstert.2007.02.062" target="_blank">10.1016/j.fertnstert.2007.02.062</a>)<span style="background-color: white; color: #222222;"> in which the authors claimed that their measure of variability (typically written as ± and a number after the mean) was indeed the SEM. But the sample sizes in these papers are such that the implied SDs would need to be huge. Across these seven papers, the implied SD for age would range from 29.50 to 62.39 years, and for height from 54.79 to 124.77 cm. I therefore stand by my interpretation that these values were intended by the authors to be standard deviations, although that then gives them another question to answer, namely why they claimed them to be SEM when those numbers are patently absurd.</span></div><div style="line-height: 24px;"><span style="background-color: white; color: #222222; font-family: inherit;"><div style="color: black; line-height: 24px;"><span style="color: #222222; font-family: inherit;">[End update 2021-10-28 19:57 UTC]</span></div></span></div><p><br /></p><h4><span style="background-color: white; color: #222222; font-family: inherit;">The </span><i style="background-color: white; font-family: inherit;">Χ</i><span style="background-color: white; font-family: inherit;">² (etc) </span>statistics</h4><div><br /></div><div style="line-height: 24px;"><span style="background-color: white;"><div><span style="color: #222222; font-family: inherit;">In examining whether the tests of proportions had been reported correctly, I included only those articles (30 out of the total set of 46) that contained at least one exact numerical (i.e., not "NS" or "<0.001") </span><i style="color: #222222; font-family: inherit;">p</i><span style="color: #222222; font-family: inherit;"> value from a Pearson chi-square test or Fisher's exact test of a 2x2 contingency table. If the authors also reported </span><span style="color: #222222;"><i>Χ</i>²</span><span style="color: #222222; font-family: inherit;"> statistics and/or odds ratios, I also included those numbers. I then examined the extent to which these statistics matched the values that I calculated from the underlying data. When the subsample sizes were very small, I allowed the authors some more leeway, as the Pearson chi-square test does not always perform well in these cases.</span></div><div><span style="color: #222222; font-family: inherit;"><br /></span></div><div><span style="color: #222222; font-family: inherit;">As with the <i>t</i> test results (see previous section), the overall results revealed a large number of incorrect <i>p</i> values in almost every article for which I recalculated the tests of proportions.</span></div><div style="color: #222222; font-family: inherit;"><span style="color: #222222; font-family: inherit;"><br /></span></div><div><span style="color: #222222;">Perhaps the most indisputable source of errors is situations in which what is effectively the same test is reported twice, with different chi-square statistics (if those are reported) and different <i>p</i> values, even though those values are necessarily identical. I counted 8 examples of this across 7 different articles. For example, consider this table from <a href="https://doi.org/10.3109/01443615.2010.508850" target="_blank">10.3109/01443615.2010.508850</a>:<br /></span></div><div><span style="color: #222222;"><br /></span></div><div><span style="color: #222222;"><div class="separator" style="clear: both; text-align: center;"><div class="separator" style="clear: both; text-align: center;"><a href="https://lh3.googleusercontent.com/-kKm_S3_t5q4/YXa8IBkG7AI/AAAAAAAAGV0/AtZuaIw7FpEzU1mOiGOFkIAs2EB30zsSACLcBGAsYHQ/image.png" style="margin-left: 1em; margin-right: 1em;"><img alt="" data-original-height="196" data-original-width="591" height="212" src="https://lh3.googleusercontent.com/-kKm_S3_t5q4/YXa8IBkG7AI/AAAAAAAAGV0/AtZuaIw7FpEzU1mOiGOFkIAs2EB30zsSACLcBGAsYHQ/w640-h212/image.png" width="640" /></a>You're either pregnant or you aren't. I don't make the rules.</div><div style="text-align: left;"><br /></div></div>After 6 months in the study, every participant either had, or had not, become pregnant. So the contingency table for the first outcome ("No pregnancy") is ((141, 114), (150, 107)) and for the second outcome ("Clinical pregnancy") it is </span><span style="color: #222222;">((114, 141), (107, 150)). Those will of course give exactly the same result, meaning that at least one of the </span><span style="color: #222222;"><i>Χ</i>²/<i>p</i>-value pairs <a href="https://www.youtube.com/watch?v=D1XGQX98reM#t=3m23s" target="_blank">must be wrong</a>. In fact the correct numbers are </span><span style="color: #222222;"><i>Χ</i>²(1) = 0.49, <i>p</i> = 0.48, which means that neither of the </span><span style="color: #222222;"><i>Χ</i>² </span><span style="color: #222222;">test statistics, nor the <i>p</i> values, in the first two lines of the table match are even remotely valid. For that matter, neither (incorrect) <i>p</i> value in the table even matches its corresponding (incorrect) </span><span style="color: #222222;"><i>Χ</i>² statistic; I will let you check this for yourself as an exercise.</span></div></span><p><br /></p><h4><span style="background-color: white; color: #222222; font-family: inherit;">The <i>p</i> values</span></h4><p><span style="background-color: white; color: #222222;">There are a couple of basic things to keep in mind when reading tables of statistics:</span></p><div><ul style="text-align: left;"><li>A <i>t</i> statistic of 1.96 with 100 or more degrees of freedom gives a rounded two-tailed <i>p</i> value of 0.05 (although if you want it to be strictly less than 0.05000, you need a t statistic of 1.984 with 100 <i>df</i>s).</li><li>For any given number of degrees of freedom, a larger <i>t</i> or <span face="arial, sans-serif" style="background-color: white; color: #4d5156; font-size: 14px;"><i>Χ</i>² </span>statistic gives a smaller <i>p</i> value.</li></ul><div>With that in mind, let's look at this table (from <a href="https://doi.org/10.1016/j.fertnstert.2007.05.010" target="_blank">10.1016/j.fertnstert.2007.05.010</a>), which I believe to be entirely typical of the articles under discussion here:</div><div><div class="separator" style="clear: both; text-align: left;"><br /></div><div class="separator" style="clear: both; text-align: center;"><div class="separator" style="clear: both; text-align: center;"><a href="https://lh3.googleusercontent.com/-6ZCNjcbTAzE/YXg3Gm5GcdI/AAAAAAAAGWw/gITp_H9jNwEYPGYoxbpK0eGbEXg-aBCawCLcBGAsYHQ/image.png" style="margin-left: 1em; margin-right: 1em;"><img alt="" data-original-height="466" data-original-width="882" height="338" src="https://lh3.googleusercontent.com/-6ZCNjcbTAzE/YXg3Gm5GcdI/AAAAAAAAGWw/gITp_H9jNwEYPGYoxbpK0eGbEXg-aBCawCLcBGAsYHQ/w640-h338/image.png" width="640" /></a></div>The good news is that the percentages and the <span face="arial, sans-serif" style="background-color: white; color: #4d5156; font-size: 14px; text-align: left;"><i>Χ</i>² </span><span style="text-align: left;">statistics </span>check out OK!</div><div class="separator" style="clear: both; text-align: center;"><div style="text-align: left;"><br /></div></div><div class="separator" style="clear: both; text-align: left;"><br /></div><div class="separator" style="clear: both; text-align: left;">Nether of the above bits of very basic knowledge is respected here. First we have <i>t</i>(228) = 2.66, <i>p</i> = 0.1 (when the correct <i>p</i> value for that <i>t</i> is 0.008—although in any case the correct <i>t</i> statistic for the given means and SDs would be between 5.05 and 5.83). Second, between FSH (follicle-stimulating hormone) and LH (luteinizing hormone), the <i>t</i> statistic goes up and so does the <i>p</i> value (which is also clearly incorrect in both cases).</div><br /><div>A number of the articles contain <i>p</i> values that are literally impossible (i.e., greater than 1.0; don't @ me to tell me about that time you did a Bonferroni correction by multiplying the <i>p</i> value instead of dividing the alpha). See <a href="https://doi.org/10.3109/01443615.2010.497873" target="_blank">10.3109/01443615.2010.497873</a> (Table 1, "Parity", p = 1.13; see also "Other inconsistencies", below), <a href="https://doi.org/10.1016/j.fertnstert.2008.04.065" target="_blank">10.1016/j.fertnstert.2008.04.065</a> (Table 1, "Height", <i>p </i>= 1.01), and <a href="https://doi.org/10.1007/s00404-013-2866-0" target="_blank">10.1007/s00404-013-2866-0</a>, which contains no less than four examples across its Tables 1 and 2:</div><div><br /></div><div><div class="separator" style="clear: both; text-align: center;"><div class="separator" style="clear: both; text-align: center;"><a href="https://lh3.googleusercontent.com/-erX1IH1z848/YXg9BoMN0dI/AAAAAAAAGXA/wEhu3SJeX6M6z_5sKEjuaUQ9FYRZWwUHACLcBGAsYHQ/image.png" style="margin-left: 1em; margin-right: 1em;"><img alt="" data-original-height="882" data-original-width="582" src="https://lh3.googleusercontent.com/-erX1IH1z848/YXg9BoMN0dI/AAAAAAAAGXA/wEhu3SJeX6M6z_5sKEjuaUQ9FYRZWwUHACLcBGAsYHQ/s16000/image.png" /></a></div></div></div><div class="separator" style="clear: both; line-height: 16px; text-align: center;">1.12, 1.22, 1.32, 1.42: The impossible <i>p</i> values form a nice pattern.</div></div></div><div><span style="font-weight: normal;"><br /></span></div><div><span style="font-weight: normal;"><br /></span></div><h4 style="text-align: left;">The confidence intervals</h4><p>A few of the articles have confidence intervals in the tables, perhaps added at the insistence of a reviewer or editor. But in most cases the point estimate falls <i>outside</i> the confidence interval. Sometimes this can become quite absurd, as in the following example (from <a href="https://doi.org/10.1016/j.fertnstert.2007.08.034" target="_blank">10.1016/j.fertnstert.2007.08.034</a>). Those CI limits are ± 1.96 standard errors either side of... what exactly?</p><div class="separator" style="clear: both; text-align: center;"><div class="separator" style="clear: both; text-align: center;"><div class="separator" style="clear: both; text-align: center;"><div style="text-align: left;"><div class="separator" style="clear: both; text-align: center;"><div class="separator" style="clear: both; text-align: center;"><a href="https://lh3.googleusercontent.com/-PmDwPXJfTw4/YXg95eYOgNI/AAAAAAAAGXI/DH7jcdOarV4U5gd6P_cAxW8rKEA6NHQPwCLcBGAsYHQ/image.png" style="margin-left: 1em; margin-right: 1em;"><img alt="" data-original-height="435" data-original-width="788" height="354" src="https://lh3.googleusercontent.com/-PmDwPXJfTw4/YXg95eYOgNI/AAAAAAAAGXI/DH7jcdOarV4U5gd6P_cAxW8rKEA6NHQPwCLcBGAsYHQ/w640-h354/image.png" width="640" /></a></div></div><div class="separator" style="clear: both; text-align: center;">(Once you look beyond the CIs, there is a "bonus" waiting in the <i>p</i> value column here.)</div></div></div></div></div><p></p><h4 style="text-align: left;"><br /></h4><h4 style="text-align: left;">Other inconsistencies</h4><p style="text-align: left;">Within these 46 articles it is hard not to notice a considerable number of other inconsistencies, which make the reader wonder how much care and attention went into both the writing and review processes. These tables from <a href="https://doi.org/10.3109/01443615.2010.497873" target="_blank">10.3109/01443615.2010.497873</a> provide a particularly egregious example, with the appearance of 41 and 42 extra patients in the respective groups between baseline and outcome. (As a bonus, we also have a <i>p</i> value of 1.13.)</p><div><div class="separator" style="clear: both; text-align: center;"><div class="separator" style="clear: both; text-align: center;"><div class="separator" style="clear: both; text-align: center;"><a href="https://lh3.googleusercontent.com/-FNFspMsqgqE/YXcSMFZA0fI/AAAAAAAAGWk/U5N8YYpnQGgnSBv7LVJeQQH9-A6-IYFcACLcBGAsYHQ/image.png" style="margin-left: 1em; margin-right: 1em;"><img alt="" data-original-height="606" data-original-width="783" height="495" src="https://lh3.googleusercontent.com/-FNFspMsqgqE/YXcSMFZA0fI/AAAAAAAAGWk/U5N8YYpnQGgnSBv7LVJeQQH9-A6-IYFcACLcBGAsYHQ/w640-h495/image.png" width="640" /></a></div></div>(Some white space has been removed from this table.)</div><div><br /></div><div><br /></div><div>These results from <a href="https://doi.org/10.1016/j.ejogrb.2012.09.014" target="_blank">10.1016/j.ejogrb.2012.09.014</a> make no sense. The first comparison was apparently done using Fisher's exact test, the second with Pearson's <i style="background-color: white; color: #222222;">Χ</i><span style="background-color: white; color: #222222;">² test, and the third, well, your guess is as good as mine. But there is no reason to use the two different types of test here, and even less reason to use Fisher's test for the larger case numbers and Pearson's for the smaller ones. (The <i>p</i> values are all incorrect, and would be even if the other test were to have been used for every variable.)</span></div><div><br /></div><div><div class="separator" style="clear: both; text-align: center;"><a href="https://lh3.googleusercontent.com/-VXgKeHfHhZU/YXhCJQ71khI/AAAAAAAAGXQ/4X85tJ3tGWgDc1cLRpBPtfImumF4aSr-wCLcBGAsYHQ/image.png" style="margin-left: 1em; margin-right: 1em;"><img alt="" data-original-height="480" data-original-width="1122" height="274" src="https://lh3.googleusercontent.com/-VXgKeHfHhZU/YXhCJQ71khI/AAAAAAAAGXQ/4X85tJ3tGWgDc1cLRpBPtfImumF4aSr-wCLcBGAsYHQ/w640-h274/image.png" width="640" /></a></div><div class="separator" style="clear: both; text-align: center;"><div class="separator" style="clear: both;">(Some white space and some results that are not relevant to the point being illustrated have been removed from this table.)</div></div><br /><br /></div>Finally, fans of <a href="https://journals.sagepub.com/doi/abs/10.1177/1948550616673876" target="_blank">GRIM</a> might also be interested to learn that two articles show signs of possible inconsistencies in their reported means:</div><div><a href="https://doi.org/10.1016/j.rbmo.2014.03.011" target="_blank">10.1016/j.rbmo.2014.03.011</a>, "Age" (both groups) and "Infertility" (control group)</div><div><a href="https://doi.org/10.1111/j.1447-0756.2012.02016.x" target="_blank">10.1111/j.1447-0756.2012.02016.x</a>, "Age" (both groups)</div><div><br /></div><div><br /></div><h4 style="text-align: left;">Conclusion</h4><p>The results of these analyses seem to indicate that something has gone very badly wrong in the writing, reviewing, and publication of these articles. Even though I tried to give the published numbers the benefit of the doubt as far as possible, I estimate that across these 46 articles, 346 (64%) of the 542 parametric tests (unpaired <i>t</i> tests, or, occasionally, ANOVA) and 151 (61%) of the 247 contingency table test (Pearson's <i style="background-color: white; color: #222222;">Χ</i><span style="background-color: white; color: #222222;">² or Fisher's exact test</span>) that I was able to check were incorrectly reported. I don't think that anybody should be relying on the conclusions of these articles as a guide to practice, and I suspect that the only solution for most of them will be retraction. (As already mentioned, five have already been retracted following the publication of Bordewijk et al.'s commentary.)</p><p>I have a few aims in writing this post.</p><p>First, I want to do whatever I can to help get these misleading (at best) papers retracted from the medical literature, where they would seem to have considerable potential to do serious harm to the health of women, especially those who are pregnant or trying to overcome infertility.</p><p>Second, I aim to show some of the techniques that can be used to detect obvious errors in published articles (or in manuscripts that you might be reviewing).</p><p>Third, and the important reason for doing all this work (it took a lot of hours to do these analyses, as you will see if you download the Excel file!), is to draw attention to the utter failure of peer review that was required in order for most of these articles to get published. They appeared in 13 different journals, none of which would appear to correspond to most people's idea of a "predatory" outlet. It is very tempting to imagine that nobody—editors, reviewers, Dr Badawy's thesis committee at the University of Utrecht, or readers of the journals (until Ben Mol and Esmée Bordewijk came along)—even so much glanced at the tables of results in these articles, given that they almost all contain multiple impossible numbers.</p><p>It is true that the majority of these articles are more than 10 years old, but I wonder how much has changed in the publication processes of medical journals since then. The reality of scientific peer review seems to be that, to a first approximation, <i>nobody ever checks any of the numbers</i>. I find that deeply worrying.</p><p><br /></p><h4 style="text-align: left;">Supporting documents</h4><p>The majority of the analyses underlying this post have been done with Microsoft Excel 2003. I have some R code that can do the same thing, but it seemed to me to make more sense to use Excel as the process of copying and pasting numbers from the tables in the articles was a lot more reliable, requiring only a text editor to replace the column separators with tab characters. I used my R code to compute the test statistics in a couple of cases where there were more than two groups and so I had to use the <i>rpsychi</i> package to calculate the results of a one-way ANOVA.</p><p>In my Excel file, each unpaired <i>t</i> test is performed on a separate line. The user enters the mean, standard deviation, and sample size for each of the two conditions, plus an indication of the rounding precision (i.e, the number of decimal places) for the means and SDs separately. The spreadsheet then calculates (using formulas that you can find in columns whose width I have in most cases reduced to zero) the minimum and maximum possible (i.e., before rounding) means and SDs, and from that it determines the minimum, notional (i.e., assuming that the rounded input values are exact), and maximum <i>t</i> statistics and the corresponding <i>p</i> values. It then highlights those <i>t</i> statistics (if available) and <i>p</i> values (or "significant/not-significant" claims) from the article that are not compatible with any point in the possible ranges of values. That is, at all times, I give the maximum benefit of the doubt to the authors. (Similar considerations apply, <i>mutatis mutandis</i>, to the table of chi-square tests in the same Excel file.)</p><p>The documents (an Excel file for the main analyses and some R code for the bits that I couldn't work out how to do easily in Excel) are available <a href="http://nickbrown.fr/blog/mansoura" target="_blank">here</a>. The article PDFs are all copyrighted and I cannot share them, but if you do not have institutional access then there is always the site whose name rhymes with Dry Club.</p><p><br /></p><h4 style="text-align: left;">Appendix: List of examined articles</h4><p><u>Articles that have not been retracted and have no expression of concern</u></p><p>Badawy et al. (2009). Gonadotropin-releasing hormone agonists for prevention of chemotherapy-induced ovarian damage: Prospective randomized study. <i>Fertility and Sterility</i>. <a href="https://doi.org/10.1016/j.fertnstert.2007.12.044" target="_blank">https://doi.org/10.1016/j.fertnstert.2007.12.044</a></p><p>Badawy et al. (2007). Induction of ovulation in idiopathic premature ovarian failure: A randomized double-blind trial. <i>Reproductive Biomedicine Online</i>. <a href="https://doi.org/10.1016/s1472-6483(10)60711-0" target="_blank">https://doi.org/10.1016/s1472-6483(10)60711-0</a></p><p>Badawy et al. (2010). Clomiphene citrate or aromatase inhibitors combined with gonadotropins for superovulation in women undergoing intrauterine insemination: A prospective randomised trial. <i>Journal of Obstetrics and Gynaecology</i>. <a href="https://doi.org/10.3109/01443615.2010.497873" target="_blank">https://doi.org/10.3109/01443615.2010.497873</a></p><p>Badawy et al. (2009). Ultrasound-guided transvaginal ovarian needle drilling (UTND) for treatment of polycystic ovary syndrome: A randomized controlled trial. <i>Fertility and Sterility</i>. <a href="https://doi.org/10.1016/j.fertnstert.2008.01.044" target="_blank">https://doi.org/10.1016/j.fertnstert.2008.01.044</a></p><p>Badawy et al. (2008). Low-molecular weight heparin in patients with recurrent early miscarriages of unknown aetiology. <i>Journal of Obstetrics and Gynaecology</i>. <a href="https://doi.org/10.1080/01443610802042688" target="_blank">https://doi.org/10.1080/01443610802042688</a></p><p>Badawy et al. (2010). Laparoscopy--or not--for management of unexplained infertility. <i>Journal of Obstetrics and Gynaecology</i>. <a href="https://doi.org/10.3109/01443615.2010.508850" target="_blank">https://doi.org/10.3109/01443615.2010.508850</a></p><p>Badawy et al. (2007). Plasma homocysteine and polycystic ovary syndrome: The missed link. European <i>Journal of Obstetrics & Gynecology and Reproductive Biology</i>. <a href="https://doi.org/10.1016/j.ejogrb.2006.10.015" target="_blank">https://doi.org/10.1016/j.ejogrb.2006.10.015</a></p><p>Badawy et al. (2008). Extending clomiphene treatment in clomiphene-resistant women with PCOS: A randomized controlled trial. <i>Reproductive Biomedicine Online</i>. <a href="https://doi.org/10.1016/s1472-6483(10)60148-4" target="_blank">https://doi.org/10.1016/s1472-6483(10)60148-4</a></p><p>Badawy et al. (2006). Clomiphene citrate plus N-acetyl cysteine versus clomiphene citrate for augmenting ovulation in the management of unexplained infertility: A randomized double-blind controlled trial. <i>Fertility and Sterility</i>. <a href="https://doi.org/10.1016/j.fertnstert.2006.02.097" target="_blank">https://doi.org/10.1016/j.fertnstert.2006.02.097</a></p><p>Badawy et al. (2007). Randomized controlled trial of three doses of letrozole for ovulation induction in patients with unexplained infertility. <i>Reproductive Biomedicine Online</i>. <a href="https://doi.org/10.1016/s1472-6483(10)61046-2" target="_blank">https://doi.org/10.1016/s1472-6483(10)61046-2</a></p><p>Fawzy et al. (2007). Treatment options and pregnancy outcome in women with idiopathic recurrent miscarriage: A randomized placebo-controlled study. <i>Archives of Gynecology and Obstetrics</i>. <a href="https://doi.org/10.1007/s00404-007-0527-x" target="_blank">https://doi.org/10.1007/s00404-007-0527-x</a></p><p>Gibreel et al. (2012). Endometrial scratching to improve pregnancy rate in couples with unexplained subfertility: A randomized controlled trial. <i>Journal of Obstetrics and Gynaecology Research</i>. <a href="https://doi.org/10.1111/j.1447-0756.2012.02016.x" target="_blank">https://doi.org/10.1111/j.1447-0756.2012.02016.x</a></p><p>Abu Hashim et al. (2010). Combined metformin and clomiphene citrate versus highly purified FSH for ovulation induction in clomiphene-resistant PCOS women: A randomised controlled trial. <i>Gynecological Endocrinology</i>. <a href="https://doi.org/10.3109/09513590.2010.488771" target="_blank">https://doi.org/10.3109/09513590.2010.488771</a></p><p>Abu Hashim et al. (2010). Letrozole versus laparoscopic ovarian diathermy for ovulation induction in clomiphene-resistant women with polycystic ovary syndrome: A randomized controlled trial. <i>Archives of Gynecology and Obstetrics</i>. <a href="https://doi.org/10.1007/s00404-010-1566-2" target="_blank">https://doi.org/10.1007/s00404-010-1566-2</a></p><p>Abu Hashim et al. (2011). Laparoscopic ovarian diathermy after clomiphene failure in polycystic ovary syndrome: is it worthwhile? A randomized controlled trial. <i>Archives of Gynecology and Obstetrics</i>. <a href="https://doi.org/10.1007/s00404-011-1983-x" target="_blank">https://doi.org/10.1007/s00404-011-1983-x</a></p><p>Abu Hashim et al. (2012). Contraceptive vaginal ring treatment of heavy menstrual bleeding: A randomized controlled trial with norethisterone. <i>Contraception</i>. <a href="https://doi.org/10.1016/j.contraception.2011.07.012" target="_blank">https://doi.org/10.1016/j.contraception.2011.07.012</a></p><p>Abu Hashim et al. (2010). N-acetyl cysteine plus clomiphene citrate versus metformin and clomiphene citrate in treatment of clomiphene-resistant polycystic ovary syndrome: A randomized controlled trial. <i>Journal of Women's Health</i>. <a href="https://doi.org/10.1089/jwh.2009.1920" target="_blank">https://doi.org/10.1089/jwh.2009.1920</a></p><p>Abu Hashim et al. (2010). Combined metformin and clomiphene citrate versus laparoscopic ovarian diathermy for ovulation induction in clomiphene-resistant women with polycystic ovary syndrome: A randomized controlled trial. <i>Journal of Obstetrics and Gynaecology Research</i>. <a href="https://doi.org/10.1111/j.1447-0756.2010.01383.x" target="_blank">https://doi.org/10.1111/j.1447-0756.2010.01383.x</a></p><p>Abu Hashim et al. (2011). Minimal stimulation or clomiphene citrate as first-line therapy in women with polycystic ovary syndrome: A randomized controlled trial. <i>Gynecological Endocrinology</i>. <a href="https://doi.org/10.3109/09513590.2011.589924" target="_blank">https://doi.org/10.3109/09513590.2011.589924</a></p><p>Abu Hashim et al. (2011). Does laparoscopic ovarian diathermy change clomiphene-resistant PCOS into clomiphene-sensitive? <i>Archives of Gynecology and Obstetrics</i>. <a href="https://doi.org/10.1007/s00404-011-1931-9" target="_blank">https://doi.org/10.1007/s00404-011-1931-9</a></p><p>Abu Hashim et al. (2013). LNG-IUS treatment of non-atypical endometrial hyperplasia in perimenopausal women: A randomized controlled trial. <i>Journal of Gynecologic Oncology</i>. <a href="https://doi.org/10.3802/jgo.2013.24.2.128" target="_blank">https://doi.org/10.3802/jgo.2013.24.2.128</a></p><p>Marzouk et al. (2014). Lavender-thymol as a new topical aromatherapy preparation for episiotomy: A randomised clinical trial. <i>Journal of Obstetrics and Gynaecology</i>. <a href="https://doi.org/10.3109/01443615.2014.970522" target="_blank">https://doi.org/10.3109/01443615.2014.970522</a></p><p>Ragab et al. (2013). Does immediate postpartum curettage of the endometrium accelerate recovery from preeclampsia-eclampsia? A randomized controlled trial. <i>Archives of Gynecology and Obstetrics</i>. <a href="https://doi.org/10.1007/s00404-013-2866-0" target="_blank">https://doi.org/10.1007/s00404-013-2866-0</a></p><p>El Refaeey et al. (2014). Combined coenzyme Q10 and clomiphene citrate for ovulation induction in clomiphene-citrate-resistant polycystic ovary syndrome. <i>Reproductive Biomedicine Online</i>. <a href="https://doi.org/10.1016/j.rbmo.2014.03.011" target="_blank">https://doi.org/10.1016/j.rbmo.2014.03.011</a></p><p>Seleem et al. (2014). Superoxide dismutase in polycystic ovary syndrome patients undergoing intracytoplasmic sperm injection. <i>Archives of Gynecology and Obstetrics</i>. <a href="https://doi.org/10.1007/s10815-014-0190-7">https://doi.org/10.1007/s10815-014-0190-7</a></p><p>Shokeir (2006). Tamoxifen citrate for women with unexplained infertility. <i>Archives of Gynecology and Obstetrics</i>. <a href="https://doi.org/10.1007/s00404-006-0181-8" target="_blank">https://doi.org/10.1007/s00404-006-0181-8</a> *</p><p>Shokeir et al. (2016). Hysteroscopic-guided local endometrial injury does not improve natural cycle pregnancy rate in women with unexplained infertility: Randomized controlled trial. <i>Journal of Obstetrics and Gynaecology Research</i>. <a href="https://doi.org/10.1111/jog.13077" target="_blank">https://doi.org/10.1111/jog.13077</a></p><p>Shokeir et al. (2009). The efficacy of Implanon for the treatment of chronic pelvic pain associated with pelvic congestion: 1-year randomized controlled pilot study. <i>Archives of Gynecology and Obstetrics</i>. <a href="https://doi.org/10.1007/s00404-009-0951-1" target="_blank">https://doi.org/10.1007/s00404-009-0951-1</a> *</p><p>Shokeir & Mousa (2015). A randomized, placebo-controlled, double-blind study of hysteroscopic-guided pertubal diluted bupivacaine infusion for endometriosis-associated chronic pelvic pain. <i>International Journal of Gynecology & Obstetrics</i>. <a href="https://doi.org/10.1016/j.ijgo.2015.03.043" target="_blank">https://doi.org/10.1016/j.ijgo.2015.03.043</a></p><p><u>Articles that are subject to an editorial </u><u>Expression of Concern</u></p><p>Badawy et al. (2012). Aromatase inhibitors or gonadotropin-releasing hormone agonists for the management of uterine adenomyosis: A randomized controlled trial. <i>Acta Obstetricia et Gynecologica Scandinavica</i>. <a href="https://doi.org/10.1111/j.1600-0412.2012.01350.x" target="_blank">https://doi.org/10.1111/j.1600-0412.2012.01350.x</a></p><p>Badawy et al. (2009). Extended letrozole therapy for ovulation induction in clomiphene-resistant women with polycystic ovary syndrome: A novel protocol. <i>Fertility and Sterility</i>. <a href="https://doi.org/10.1016/j.fertnstert.2008.04.065" target="_blank">https://doi.org/10.1016/j.fertnstert.2008.04.065</a></p><p>Badawy et al. (2008). Luteal phase clomiphene citrate for ovulation induction in women with polycystic ovary syndrome: A novel protocol. <i>Fertility and Sterility</i>. <a href="https://doi.org/10.1016/j.fertnstert.2008.01.016" target="_blank">https://doi.org/10.1016/j.fertnstert.2008.01.016</a></p><p>Badawy et al. (2007). N-Acetyl cysteine and clomiphene citrate for induction of ovulation in polycystic ovary syndrome: A cross-over trial. <i>Acta Obstetricia et Gynecologica Scandinavica</i>. <a href="https://doi.org/10.1080/00016340601090337" target="_blank">https://doi.org/10.1080/00016340601090337</a></p><p>Badawy et al. (2009). Clomiphene citrate or anastrozole for ovulation induction in women with polycystic ovary syndrome? A prospective controlled trial. <i>Fertility and Sterility</i>. <a href="https://doi.org/10.1016/j.fertnstert.2007.08.034" target="_blank">https://doi.org/10.1016/j.fertnstert.2007.08.034</a></p><p>Badawy et al. (2009). Pregnancy outcome after ovulation induction with aromatase inhibitors or clomiphene citrate in unexplained infertility. <i>Acta Obstetricia et Gynecologica Scandinavica</i>. <a href="https://doi.org/10.1080/00016340802638199" target="_blank">https://doi.org/10.1080/00016340802638199</a></p><p>Abu Hashim et al. (2011). Intrauterine insemination versus timed intercourse with clomiphene citrate in polycystic ovary syndrome: A randomized controlled trial. <i>Acta Obstetricia et Gynecologica Scandinavica</i>. <a href="https://doi.org/10.1111/j.1600-0412.2010.01063.x" target="_blank">https://doi.org/10.1111/j.1600-0412.2010.01063.x</a></p><p>Abu Hashim et al. (2012). Randomized comparison of superovulation with letrozole versus clomiphene citrate in an IUI program for women with recently surgically treated minimal to mild endometriosis. <i>Acta Obstetricia et Gynecologica Scandinavica</i>. <a href="https://doi.org/10.1111/j.1600-0412.2011.01346.x" target="_blank">https://doi.org/10.1111/j.1600-0412.2011.01346.x</a></p><p>Shokeir et al. (2011). An RCT: use of oxytocin drip during hysteroscopic endometrial resection and its effect on operative blood loss and glycine deficit. <i>Journal of Minimally Invasive Gynecology</i>. <a href="https://doi.org/10.1016/j.jmig.2011.03.015" target="_blank">https://doi.org/10.1016/j.jmig.2011.03.015</a></p><p>Shokeir et al. (2013). Does adjuvant long-acting gestagen therapy improve the outcome of hysteroscopic endometrial resection in women of low-resource settings with heavy menstrual bleeding? <i>Journal of Minimally Invasive Gynecology</i>. <a href="https://doi.org/10.1016/j.jmig.2012.11.006" target="_blank">https://doi.org/10.1016/j.jmig.2012.11.006</a></p><p>Badawy & Gibreal (2011). Clomiphene citrate versus tamoxifen for ovulation induction in women with PCOS: A prospective randomized trial. <i>European Journal of Obstetrics & Gynecology and Reproductive Biology</i>. <a href="https://doi.org/10.1016/j.ejogrb.2011.07.015" target="_blank">https://doi.org/10.1016/j.ejogrb.2011.07.015</a></p><p>Shokeir et al. (2013). Reducing blood loss at abdominal myomectomy with preoperative use of dinoprostone intravaginal suppository: A randomized placebo-controlled pilot study. <i>European Journal of Obstetrics & Gynecology and Reproductive Biology</i>. <a href="https://doi.org/10.1016/j.ejogrb.2012.09.014" target="_blank">https://doi.org/10.1016/j.ejogrb.2012.09.014</a></p><p><u>Articles that have been retracted</u></p><p>Badawy et al. (2009). Clomiphene citrate or aromatase inhibitors for superovulation in women with unexplained infertility undergoing intrauterine insemination: A prospective randomized trial. <i>Fertility and Sterility</i>. <a href="https://doi.org/10.1016/j.fertnstert.2008.06.013" target="_blank">https://doi.org/10.1016/j.fertnstert.2008.06.013</a></p><p>Badawy et al. (2009). Clomiphene citrate or letrozole for ovulation induction in women with polycystic ovarian syndrome: A prospective randomized trial. <i>Fertility and Sterility</i>. <a href="https://doi.org/10.1016/j.fertnstert.2007.02.062" target="_blank">https://doi.org/10.1016/j.fertnstert.2007.02.062</a></p><p>Badawy et al. (2008). Anastrozole or letrozole for ovulation induction in clomiphene-resistant women with polycystic ovarian syndrome: A prospective randomized trial. <i>Fertility and Sterility</i>. <a href="https://doi.org/10.1016/j.fertnstert.2007.05.010" target="_blank">https://doi.org/10.1016/j.fertnstert.2007.05.010</a></p><p>Abu Hashim et al. (2010). Letrozole versus combined metformin and clomiphene citrate for ovulation induction in clomiphene-resistant women with polycystic ovary syndrome: A randomized controlled trial. <i>Fertility and Sterility</i>. <a href="https://doi.org/10.1016/j.fertnstert.2009.07.985" target="_blank">https://doi.org/10.1016/j.fertnstert.2009.07.985</a></p><p>El-Refaie et al. (2015). Vaginal progesterone for prevention of preterm labor in asymptomatic twin pregnancies with sonographic short cervix: A randomized clinical trial of efficacy and safety. <i>Archives of Gynecology and Obstetrics</i>. <a href="https://doi.org/10.1007/s00404-015-3767-1" target="_blank">https://doi.org/10.1007/s00404-015-3767-1</a></p><div><u>Note</u></div><p>The two articles marked with a * above are the only ones in which I did not identify any problems; in each of these articles all of the statistical tests are marked as either "S" (significant) or "NS" (not significant) and none of the calculations that I performed resulted in the opposite verdict for any test.</p><p><br /></p><p><span style="background-color: white; color: #222222;"><br /></span></p><p><br /></p><p><br /></p>
</div>Nick Brownhttp://www.blogger.com/profile/07481236547943428014noreply@blogger.com2tag:blogger.com,1999:blog-7890764972166411105.post-14647530312170230592021-10-02T21:18:00.155+02:002023-06-04T17:41:26.541+02:00 Applying John Carlisle's "Table 1" analysis to some claims about the treatment of Covid-19 using anti-androgens<div style="line-height: 24px;">
<p><span style="font-family: inherit;">Back in 2017, Dr. John Carlisle published<span style="background-color: white; color: #222222;"> </span><a href="http://onlinelibrary.wiley.com/doi/10.1111/anae.13938/full" style="background-color: white; color: #cc6611; text-decoration-line: none;">this article</a><span style="background-color: white; color: #222222;"> in which he described a novel method for examining the table of participants' baseline characteristics (usually Table 1) in reports of randomised trials. Table 1 is typically used to show that the randomisation "worked", which in general means that the groups do not differ on important variables more often than we would expect by chance (with the <i>p</i> value of the final comparison expected to, in effect, "mop up" any differences that do occur).</span></span></p><p><span style="background-color: white; color: #222222; font-family: inherit;"></span></p><p>[[ Update 2023-06-04 15:40 UTC: Please read my additional comments just before the "Materials" section of this post. ]]</p><div><span style="background-color: white; color: #222222; font-family: inherit;">Carlisle's insight was that it is possible for the baseline characteristics to be too similar across groups. That is, in some cases, we do not see the random variation that would be expect if the assignment to groups is truly random. For example, if you have 100 participants and 20 of them have some condition (say, diabetes), while a 10–10 split across two groups is the most likely individual outcome (about 17.6% of the time), you would expect a 7–13 </span><span style="background-color: white; color: #222222;">(or more extreme) split about 26.5% of the time. If Table 1 contains a large number of even or near-even splits, that can be a sign that the randomisation was not done as reported, because there is just not enough genuine randomness in the data.</span></div><p><span style="background-color: white; color: #222222;">We can quantify the degree of randomness that is present by looking at the <i>p</i> values for the different between-group baseline tests in Table 1. If all of the variables are truly randomised, we would expect these <i>p</i> values to be uniformly distributed with a mean of 0.50. With a sufficiently large number of variables, we would expect 10% of the <i>p</i> values to be between 0 and 0.1, a further 10% to be between 0.1 and 0.2, and so on. If we see mostly <i>p</i> values above 0.8 or 0.9, this suggests that the baseline similarities between the groups could be too good to be true. A statistical test <a href="https://en.wikipedia.org/wiki/Fisher%27s_method#Relation_to_Stouffer's_Z-score_method" target="_blank">attributed to Stouffer and Fisher</a> can be used to determine the probability of the set of <i>p</i> values that we observe being due to chance.</span></p><p>Carlisle's method has limitations, some of which I <a href="https://steamtraen.blogspot.com/2017/06/exploring-john-carlisles-bombshell.html" target="_blank">blogged about at the time</a> (see also the comments under that post, including a nice reply from John Carlisle himself) and some of which were mentioned by proper methodologists and statisticians, for example <a href="https://www.methodsman.com/blog/fraud-carlisle" target="_blank">here</a> and <a href="https://errorstatistics.com/2017/07/01/s-senn-fishing-for-fakes-with-fisher-guest-post/amp/" target="_blank">here</a>. Those limitations (principally, the possibility of non-independence of the observations being combined from Table 1, and their sometimes limited number) should be borne in mind when reading what follows, but (spoiler alert) I do not think they are sufficient to explain the issues that I report here.</p><p>In this post, I'm going to apply Carlisle's method, and the associated Stouffer-Fisher test, to two articles that have made claims of remarkably large positive effects of anti-androgenic drugs for the treatment of Covid-19:</p><div style="line-height: 24px; margin-left: 36pt; text-indent: -36pt;">Cadegiani, F. A., McCoy, J., Wambier, C. G., <span style="text-indent: -36pt;">Vaño-Galván, S., Shapiro, J., </span><span style="text-indent: -36pt;">Tosti, A., Zimerman</span><span style="text-indent: -36pt;">, R. A., & </span>Goren, A. (2021). <span style="text-indent: -36pt;">Proxalutamide significantly accelerates viral c</span><span style="text-indent: -36pt;">learance and reduces time to clinical r</span><span style="text-indent: -36pt;">emission in patients with mild to moderate </span><span style="text-indent: -36pt;">COVID-19: Results from a randomized, double-b</span><span style="text-indent: -36pt;">linded, placebo-controlled trial</span><span style="text-indent: -36pt;">. </span><i style="text-indent: -36pt;">Cureus</i><span style="text-indent: -36pt;">, </span><i style="text-indent: -36pt;">13</i><span style="text-indent: -36pt;">(2), e13492. </span><span style="text-indent: -36pt;">https://doi.org/10.7759/cureus.13492</span></div><div style="line-height: 24px; margin-left: 36pt; text-indent: -36pt;"><br /></div><div style="line-height: 24px; margin-left: 36pt; text-indent: -36pt;">Cadegiani, F. A., McCoy, J., Wambier, C. G., & Goren, A. (2021). <span style="text-indent: -36pt;">Early antiandrogen therapy with dutasteride r</span><span style="text-indent: -36pt;">educes viral shedding, inflammatory r</span><span style="text-indent: -36pt;">esponses, and time-to-remission in males with </span><span style="text-indent: -36pt;">COVID-19: A randomized, double-blind, placebo-c</span><span style="text-indent: -36pt;">ontrolled interventional trial (EAT-DUTA </span><span style="text-indent: -36pt;">AndroCoV Trial – Biochemical)</span>. <i>Cureus</i>, <i>13</i>(2), e13047. https://doi.org/10.7759/cureus.13047</div><div style="line-height: 24px; margin-left: 36pt; text-indent: -36pt;"><br /></div><div style="line-height: 24px;">I will refer to these as Article 1 and Article 2, respectively. Article 2 also seems to be closely related to this preprint, which reports results from what appear to be a superset of its participants; I will refer to this as Article 3. <span style="text-indent: -48px;">Below, I also report the results of the analyses using Carlisle's method on this preprint; h</span>owever, because of the similarity between the two samples I don't think that it would be fair to the authors to claim that the issues that I report here have been found in three, rather than two, articles.</div><div style="line-height: 24px;"><br /></div><div style="line-height: 24px; margin-left: 36pt; text-indent: -36pt;">Cadegiani, F. A., McCoy, J., Wambier, C. G., & Goren, A. (2020). 5-alpha-reductase inhibitors reduce remission time of COVID-19: Results from a randomized double blind placebo controlled interventional trial in 130 SARS-CoV-2 positive men. <i>medRxiv</i>. https://doi.org/10.1101/2020.11.16.20232512</div><div style="line-height: 24px; margin-left: 36pt; text-indent: -36pt;"><br /></div><h3 style="line-height: 24px; margin-left: 36pt; text-align: left; text-indent: -36pt;">Method</h3><div>For each article, I extracted the contents of Table 1 to a text file and, using global commands in the "vim" text editor as far as possible, converted each line into a call to a custom-written function that calculated the <i>p</i> value for that variable.</div><div><br /></div><div>For variables where the <i>p</i> value is calculated from an independent-samples <i>t</i> test, my custom function determined the maximum possible <i>t</i> statistic (and, hence, the minimum possible <i>p</i> value), by adding the maximum possible rounding error to the larger mean, subtracting the same maximum possible rounding error from the smaller mean, and subtracting the maximum possible rounding error from the standard deviation of each group (cf. Brown & Heathers, 2019, "Rounded Input Variables, Exact Test Statistics (RIVETS)", https://psyarxiv.com/ctu9z/). I believe that doing this works in the authors' favour, as the majority of the <i>p</i> values in these analyses come from contingency tables and are rather large; that is, getting the smallest possible <i>p</i> value from the <i>t</i> tests tends to increase the overall Stouffer-Fisher test <i>p</i> value.</div><div><br /></div><div>For the majority of the variables, where the <i>p</i> value is derived from an analysis of 2x2 contingency tables, my custom function applied the following rules:</div><div><ul style="line-height: 24px; text-align: left;"><li>If any cell of the table contains 0, return NULL; the variable will not be considered to have returned a <i>p</i> value).</li><li>If any cell of the table contains a number less than 5, apply Fisher's exact test. This is of course an arbitrary distinction (but it doesn't make too much difference anyway).</li><li>Otherwise, apply a chi-square test with Yates' continuity correction for 2x2 tables.</li></ul></div><div>Other analyses are possible, but this was what I decided to do a priori. I call this "Analysis 1a" below. In Analysis 1b I included tables where one or more cells are zero (also using Fisher's exact test). In Analysis 1c I excluded all variables where one or more cells had a value less then 3, so that any variable where 2 or fewer people in either condition had or did not have the attribute in question we excluded. In Analyses 2a, 2b, and 2c I applied the same rules for inclusion as in 1a, 1b, and 1c, respectively, but I used the chi-square test throughout. (This means that Analyses 1c and 2c are identical, as there are no variable in 1c for which my rules would mean that Fisher's exact test would be used.) In Analyses 3a, 3b, and 3c I used Fisher's exact test throughout.</div><div><br /></div><div>After calculating the <i>p</i> values, I replaced any that were greater than 0.98 with exactly 0.98, which avoids problems with the calculation of the Stouffer-Fisher formula in R with values of exactly 1.0 (which will occur, for example, if the number of cases in a contingency table is identical across conditions, or differs only by 1). Again, I believe that this choice works in the authors' favour. Then I calculated the overall Stouffer-Fisher test <i>p</i> value formula using the method that I described in my blog post about Carlisle's article:</div></div><ol style="line-height: 24px; text-align: left;"><li>Convert each <i>p</i> value into a <i>z</i> score.</li><li>Sum the <i>z</i> scores.</li><li>If there are <i>k</i> scores, divide the sum of the <i>z</i> scores from step 2 by the square root of <i>k</i>.</li><li>Calculate the one-tailed <i>p</i> value associated with the overall <i>z</i> score from step 3.</li></ol><p></p><div style="line-height: 24px;">For example, for Article 1 with Analysis 2c (see the table below), 26 <i>p</i> values are retained from 55 variables: (0.048, 0.431, 0.703, 0.782, 0.973, 0.298, 0.980, 0.682, 0.817, 0.897, 0.826, 0.328, 0.980, 0.980, 0.227, 0.424, 0.918, 0.884, 0.959, 0.353, 0.980, 0.511, 0.980, 0.512, 0.980, 0.980). These correspond to the z scores (-1.662, -0.175, 0.533, 0.779, 1.924, -0.531, 2.054, 0.473, 0.904, 1.265, 0.939, -0.446, 2.054, 2.054, -0.749, -0.193, 1.393, 1.198, 1.739, -0.376, 2.054, 0.028, 2.054, 0.030, 2.054, 2.054), which sum to 21.448. We divide this by the square root of 26 (i.e., 5.099) to get an overall <i>z</i> score of 4.206, which in turn gives a <i>p</i> value of 0.000013 (1.30E-05).</div><div style="line-height: 24px;"><br /></div><div style="line-height: 24px;">Note that we do not take the absolute value of each <i>z</i> score, because the sign is important. A <i>p</i> value below/above 0.5 corresponds to a negative/positive <i>z</i> score. If the positive and negative z scores cancel each other out, the overall Stouffer-Fisher <i>p</i> value will be 0.5, which is what we expect to see on average with perfect randomisation.</div><p></p><div style="line-height: 24px;"><h3 style="line-height: 24px; margin-left: 36pt; text-align: left; text-indent: -36pt;">Results</h3><div style="line-height: 24px;">Here are the <i>p</i> values that I obtained from each article using each of the analysis methods described above. Note that a value of zero corresponds to a p value below 2.2E-16, the smallest value that R can calculate (on my computer anyway). The 10 columns to the right of the overall <i>p</i> value show the deciles of the distribution of <i>p</i> values of the individual tests that make up the overall score.</div><div style="line-height: 24px;"><br /></div><div style="line-height: 24px;"><div class="separator" style="clear: both; text-align: center;"><div class="separator" style="clear: both; text-align: center;"><div class="separator" style="clear: both; text-align: center;"><div class="separator" style="clear: both; text-align: center;"><div class="separator" style="clear: both; text-align: center;"><div class="separator" style="clear: both; text-align: center;"><div class="separator" style="clear: both; text-align: center;"><div class="separator" style="clear: both; text-align: center;"><div class="separator" style="clear: both; text-align: center;"><div class="separator" style="clear: both; text-align: center;"><a href="https://lh3.googleusercontent.com/-pPo-NFTj_9Y/YVmowqdS7BI/AAAAAAAAGRg/gIgA80fc-ZcLw3rKPRRENSsxdM2ChmjWQCLcBGAsYHQ/image.png" style="margin-left: 1em; margin-right: 1em;"><img alt="" data-original-height="505" data-original-width="534" height="605" src="https://lh3.googleusercontent.com/-pPo-NFTj_9Y/YVmowqdS7BI/AAAAAAAAGRg/gIgA80fc-ZcLw3rKPRRENSsxdM2ChmjWQCLcBGAsYHQ/w640-h605/image.png" width="640" /></a></div><br /></div></div></div></div></div></div></div></div></div>It can be readily seen that in Articles 1 and 2, which are the main focus of this post, the <i>p</i> value is very small, whatever the combination of analyses and exclusions that were performed, and the distribution of <i>p</i> values from the individual comparisons is very heavily skewed towards values above .7 and especially .9. Even when I excluded a large number of comparisons because there were fewer than 3 cases or non-cases in one or more conditions (Analyses 1c, 2c, and 3c), a decision that is heavily in favour of the authors, the largest <i>p</i> value I obtained for Article 1 was 0.0000337, and for Article 2 the largest <i>p</i> value was 0.00000000819. (As mentioned above, the results for Article 3 should not be considered to be independent from those of Article 2, but they do tend to confirm that there are potentially severe problems with the randomisation of participants in both reported versions of that study.)</div><div style="line-height: 24px;"><br /></div><h3 style="text-align: left;">Conclusion</h3><p>The use of novel statistical methods should always be approached with an abundance of caution; indeed, as noted earlier, I have had my own criticisms of John Carlisle's approach in the past. (Now that I have adopted Carlisle's method in this post, I hope that someone else will take the part of the "Red Team" and perhaps show me where my application of it might be invalid!)</p><p>However, in this case, I believe that my comments from 2017 about the non-independence (and low number) of baseline measures do not apply here to anything like the same extent. In addition, I have run multiple analyses of each of the articles in question here, and tried to make whatever choice favoured the authors when the opportunity presented itself. (Of course, I could have made even more conservative choices; for example, I could have excluded all of the contingency tables where any of the cells were less than 6, or 10. But at some point, one has to accept at face value the authors' claims that their Table 1 results (a) are worth reporting and (b) demonstrate the success of their randomisation.)</p><p>Despite that, I obtained what seems to be substantial evidence that the randomisation of participants to conditions in the two studies that are the respective subjects of these articles may not have taken place exactly as reported. This adds to a growing volume of public (and, in some cases, official) critique of the Covid-19 research coming from this laboratory (e.g., in English, <a href="https://www.science.org/news/2021/07/too-good-be-true-doubts-swirl-around-trial-saw-77-reduction-covid-19-mortality" target="_blank">here</a>, <a href="https://www.swissinfo.ch/eng/reuters/the-man-behind-brazil-s-search-for-miracle-covid-19-cures/46618576" target="_blank">here</a>, <a href="https://kylesheldrick.blogspot.com/2021/08/data-from-cadegiani-et-al-contains.html" target="_blank">here</a>, <a href="https://gidmk.medium.com/is-ivermectin-for-covid-19-based-on-fraudulent-research-part-2-a4475523b4e4" target="_blank">here</a>, and <a href="https://forbetterscience.com/2021/07/21/frontiers-in-homicidal-quackery/" target="_blank">here</a>, or in Portuguese <a href="https://www.gov.br/anvisa/pt-br/assuntos/noticias-anvisa/2021/esclarecimentos-sobre-estudos-com-proxalutamida-aprovados-pela-anvisa" target="_blank">here</a>, <a href="http://conselho.saude.gov.br/ultimas-noticias-cns/2033-conep-cns-pede-que-procuradoria-geral-investigue-200-mortes-em-estudo-irregular-com-proxalutamida-para-tratar-covid-19" target="_blank">here</a>, <a href="https://www.matinaljornalismo.com.br/matinal/reportagem-matinal/proxalutamida-hospital-militar-covid-porto-alegre/?__cf_chl_captcha_tk__=pmd_R.rCGpLa8ocaXR35A_QSNBc3J2w4pJ7sparBKHqMxXo-1633212095-0-gqNtZGzNA1CjcnBszQhR" target="_blank">here</a>, <a href="https://saude.estadao.com.br/noticias/geral,proxalutamida-conselho-e-mpf-vao-apurar-teste-sem-autorizacao-com-pacientes-de-covid,70003821117" target="_blank">here</a>, <a href="https://oglobo.globo.com/politica/juiz-censura-reportagem-do-globo-sobre-remedio-sem-eficacia-comprovada-25165048" target="_blank">here</a>, <a href="https://www.bbc.com/portuguese/brasil-57897387" target="_blank">here</a>, and <a href="https://www.gov.br/anvisa/pt-br/assuntos/noticias-anvisa/2021/esclarecimentos-sobre-estudos-com-proxalutamida-aprovados-pela-anvisa" target="_blank">here</a>).</p><p>[[ Update 2023-06-04 15:40 UTC: Two alert comments mentioned the preprint by Daniel Tausk which points out a limitation of the use of Carlisle's method with dichotomous variables; my comment that "The use of novel statistical methods should always be approached with an abundance of caution" was apparently prescient. Tausk's analysis suggests that the <i>p</i> values that I calculated above may have been rather too small; he suggests 0.017 for the first and 0.0024 for the second. These might not, on their own, constitute strong enough evidence to launch an inquiry into these studies, but it seems to me that they are still of interest in view of the other critiques of the work from this laboratory. I will let readers make up their own minds. ]]</p><h3 style="text-align: left;">Materials</h3><p>My R code is available <a href="https://gist.github.com/sTeamTraen/83d54a97c6a2ff4be2ef5a23c56bd7fc" target="_blank">here</a>. The three referenced articles are published either under open access conditions or as a preprint, so I hope they will be easy for readers to download when checking my working.</p><p>[[ Update 2021-10-03 12:56 UTC: Added the columns of <i>p</i> value deciles to the table of results. ]]</p>
</div>
Nick Brownhttp://www.blogger.com/profile/07481236547943428014noreply@blogger.com6tag:blogger.com,1999:blog-7890764972166411105.post-1587940941082439332021-09-14T23:44:00.006+02:002021-09-15T01:36:18.712+02:00Some problems in a study of Nitazoxanide as a treatment for Covid-19<div style="line-height: 24px;">
<p><span style="font-size: small;">Nitazoxanide (NTZ) is an antiparasitic medication that has been proposed as a possible treatment for Covid-19. (The currently most widely-touted repurposed drug for Covid-19, Ivermectin, is also an antiparasitic.) The effects of NTZ were recently examined in this paper:<br /></span></p>
<div style="line-height: 24px; margin-left: 36pt; text-indent: -36pt;"><span style="font-size: small;"> Blum, V. F., Cimerman, S., Hunter, J. R., Tierno, P., Lacerda, A., Soeiro, A., Cardoso, F., Bellei, N. C., Maricato, J., Mantovani, N., Vassao, M., Dias, D., Galinskas, J., <br />Janini, L. M. R., Santos-Oliveira, J. R., Da-Cruz, A. M., & Diaz, R. S. (2021). Nitazoxanide superiority to placebo to treat moderate COVID-19: A pilot prove [<i>sic</i>] of concept randomized double-blind clinical trial. <i>EClinicalMedicine</i>, <i>37</i>, 100981.<i> </i>https://doi.org/10.1016/j.eclinm.2021.100981</span></div>
<p><span style="font-size: small;">The authors reported (p. 9) that "The study protocol and all data files from the study results and<br />programs to process the files are deposited in a publicly available repository on the GitLab system for this clinical trial", with a <a href="https://gitlab.com/jameshunterbr/ntz-clinical-trial" target="_blank">link</a>. To me, this implies that the data and code should have been available on the date when the article was published, i.e., June 27, 2021. However, it appears that no files were actually uploaded to that repository until August 29, 2021, which perhaps not uncoincidentally was shortly after my colleague Kyle Sheldrick had asked the authors to post their data.</span></p><p><span style="font-size: small;">The GitLab repository contains the data divided into two sets, one with the participant characteristics ("Demographic_NTZ_Trial_Data") and one with the observations from each day of the trial ("Clinical_Laboratory_NTZ_Trial_Data"). Each of these is ostensibly provided in two formats, as both an Excel sheet (with the extension ".xlsx") and as the binary dump of an R data structure (".rds"). For "Demographic_NTZ_Trial_Data" the Excel and R versions appear to correctly contain the same data. Unfortunately, the two R binary files are byte-for-byte identical to each other, suggesting that the wrong file has been uploaded for “Clinical_Laboratory_NTZ_Trial_Data"; hence, I relied on the Excel version of both files.</span></p><p><span style="font-size: small;">Also perhaps of note is that the metadata (“File/Properties/Summary” and “File/Properties/Statistics”) of the Excel files reveal that they were created by “Apache POI” on August 29, 2021 at 19:13:34 UTC (“Demographic_NTZ_Trial_Data”) and 20:27:18 UTC (“Clinical_Laboratory_NTZ_Trial_Data”). The file creator name “Apache POI” suggests that these files were created by the R function <span style="font-family: courier;">write.xlsx()</span>, which leaves this “signature” from the Java library that it uses to create Excel files. This in turn suggests that there ought to be a more fundamental version of the data files somewhere, perhaps in CSV format. It would be nice if the authors could share this “rawer” form of the data with us, rather than the output of an R program after some unknown amount of preprocessing.</span></p><p><span style="font-size: small;">The GitLab repository contains only the above-mentioned data files, plus a "data dictionary" containing a one-sentence description of each variable, and a brief "Readme"file. The authors have therefore apparently still not fulfilled their promise to provide the study protocols or their analysis code.<br /></span></p><p><span style="font-size: small;">All of the following analyses are based on the two Excel files, in which several of the variables are labeled and/or coded in Portuguese; I believe that I have appropriately interpreted these labels, so that for example the variable “sexo” with the value “Masculino” corresponds to a male patient (reported by the authors in English as "Masculine"). <br /></span></p><p><span style="font-size: small;"><u>The table of results</u><br /></span></p><p><span style="font-size: small;">The basis of the majority of the authors' claims regarding the superiority of NTZ is the main table of results (Table 1, p. 5). Unfortunately, this table seems to contain a large number of errors and inconsistencies with the supplied data files. Some of these are more serious than others; I will describe them here from the top to the bottom of the table, rather than attempting to order them by some subjective measure of severity.</span></p><ol style="text-align: left;"><li><span style="font-size: small;">Table 1 reports that there were 7 men ("Gender: Masculine") and 18 women ("Gender: Feminine") in the NTZ group, versus 8 men and 17 women in the placebo group. However, in the data these numbers show the opposite result, not by group but by gender: In the NTZ group, 18 patient records have the value "Masculino" and 7 have "Feminino" for the variable named "sexo", whereas in the placebo group, 17 records have "Masculino" and 8 contain "Feminino". That is, the data say that two-thirds of the patients were men and the table of results says that two-thirds were women. The article text does not discuss the sex or gender of the patients as a group, but it does not seem unreasonable to me that one of the 17 authors of the article might have been expected to notice that they reported the proportions of men and women entirely incorrectly.</span></li><li><span style="font-size: small;">The variables "Race" and "Age group" list counts of the patients in each of these categories, and express these as percentages. However, these percentages are not of the number of people in each experimental condition, but of the total number of patients. For example, the first value for Race is White: 21 (42%), but 21 is 84% of the 25 people in the NTZ condition.</span></li><li><span style="font-size: small;">The calculation of "RT-PCR Difference Day 1 - 21" does not seem to be correct. In the NTZ group, there are no records with measurements of PCR viral load at day 21, so there is no way to calculate the mean or SD of the difference between the two timepoints. In the placebo group there are four such records; I calculate the mean difference between the day-1 and day-21 viral load for these cases to be 0.79 with an SD of 3.45.<br /></span></li><li><span style="font-size: small;">The calculation of mean and SD values for "Removed from Supplemental O2" does not seem to be possible, as there is no variable corresponding to this in the data file. (The authors provided a data dictionary in their GitLab repository, which does not apparently mention any variable corresponding to supplemental oxygen.) Perhaps this variable is based on the clinical condition score changing from 3 to a lower value on a subsequent day, but this only occurs in three cases in the data file (patient 101008, day 7, change from “3” (supplemental O₂) to “2” (hospitalised); patient 103035, day 7, change from “3” to “2”; patient 104001, day 14, change from “3” to “2”), and all three of these participants are in the placebo group. It does not need to be possible to calculate a meaningful mean or standard deviation for these values even for the placebo group, let alone for the NTZ group where there are zero cases.)<br /></span></li><li><span style="font-size: small;">The <i>p</i> values for "Viral Load at Day 1" and "O2 Saturation at Day 1" in the table are both reported as 0.984. Using R's <span style="font-family: courier;">wilcox.test()</span> function, I calculated these values as 0.168 and 0.381 respectively. The authors also mention (p. 4) that they used the Kruskal-Wallis test in some cases; with R's <span style="font-family: courier;">kruskal.test()</span> function, I obtained <i>p</i> values of 0.167 and 0.380 respectively.</span></li><li><span style="font-size: small;">Many of the numbers for the patient clinical condition scores in the lines "Patients Hospitalized" (6/8), "Oxygen supplementation" (8/8), and "Death" (3/6) are different between the data file and the article. Interestingly, all eight numbers for "Invasive mechanical ventilation" seem to be correct. See my annotated version of the table for more information.<br /></span></li><li><span style="font-size: small;">It is not clear how the <i>p</i> values for the patient clinical conditions were calculated. This could have been a simple two-count chi-square test (e.g., in R, <span style="font-family: courier;">chisq.test(c(2, 1))</span> gives a p value of 0.564, which appears a couple of times in the table), or it could perhaps have been a 2 x 2 contingency table, with the denominator being the number of patients in the other conditions. In any case, these p values do not appear to match either the calculated or reported frequencies.</span></li><li><span style="font-size: small;">The mean and SD values for Lymphocytes at day 1 in the placebo condition and at day 7 in the NTZ condition have been exchanged. That is, according to the data file, the mean lymphocyte value on day 1 in the placebo condition was 1162.36, and the mean value on day 7 in the NTZ condition was 944.88 (a decrease on the day 1 value, rather than the increase that was reported in Table 1).</span></li><li><span style="font-size: small;">For the measures "D-Dimer", "US- C Reactive Protein", "TNF-<span class="ILfuVd"><span class="hgKElc">α</span></span>", "IL-6", and "IL-8", there are several problems. First, the table reports values at day 10, but there are no records of any kind with this day in the data files. Second, for "TNF-<span class="ILfuVd"><span class="hgKElc">α</span></span>", "IL-6", and "IL-8", the table reports mean values on day 1, but none of the patient records for day 1 have any values other than "#N/A" for any of these three variables. Third, none of the remaining mean or SD values come close to the values that I calculated from the data file.<br /></span></li><li><span style="font-size: small;">A column in Table 1 shows the ratio between the means or counts in the two groups and describes this as the "Rate ratio". However, there are some rather strange numbers in this column. In many cases the decimal point (or comma) is missing, so that for example the ratio of 7 "Mixed" (race) patients in the placebo group to 3 in the NTZ group is reported as 2333. In other cases, even with a decimal separator assumed, the numbers make little sense. For example, the "ratios" between the numbers of patients on intensive mechanical ventilation at days 4 (3 placebo, 1 NTZ) and 7 (4 placebo, 1 NTZ) are reported as 3333 and 4444. One wonders what kind of analysis code generated these numbers.</span></li><li><span style="font-size: small;">The "Rate ratio" values under "Day 21 Difference" for the measures "D-Dimer" and "US- C Reactive Protein" have been exchanged in the table, relative to what is in the data file.</span></li></ol><p><span style="font-size: small;"><u>The regression lines</u><br /></span></p><p><span style="font-size: small;">Another set of problems is apparent in Blum et al.'s Figure 2, which the authors claim (p. 6) “shows that the viral load for the NTZ arm of the study dropped slightly faster than the Placebo arm over the 21 days of the study (slope of 1.19 for NTZ against 1.08 for the placebo)”.</span></p><p><span style="font-size: small;">First, the variable corresponding to the PCR tests in the data files is the PCR Ct (cycle threshold, cf. the data file “ntz_data_dict_2808.html” in the authors’ GitLab repository), and a higher cycle threshold corresponds to a lower viral load. That is, both of these slopes ostensibly show the viral load increasing over the course of the study, with (on the authors’ account) the NTZ group getting worse to a greater degree than the placebo group. This apparent failure to understand the meaning of their own data might be considered quite concerning.<br /></span></p><p><span style="font-size: small;"><img alt="" height="561" src="data:image/png;base64,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" width="640" /></span></p><p><span style="font-size: small;">Second, it can be readily seen that the great majority of the visible data points here are above the two lines, with a considerable number being both above and far off to the right. A moment's thought suggests a solution, which is that very large numbers of data points with a viral load value of zero have been drawn on top of each other. However, in the data file, the lowest PCR value is 17. What seems to have happened is that, as well as 115 actual PCR values, the authors have coded 136 NA values from days 4 through 21—corresponding either to missing tests or patient dropout—as a value of zero. While some of these NA values might correspond to negative PCR tests, others must be missing data as they occur later than the date at which the patient’s condition code changed to “5” (death). When I omitted these NA values, I obtained the following plot:</span></p><p><span style="font-size: small;"><img alt="" height="546" src="data:image/png;base64,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" width="640" /></span></p><p><span style="font-size: small;">On its own terms (but see my next point), this plot shows—rather more plausibly than the published Figure 2—that the PCR cycle threshold increased (and, hence, the viral load decreased) throughout the study, with a bigger decrease in the NTZ group.</span></p><p><span style="font-size: small;">The third problem with Figure 2 is that attempting to interpret the slope of a linear regression line (one of the assumptions of which is that each data point is independent of the others) in this way for time series data with multiple repeated measures on the same patients is completely invalid. No conclusions can be drawn on the basis of this model. The authors ought to have used some kind of generalized linear model, accounting for correlation between observations on the same patients and also the censoring effects of patient death or discharge.<br /></span></p><p><span style="font-size: small;"><u>Repeated-measures "ANOVA"</u> <br /></span></p><p><span style="font-size: small;">Figure 3 of Blum et al.'s article shows the results of analyses that they describe as "Two-way-ANOVA". It is not clear exactly what analyses were performed here. On p. 4 the authors say "Statistical analyses included one-way analyses of variances using the Kruskal-Wallis non-parametric ranks test. Additional statistical questions were addressed with the Wilcoxon Rank Sum Test for numerical variables and Chi-Squared tests for categorical variables." Since the Wilcoxon rank-sum test is sufficient to conduct all of the analyses in Table 2, one might imagine that "Two-way-ANOVA" could include the Kruskal-Wallis test, although that is not appropriate for repeated measures.</span></p><p><span style="font-size: small;">In any case, some of the numbers reported in Figure 3 are rather strange. For example, in panels B and D, a significance star is assigned to <i>p</i> values less than 0.5, which does not seem to be a very high bar to meet. Furthermore, the caption for Figure 3 includes the sentence "The MFI of CD38+CD4+ T cells (3A), HLA-DR.+CD4+ T cells (3B), ... are expressed as <i>the median ± standard deviation</i>" (emphasis added), which is a rather unusual combination.</span></p><p><span style="font-size: small;">I was unable to reproduce Figure 3 for two reasons. First, there are no data in the files for day 10 or with any label that corresponds to the Y axis of panels E and F. Second, the variables that do exist in the data file with names that resemble the labels of the Y axis of panels A through D (i.e., "cd38_cd4", "cd38_cd8", "hla_dr_cd4", and "hla_dr_cd8") have mean values in the range 0.5–2.0, whereas the Y-axes in Figure 3 are two or more orders of magnitude larger; it is not clear to me if I haven't understood what these variables mean, or if they are simply missing from the data files.</span></p><p><span style="font-size: small;"><span style="font-family: inherit;"><b id="docs-internal-guid-c1e1e581-7fff-ef2b-1d94-1d37ae9eed4d" style="font-weight: normal;"></b></span></span></p><p dir="ltr" style="line-height: 24px; margin-bottom: 12pt; margin-top: 12pt;"><span style="font-size: small;"><b id="docs-internal-guid-c1e1e581-7fff-ef2b-1d94-1d37ae9eed4d" style="font-weight: normal;"><span style="background-color: transparent; color: black; font-style: normal; font-variant: normal; font-weight: 400; text-decoration-skip-ink: none; text-decoration: underline; vertical-align: baseline; white-space: pre-wrap;">
Analyses of individual patients</span></b></span></p><p dir="ltr" style="line-height: 24px; margin-bottom: 12pt; margin-top: 12pt;"><span style="font-size: small;"><b id="docs-internal-guid-c1e1e581-7fff-ef2b-1d94-1d37ae9eed4d" style="font-weight: normal;"><span style="background-color: transparent; color: black; font-style: normal; font-variant: normal; font-weight: 400; text-decoration: none; vertical-align: baseline; white-space: pre-wrap;">The authors provided detailed information in the text of their article about the patients who died in the study. However, in some cases it is difficult to identify these individuals in the data, as there are a considerable number of discrepancies between the numbers reported in the text and those in the data file..</span></b></span></p><ol style="margin-bottom: 0px; margin-top: 0px; padding-inline-start: 48px;"><li aria-level="1" dir="ltr" style="background-color: transparent; color: black; font-style: normal; font-variant: normal; font-weight: 400; list-style-type: decimal; text-decoration: none; vertical-align: baseline; white-space: pre;"><p dir="ltr" role="presentation" style="line-height: 24px; margin-bottom: 12pt; margin-top: 12pt;"><span style="font-size: small;"><b id="docs-internal-guid-c1e1e581-7fff-ef2b-1d94-1d37ae9eed4d" style="font-weight: normal;"><span style="background-color: transparent; color: black; font-style: normal; font-variant: normal; font-weight: 400; text-decoration: none; vertical-align: baseline; white-space: pre-wrap;">Male deaths in the NTZ arm</span></b></span></p></li></ol><p dir="ltr" style="line-height: 24px; margin-bottom: 12pt; margin-left: 36pt; margin-top: 12pt;"><span style="font-size: small;"><b id="docs-internal-guid-c1e1e581-7fff-ef2b-1d94-1d37ae9eed4d" style="font-weight: normal;"><span style="background-color: transparent; color: black; font-style: normal; font-variant: normal; font-weight: 400; text-decoration: none; vertical-align: baseline; white-space: pre-wrap;">“Two patients died in the NTZ arm of the study. One of these, patient P3 was a 67-year-old male with systemic arterial hypertension (SAH) and a BMI (body mass index) of 25. … Another patient in the NTZ group, P6, was a 63-year-old male with type 2 diabetes mellitus and dyslipidemia, with a BMI of 31.”</span></b></span></p><p dir="ltr" style="line-height: 24px; margin-bottom: 12pt; margin-left: 36pt; margin-top: 12pt;"><span style="font-size: small;"><b id="docs-internal-guid-c1e1e581-7fff-ef2b-1d94-1d37ae9eed4d" style="font-weight: normal;"><span style="background-color: transparent; color: black; font-style: normal; font-variant: normal; font-weight: 400; text-decoration: none; vertical-align: baseline; white-space: pre-wrap;">The data contain two records of patients dying in the NTZ arm. One (numerical ID 101004) was aged 77 with a BMI of 24.9, while the other (numerical ID 106016) was aged 63 with a BMI of 30.74. We might tentatively assume that these individuals correspond to the two mentioned in the text. However, 101004, with “systemic arterial hypertension” had baseline systolic blood pressure of 120, whereas 106016, with no mention of hypertension, had a systolic BP of 150. The reported C-reactive protein for the latter also does not match the data. Furthermore, for both of these patients (and indeed for all of the patients who died), IL-6 values are reported, although the IL-6 numbers are NAs in all cases (except for one reading for patient 103004; see next paragraph).</span></b></span></p><ol start="2" style="margin-bottom: 0px; margin-top: 0px; padding-inline-start: 48px;"><li aria-level="1" dir="ltr" style="background-color: transparent; color: black; font-style: normal; font-variant: normal; font-weight: 400; list-style-type: decimal; text-decoration: none; vertical-align: baseline; white-space: pre;"><p dir="ltr" role="presentation" style="line-height: 24px; margin-bottom: 12pt; margin-top: 12pt;"><span style="font-size: small;"><b id="docs-internal-guid-c1e1e581-7fff-ef2b-1d94-1d37ae9eed4d" style="font-weight: normal;"><span style="background-color: transparent; color: black; font-style: normal; font-variant: normal; font-weight: 400; text-decoration: none; vertical-align: baseline; white-space: pre-wrap;">Female deaths in the placebo arm</span></b></span></p></li></ol><p dir="ltr" style="line-height: 24px; margin-bottom: 12pt; margin-left: 36pt; margin-top: 12pt;"><span style="font-size: small;"><b id="docs-internal-guid-c1e1e581-7fff-ef2b-1d94-1d37ae9eed4d" style="font-weight: normal;"><span style="background-color: transparent; color: black; font-style: normal; font-variant: normal; font-weight: 400; text-decoration: none; vertical-align: baseline; white-space: pre-wrap;">“P8 was an 88-year-old female with SAH, BMI of 24 … and IL-6 of 4872 MFI”. There is one 88-year-old female patient in the dataset (103004); her BMI is recorded as 28.8. For this patient, unlike all the others who died, there is one IL-6 reading, of 225.5 on day 21. That is the date on which she is reported as having died in the text of the article, although the data file “Demographic_NTZ_Trial_Data” states that she left the trial early (due to death) on day 7.</span></b></span></p><p dir="ltr" style="line-height: 24px; margin-bottom: 12pt; margin-left: 36pt; margin-top: 12pt;"><span style="font-size: small;"><b id="docs-internal-guid-c1e1e581-7fff-ef2b-1d94-1d37ae9eed4d" style="font-weight: normal;"><span style="background-color: transparent; color: black; font-style: normal; font-variant: normal; font-weight: 400; text-decoration: none; vertical-align: baseline; white-space: pre-wrap;">“Patient P14 was a 65-year-old female with previously not welldefined cardiomyopathy, BMI of 25. … and IL-6 of 4658 MFI. She died on D4...” There are no 65-year-old females in the data, whatever their outcome. By elimination (see next paragraph) it appears that P14 was in fact the patient with numerical ID 106010, aged 76 who according to the data file “Demographic_NTZ_Trial_Data” died on day 6. Again, no IL-6 data were reported for this patient.</span></b></span></p><p dir="ltr" style="line-height: 24px; margin-bottom: 12pt; margin-left: 36pt; margin-top: 12pt;"><span style="font-size: small;"><b id="docs-internal-guid-c1e1e581-7fff-ef2b-1d94-1d37ae9eed4d" style="font-weight: normal;"><span style="background-color: transparent; color: black; font-style: normal; font-variant: normal; font-weight: 400; text-decoration: none; vertical-align: baseline; white-space: pre-wrap;">“The last patient in the placebo group to die was P43, a 73-year-old female with obesity (BMI of 32). Baseline laboratory evaluation revealed a total lymphocyte count of 930 (total leucocytes13,480), D-dimer of 0.82 mcg/mL, US-RCP of 187.59mcg/L, and IL-6 of 417 MFI. She died on D12...” This appears to be patient 106038, who according to the data file was aged 72 (rather than 73). Again, no IL-6 data were reported for this patient.</span></b></span></p><ol start="3" style="margin-bottom: 0px; margin-top: 0px; padding-inline-start: 48px;"><li aria-level="1" dir="ltr" style="background-color: transparent; color: black; font-style: normal; font-variant: normal; font-weight: 400; list-style-type: decimal; text-decoration: none; vertical-align: baseline; white-space: pre;"><p dir="ltr" role="presentation" style="line-height: 24px; margin-bottom: 12pt; margin-top: 12pt;"><span style="font-size: small;"><b id="docs-internal-guid-c1e1e581-7fff-ef2b-1d94-1d37ae9eed4d" style="font-weight: normal;"><span style="background-color: transparent; color: black; font-style: normal; font-variant: normal; font-weight: 400; text-decoration: none; vertical-align: baseline; white-space: pre-wrap;">Male deaths in the placebo arm</span></b></span></p></li></ol><p dir="ltr" style="line-height: 24px; margin-bottom: 12pt; margin-left: 36pt; margin-top: 12pt;"><span style="font-size: small;"><b id="docs-internal-guid-c1e1e581-7fff-ef2b-1d94-1d37ae9eed4d" style="font-weight: normal;"><span style="background-color: transparent; color: black; font-style: normal; font-variant: normal; font-weight: 400; text-decoration: none; vertical-align: baseline; white-space: pre-wrap;">Three deaths among male participants were reported in the placebo arm. Apart from small discrepancies in their ages in two cases (P24, 106012, reported age 55, data file age 55; P27, 106014, reported age 71, data file age 70; P17, 106028, reported age 78, data file age 79), the only obvious discrepancy for these three cases is the reporting of IL-6 data in the text when these values are not present in the data.</span></b></span></p><p></p><p><span style="font-size: small;"><span style="font-family: inherit;"><u>Conclusion</u></span> <br /></span></p><p><span style="font-size: small;">Blum et al.'s article contains a large number of both major and minor statistical and reporting errors, some of which are apparent even without the data set. Whether or not these issues are too great to make the article salvageable is a matter for the authors and the journal’s editorial team.<br /></span></p><p><span style="font-size: small;">I have made my analysis code, together with full-size versions of the images, available at https://osf.io/uz875/</span></p><p><span style="font-size: small;">Thanks to Kyle Sheldrick and Gideon Meyerowitz-Katz for bringing the paper to my attention, and being the first to notice the problem with the NA values being interpreted as zeroes in Figure 2.<br /></span></p><p><span style="font-size: small;"><br /></span></p>
</div><span style="font-size: small;">
</span>Nick Brownhttp://www.blogger.com/profile/18266307287741345798noreply@blogger.com3tag:blogger.com,1999:blog-7890764972166411105.post-32177286535638561972021-07-15T22:11:00.012+02:002022-04-04T17:56:07.435+02:00Some problems in the dataset of a large study of Ivermectin for the treatment of Covid-19<div class="MsoNormal" style="line-height: 24px;"><div class="MsoNormal" style="line-height: 24px;">This post appears at the same time as <a href="https://grftr.news/why-was-a-major-study-on-ivermectin-for-covid-19-just-retracted" target="_blank">this piece at grftr.news</a> by <a href="https://twitter.com/JackMLawrence" target="_blank">Jack Lawrence</a>. Jack contacted me to ask if I could help him look at a number of issues with a prominent study of Ivermectin for the treatment of Covid-19. My speciality is <strike>forensic numerical data analysis</strike> finding errors in numbers, so I concentrated on that and suggested some other names to Jack to help him look at things like the study design, methods, and reporting. </div><div class="MsoNormal" style="line-height: 24px;"><br /></div><div class="MsoNormal" style="line-height: 24px;">Hence, this post is almost entirely about the problems with the data from this study; Jack's piece covers other topics, such as the plagiarised text, the clinical trial "pre-registration" that was <a href="https://clinicaltrials.gov/ct2/show/NCT04668469">made after the first version of the results of the study was published</a>, and many other problems. <a href="https://twitter.com/GidMK" target="_blank">Gideon Meyerowitz-Katz</a> has <a href="https://gidmk.medium.com/is-ivermectin-for-covid-19-based-on-fraudulent-research-5cc079278602">a piece over at Medium</a> that discusses the implications of this study for the whole Ivermectin-for-Covid literature, and <a href="https://twitter.com/MelissaLDavey" target="_blank">Melissa Davey</a> is covering the story <a href="https://www.theguardian.com/science/2021/jul/16/huge-study-supporting-ivermectin-as-covid-treatment-withdrawn-over-ethical-concerns" target="_blank">in the Guardian</a> today.</div><div class="MsoNormal" style="line-height: 24px;"><br /></div><div class="MsoNormal" style="line-height: 24px;">Here is the article reference. In fact, it's not in a peer-reviewed journal. It's just a preprint—in spite of which it has already acquired 30 citations in just over six months, <a href="https://scholar.google.es/scholar?q=Efficacy+and+Safety+of+Ivermectin+for+Treatment+and+prophylaxis+of+COVID-19+Pandemic&hl=en&as_sdt=0&as_vis=1&oi=scholart" target="_blank">according to Google Scholar</a>, and—as reported by Jack—has also become a major component of the weight of evidence for the efficacy of Ivermectin in several meta-analyses (see Gideon's Medium piece, linked above, for more on this).</div><div class="MsoNormal" style="line-height: 24px;"><br /></div><div class="MsoNormal" style="line-height: 24px; margin-left: 36pt; text-indent: -36pt;"><span style="mso-no-proof: yes;">E</span>lgazzar, A., Eltaweel, A., Youssef, S. A., Hany, B., Hafez, M., & Moussa, H. (2020). Efficacy and safety of Ivermectin for treatment and prophylaxis of COVID-19 pandemic. <i>Research Square</i>, <i>100956</i>. <a href="https://doi.org/10.21203/rs.3.rs-100956/v3">https://doi.org/10.21203/rs.3.rs-100956/v3</a></div><div class="MsoNormal" style="line-height: 24px;"><br /></div><div class="MsoNormal" style="line-height: 24px;">The preprint is currently in its third revision(*). You can download it, plus the two previous revisions if you want to compare those, from the <a href="https://www.researchsquare.com/article/rs-100956/" target="_blank">preprint hosting service Research Square</a>. (I use <a href="http://draftable.com">draftable.com</a> to compare PDFs.)</div><div class="MsoNormal" style="line-height: 24px;"><br /></div><div class="MsoNormal" style="line-height: 24px;">The authors have, well, "sort of" made their data available. To quote from the preprint (p. 6): "The study data master sheet are [<i>sic</i>] available on reasonable request from the corresponding auther [<i>sic</i>] from the following link. <a href="https://filetransfer.io/data-package/qGiU0mw6#link">https://filetransfer.io/data-package/qGiU0mw6#link</a>". It is tempting to imagine that one might be able to download the data file directly from that link; however, when you attempt to do that, the site says that you have to create a premium account ($9 per month), and after you have done that and downloaded the file, it turns out to be password-protected. This suggests that the authors did not want anyone to be able to read it without their approval, which is not quite in the spirit of open science. (It is, however, not incompatible with <a href="https://www.researchsquare.com/legal/editorial" target="_blank">Research Square's rather feeble data sharing policy</a>.)</div><div class="MsoNormal" style="line-height: 24px;"><br /></div><div class="MsoNormal" style="line-height: 24px;">Fortunately, Jack Lawrence did a lot of work here. Not only did he pay for a premium account at filetransfer.io, but he also guessed the password of the file, which turned out to be <a href="https://www.youtube.com/watch?v=dwhOpmdtDuk" target="_blank">1234</a>. I have never met Jack Lawrence in person, though, so as part of my due diligence for this blog post, I also paid $9 plus VAT for a one-month subscription to filetransfer.io, and downloaded the file for myself. To save you, dear reader, from having to go through that process, I have made an unlocked copy of the file available <a href="http://nickbrown.fr/blog/elgazzar" target="_blank">here</a>. It is perhaps interesting to note that, judging by the filename, the authors were apparently still editing the "study data master sheet" on 12 December 2020, when they had already posted essentially all of their results in two earlier versions of their preprint by November 16.</div><div class="MsoNormal" style="line-height: 24px;"><br /></div><h3 style="line-height: 24px; text-align: left;">Formatting problems</h3><h4 style="line-height: 24px; text-align: left;"><div class="MsoNormal" style="line-height: 24px;"><span style="font-weight: normal;">The data file is in Microsoft Excel (.XLSX) format, although the authors reported performing their analyses in SPSS 21. In the Excel metadata (File/Properties/Statistics), the creation date of the file is "16 September 2006 02:00:00", which suggests that the authors started with an older file and cleared all the cells before entering their data. Clearing the cells in this way does not remove cell formatting, which might explain one or two of the stranger cell formats that one sees when opening the file in Excel (e.g., in cell K5 the number 6.3 is formatted as a day and appears as "06-Jan", while B222 and F225:Z225 are in a different font); however, the formatting problems go a lot further than that.</span></div><div class="MsoNormal" style="line-height: 24px;"><br /></div><div class="MsoNormal" style="line-height: 24px;">Numbers containing non-numeric characters</div></h4><div class="MsoNormal" style="line-height: 24px;">Several cells that represent numbers in the Excel file appear to have been entered by someone more used to a manual typewriter than a computer. Specifically, cells K17, L318, L354, L366, L380, M38, M101, M396:M402, S272, S278, S280, S396, and S398 contain one or more occurrences of the lowercase letter "o" instead of the digit "0". As a result, these cells are text strings, rather than numbers, and any numerical calculations based on them will fail.</div><div class="MsoNormal" style="line-height: 24px;"><br /></div><div class="MsoNormal" style="line-height: 24px;">Because these cells contain strings, their values are left-aligned (the default for strings in Excel), whereas the numbers in the same column are right-aligned. In many cases it seems that the creator of the data file has attempted to remedy this visual infelicity by left-padding the non-numeric string with spaces. For example, although the value "1.o" [<i>sic</i>] in cell L318 has no padding, the same value in cells L354, L366, and L380 has been padded on the left with 33, 34, and 32 space characters, respectively.</div><div class="MsoNormal" style="line-height: 24px;"><br />Relatedly, the percentages in cells M89, M94, M128, S232, S243, S245, S250, S261, S262, S274, and S279 contain a comma as a decimal separator, instead of a dot, and so again are treated as text strings rather than numbers. These cells with commas are padded on the left with between 12 and 16 leading space characters in column S, although there is no padding in column M.</div><div class="MsoNormal" style="line-height: 24px;"><br /></div><h4 style="line-height: 24px; text-align: left;">Confusion around date formats</h4><div class="MsoNormal" style="line-height: 24px;">Columns W and X of the Excel file contain dates with the captions "<span style="font-family: courier;">symptoms date&+ ve PCR</span>" and "<span style="font-family: courier;">recovery date & -ve PCR</span>". It seems that these dates are performing multiple duties, since there is no obvious reason why a patient's date of first showing symptoms of Covid-19 should be identical to the date on which they first tested positive, or why the date of their (first?) negative PCR test should correspond with their doctor's (?) certification that they have recovered. These dates seem to have also been used to calculate the length of time during which people were hospitalised (column Y), although again, one would generally expect the dates of a participant's hospital admission to be somewhat decoupled from their dates of first symptoms and/or PCR tests. I find it very surprising that there are not more dates recorded for each patient, to account for the various milestones that are of importance in the progression and treatment of Covid-19.</div><div class="MsoNormal" style="line-height: 24px;"><br /></div><div class="MsoNormal" style="line-height: 24px;">However, it is not only the meaning of these dates that is confusing. Their format is, too. The only dates that are actually formatted as Excel dates (i.e., with an underlying number representing the count of days since December 30, 1899) are those where both the day and month are less than 13. My working hypothesis is that the creator of the file either typed in the dates by hand, or pasted them from a text file, in dd/mm/yyyy format, but that Excel was in "US date mode" at the time. Thus, the only dates that were converted to the underlying numeric date format were those that were interpretable as mm/dd/yyyy (i.e., those with a dd/mm/yyyy "day" less than 13; I assume that there were no errors with a dd/mm/yyyy "month" greater than 31).</div><div class="MsoNormal" style="line-height: 24px;"><br /></div><div class="MsoNormal" style="line-height: 24px;">As with the numbers that contain non-numeric characters (see previous section), it seems that padding with spaces has been added manually in an attempt to align the "string" dates with the "correct" (numeric) dates. In column W, there are 176 numeric dates, 115 "string" dates with no padding, and 109 "string" dates with left padding of between 19 and 27 spaces; three of these also have padding on the right, of 3, 6, and 33 spaces. In column X, after removing the text "died in ICU" from 21 dates and "died bin ICU" [<i>sic</i>] from another, there are 94 numeric dates, 305 "string" dates with left padding of between 1 and 28 spaces (three of which also have padding on the right, of 1 and 8 spaces), and one "string" date with left padding of a backquote character (`) and 22 spaces (cell X259). In 6 cases (cells X8, X15, X23, X62, X75, and X402), the discharge date in column X also includes the number of days spent in hospital, which should be in column Y; it appears that whoever was inputting the data may have thought that using spaces to move the cursor over to the right of the cell boundary between the two columns was equivalent to using the Tab key to move to the next cell. </div><div class="MsoNormal" style="line-height: 24px;"><br />Several of the "string" dates are incorrectly formatted internally, e.g., "1l6l2020" (cell W110, with lowercase "L" as the separator), "06/82020" (cell W208), "31/7//2020" (cell X223), and "6/8/20/20" (cell X230). The date in cell X155 ("31/06/2020") is ostensibly formatted correctly, but implies that the patient was discharged on the non-existent date of 31 June 2020.</div><div class="MsoNormal" style="line-height: 24px;"><br /></div><div class="MsoNormal" style="line-height: 24px;">In summary, it is impossible for the dates in the file to have been used to calculate any sort of elapsed time in SPSS. Indeed, it seems that this calculation was done by hand, with the results being reported in column Y (with the addition of the text " days"), and with different "<a href="https://en.wikipedia.org/wiki/Off-by-one_error" target="_blank">fencepost</a>" rules typically being applied for each group. For example, in groups I and III the number of days in column Y is usually one more than the difference between the dates in columns W and X (i.e., both the start and end date are counted), whereas for groups II and (especially) IV the number of days in column Y is typically equal to the difference between the dates in columns W and X. (See also "Table 6", below, for a brief discussion of the apparent confusion between columns W and X on the one hand, and column Y on the other.)</div><div class="MsoNormal" style="line-height: 24px;"><br /></div><h3 style="line-height: 24px; text-align: left;">Repeated sequences</h3><h3 style="line-height: 24px; text-align: left;"><div class="MsoNormal" style="line-height: 24px;"><span style="font-weight: normal;"><span style="font-size: small;">At several points in the Excel file, there are instances where the values of an ostensibly random variable are identical in two or more sequences of 10 or more participants, suggesting that ranges of cells or even entire rows of data have been copied and pasted.</span></span></div><div class="MsoNormal" style="line-height: 24px;"><span style="font-weight: normal;"><span style="font-size: small;"><br /></span></span></div></h3><h4 style="line-height: 24px; text-align: left;"><span>Approximately 19 cloned patients in group II</span></h4><div class="MsoNormal" style="line-height: 24px;"></div><h3 style="line-height: 24px;"><span style="font-size: small; font-weight: normal;">In cells B150:B168 and B184:B202, the patient's initials are either identical at each corresponding point (e.g., cells B150/B184) or, in almost all the remaining cases, differ in only one letter.</span></h3><h3 style="line-height: 24px;"></h3><h3 style="line-height: 24px;"></h3><div style="line-height: 24px; text-align: left;"><span style="font-weight: normal;">Cells C150:C168 are identical </span><span style="font-weight: normal;">to cells C184:C202.<br /></span>Cells D150:D168 are identical<span style="font-weight: 400;">—with one exception out of 19 cells—</span><span style="font-weight: normal;">to cells D184:D202.<br /></span><span style="font-weight: normal;">Cells I150:I167 are identical </span><span style="font-weight: normal;">to cells I184:I201.<br /></span><span style="font-weight: normal;">Cells S150:S165 are identical</span><span style="font-weight: 400;">—with one exception out of 14 cells—</span><span style="font-weight: normal;">to cells S184:S199.<br /></span><span style="font-weight: normal;">Cells U150:U168 are identical </span><span style="font-weight: normal;">to cells U184:U202.<br /></span><span style="font-weight: normal;">Cells V150:V168 are identical </span><span style="font-weight: normal;">to cells V184:V202.</span></div><div style="line-height: 24px; text-align: left;"><span style="font-weight: normal;">Cells W150:W168 are identical<span>—with three exceptions out of 19 cells—</span><span>to cells W150:W168.<br /></span></span></div><div style="line-height: 24px; text-align: left;"><span style="font-weight: normal;"><span>Cells AA150:AA168 are identical to cells AA184:AA202.</span></span></div><div style="line-height: 24px; text-align: left;"><span style="font-weight: normal;"><span><br /></span></span></div><h3 style="line-height: 24px;"><div></div></h3><h4 style="line-height: 24px;"><span>Approximately 60 cloned patients in group IV</span></h4><h3 style="line-height: 24px;"><div></div><div><span style="font-size: small;"><span style="font-weight: normal;">In cells B303:B320, B321:B338, and B339:B356, the patient's initials are either identical at each corresponding point (e.g., cells B303/B321/B339) or, in almost all the remaining cases, differ in only </span><span style="font-weight: normal;">one letter.</span></span></div><div><span style="font-weight: normal;"><span style="font-size: small;"><br /></span></span></div><div><span style="font-weight: normal;"><span style="font-size: small;">Cells I303:I320 are identical to cells I321:I338 and I339:I356, including the typo "coguh" for "cough".</span></span></div><div><span style="font-size: small;"><span style="font-weight: normal;">Cells I358:I371 are identical to cells I372:I385</span><span style="font-weight: 400;">, including the typo "coguh" for "cough".</span></span></div><div><span style="font-weight: normal;"><span style="font-size: small;">Cells I340:I349 are identical—with one exception out of 10 cells—to cells I386:I395.</span></span></div><div><span style="font-weight: normal;"><span style="font-size: small;"><br /></span></span></div><div><span style="font-weight: normal;"><span style="font-size: small;">Cells J303:J320 are identical to cells J321:J338 and J339:J356.</span></span></div><div><span style="font-weight: normal;"><span style="font-size: small;">Cells J358:J371 are identical to cells J372:J385.</span></span></div><div><span style="font-weight: normal;"><span style="font-size: small;">Cells J340:J349 are identical to cells J386:J395.</span></span></div><div><span style="font-weight: normal;"><span style="font-size: small;"><br /></span></span></div><div><span style="font-weight: normal;"><span style="font-size: small;">Cells K303:K320 are identical to cells K321:K338 and K339:K356.</span></span></div><div><span style="font-weight: normal;"><span style="font-size: small;">Cells K358:K371 are identical to cells K372:K385.</span></span></div><div><span style="font-weight: normal;"><span style="font-size: small;">Cells K340:K349 are identical to cells K386:K395.</span></span></div><div><span style="font-weight: normal;"><span style="font-size: small;"><br /></span></span></div></h3><h3 style="line-height: 24px;"><div><span style="font-weight: normal;"><span style="font-size: small;">Cells L303:L320 are identical—with two exceptions out of 18 cells—to cells L321:L338 and L339:L356.</span></span></div><div><span style="font-size: small;"><span style="font-weight: normal;">Cells L358:L371 are identical</span><span style="font-weight: 400;">—with one exception out of 14 cells—</span><span style="font-weight: normal;">to cells L372:L385.</span></span></div><div><span style="font-weight: normal;"><span style="font-size: small;">Cells L340:L349 are identical—with two exceptions out of 10 cells—to cells L386:L395.</span></span></div><div><span style="font-weight: normal;"><span style="font-size: small;"><br /></span></span></div><div><span style="font-weight: normal;"><span style="font-size: small;">Cells M303:M320 are identical to cells M321:M338 and M339:M356.</span></span></div><div><span style="font-weight: normal;"><span style="font-size: small;">Cells M358:M371 are identical to cells M372:M385.</span></span></div><div><span style="font-weight: normal;"><span style="font-size: small;">Cells M340:M349 are identical to cells M386:M395.</span></span></div><div><span style="font-weight: normal;"><span style="font-size: small;"><br /></span></span></div><div><span style="font-weight: normal;"><span style="font-size: small;">Cells S303:S320 are identical to cells S321:S338 and S339:S356.</span></span></div><div><span style="font-weight: normal;"><span style="font-size: small;">Cells S358:S371 are identical to cells S372:S385.</span></span></div><div><span style="font-weight: normal;"><span style="font-size: small;">Cells S340:S349 are identical to cells S386:S395.</span></span></div><div><span style="font-weight: normal;"><span style="font-size: small;"><br /></span></span></div><div><span style="font-weight: normal;"><span style="font-size: small;">Cells U303:U320 are identical to cells U321:U338 and U339:U356.</span></span></div><div><span style="font-weight: normal;"><span style="font-size: small;">Cells U358:U371 are identical to cells U372:U385.</span></span></div><div><span style="font-weight: normal;"><span style="font-size: small;">Cells U340:U349 are identical to cells U386:U395.</span></span></div><div><span style="font-weight: normal;"><span style="font-size: small;"><br /></span></span></div></h3><h3 style="line-height: 24px; text-align: left;"><div class="MsoNormal" style="line-height: 24px;"><span style="font-weight: normal;"><span style="font-size: small;">Cells W303:W320 are identical to cells W321:W338 and W339:W356.</span></span></div><div class="MsoNormal" style="line-height: 24px;"><span style="font-weight: normal;"><span style="font-size: small;">Cells W358:W371 are identical to cells W372:W385.</span></span></div><div class="MsoNormal" style="line-height: 24px;"><span style="font-weight: normal;"><span style="font-size: small;">Cells W340:W349 are identical to cells W386:W395.</span></span></div><div class="MsoNormal" style="line-height: 24px;"><br /></div><div class="MsoNormal" style="line-height: 24px;"><span style="font-weight: normal;"><span style="font-size: small;">Cells Y303:Y320 are identical (apart from spacing differences) to cells Y321:Y338 and</span></span></div><span style="font-size: small;">—<span style="font-weight: 400;">with one exception out of 18 cells—</span><span style="font-weight: normal;">Y339:Y356.</span></span><div class="MsoNormal" style="line-height: 24px;"><span style="font-size: small;"><span style="font-weight: normal;">Cells Y358:Y371 are identical</span><span style="font-weight: 400;">—with three exceptions out of 14 cells—</span><span style="font-weight: normal;">to cells Y372:Y385.</span></span></div><div class="MsoNormal" style="line-height: 24px;"><span style="font-weight: normal;"><span style="font-size: small;">Cells Y340:Y349 are identical to cells Y386:Y395.</span></span></div><div class="MsoNormal" style="line-height: 24px;"><span style="font-weight: normal;"><span style="font-size: small;"><br /></span></span></div></h3><h3 style="line-height: 24px;"><div class="MsoNormal" style="line-height: 24px;"><span style="font-size: small;"><span style="font-weight: normal;">Cells Z303:Z320 are identical to cells Z321:Y338 and Z339:Y356</span><span style="font-weight: normal;">.</span></span></div><div class="MsoNormal" style="line-height: 24px;"><span style="font-size: small;"><span style="font-weight: normal;">Cells Z358:Z371 are identical</span><span style="font-weight: 400;">—with three exceptions out of 14 cells—</span><span style="font-weight: normal;">to cells Z372:Z385.</span></span></div><div class="MsoNormal" style="line-height: 24px;"><span style="font-weight: normal;"><span style="font-size: small;">Cells Z340:Z349 are identical to cells Z386:Z395.</span></span></div></h3><h3 style="-webkit-text-stroke-width: 0px; color: black; font-family: "Times New Roman"; font-style: normal; font-variant-caps: normal; font-variant-ligatures: normal; letter-spacing: normal; line-height: 24px; orphans: 2; text-align: left; text-decoration-color: initial; text-decoration-style: initial; text-decoration-thickness: initial; text-indent: 0px; text-transform: none; white-space: normal; widows: 2; word-spacing: 0px;"><div class="MsoNormal" style="line-height: 24px;"><span style="font-size: medium; font-weight: normal;"><br /></span></div><div class="MsoNormal" style="line-height: 24px;"><span style="font-size: medium; font-weight: normal;"><div class="separator" style="clear: both; text-align: center;"><a href="https://lh3.googleusercontent.com/-TslVbTaUp3w/YOnHu8M3qQI/AAAAAAAAGHc/0k68Svfcrp0GZdEuFvF0ONW-OKVHasTTQCLcBGAsYHQ/image.png" style="margin-left: 1em; margin-right: 1em;"><img alt="" data-original-height="565" data-original-width="1465" height="246" src="https://lh3.googleusercontent.com/-TslVbTaUp3w/YOnHu8M3qQI/AAAAAAAAGHc/0k68Svfcrp0GZdEuFvF0ONW-OKVHasTTQCLcBGAsYHQ/w640-h246/image.png" width="640" /></a></div><br /></span></div><div class="MsoNormal" style="line-height: 24px;"><span style="font-weight: normal;"><div class="separator" style="clear: both; font-size: medium; text-align: center;"><a href="https://lh3.googleusercontent.com/-MaRWfT-03YA/YOnHJBTo2HI/AAAAAAAAGHU/LVpGAvEkwAUpipeawSAwvuhxYyRDgKg9gCLcBGAsYHQ/image.png" style="margin-left: 1em; margin-right: 1em;"><img alt="" data-original-height="909" data-original-width="1379" height="422" src="https://lh3.googleusercontent.com/-MaRWfT-03YA/YOnHJBTo2HI/AAAAAAAAGHU/LVpGAvEkwAUpipeawSAwvuhxYyRDgKg9gCLcBGAsYHQ/w640-h422/image.png" width="640" /></a></div><span style="font-size: small;"><div style="text-align: center;">Duplicated cells in groups II (top) and IV (bottom). In each column, groups of 10 or more cells with the same background colour and surrounded by a solid black border are identical. These images are screenshots from my annotated version of the Excel data file (see "Resources", below).</div></span><br /></span></div></h3><h3 style="line-height: 24px;"><span style="font-weight: normal;"><span style="font-size: small;">These patterns are not consistent with groups II and IV each containing the results of 100 different, real patients. The chances of any one of these duplications occurring by chance, let alone all of them, are astronomical. These patterns are, however, highly consistent with the idea that the Excel file has been fabricated with extensive use of copy/paste operations, followed perhaps by occasional attempts to obscure this "cloning" process by changing some numbers manually. Indeed, the slight imperfections in some of the copies would seem to exclude the possibility that these patterns are the result of an unfortunate slip of the mouse.</span></span></h3><div><span style="font-weight: normal;"><span style="font-size: small;"><br /></span></span></div><div style="line-height: 24px; text-align: left;"><span style="font-weight: normal;"><span style="font-size: small;">It seems ind</span><span style="font-size: small;">isputable that the patients in group II (mild/moderate disease, control condition) whose records are found at line 184 through 202 of the Excel file—a total of 19 people—are crude "clones" of the data of other patients (who, themselves, may or may not have actually existed). Similarly, it is hard to think of any explanation for the duplications in lines </span></span><span style="font-size: small;"><span style="font-weight: 400;">321 through 356 and 372 through 395, </span></span>other than that the records of around 32 patients in group IV (severe disease, control condition) have been "cloned", some of them multiple times. The question then naturally arises of which other records in the file may not reflect the reality of the patients in the study.</div><h3 style="line-height: 24px;"><div><span style="font-size: small; font-weight: normal;"><br /></span></div></h3><h3 style="line-height: 24px; text-align: left;">Apparent failures of randomisation</h3><div>The patients in groups I and II (mild/moderate disease, treatment and control) ought to have been similar to each other; likewise the patients in groups III and IV (severe disease, treatment and control). Indeed, the authors state (p. 3) that "A block randomization method was used to randomize the study participants into two groups that result in equal sample size. This method was used to ensure a balance in sample size across groups over the time and keep the number of participants in each group similar at all times". (Aside: I would be grateful if someone could explain to me what the second sentence there implies for the execution of the study.)</div><div><br /></div><div>However, the randomisation does not appear to have been a complete success. For example:</div><div><ul style="text-align: left;"><li>In group I, the number of patients with anosmia as an additional symptom was 25. In group II, this number was 4.</li><li>In group I, the number of patients with loss of taste as an additional symptom was 25. In group II, this number was 0.</li><li>In group III, the number of patients with vomiting as an additional symptom was 1. In group IV, this number was 12.</li><li>In group III, the number of patients with bronchial asthma as a comorbidity was 14. In group IV, this number was 0.</li><li>In group III, the number of patients with cholecystitis, chronic kidney disease, hepatitis B, hepatitis C, and open heart surgery as comorbidities was 0 in all five cases. In group IV, these numbers were 6, 5, 5, 6, and 6, respectively.</li></ul><br /></div><h3 style="line-height: 24px; text-align: left;">Descriptive statistics that do not match the preprint</h3><div>The first three paragraphs of the Results section of Elgazzar et al.'s preprint contain descriptions of the characteristics of their sample. Here, I reproduce the text of each of those paragraphs. Where the numbers that I calculated from the data set differ from those reported, I have included my calculated values in red and inside brackets. It should be apparent that while a few of the numbers calculated from the Excel sheet match those in the preprint, the great majority do not.</div><div><br /></div><div>First paragraph:</div></div><blockquote style="border: none; margin: 0px 0px 0px 40px; padding: 0px;"><div class="MsoNormal" style="line-height: 24px;"><div><div style="text-align: left;">The mean age in Group I was 56.7 <span style="color: red;">[47.5]</span> ±18.4 <span style="color: red;">[15.1]</span>; included 72 % males and 28 % females. The mean age in Group II was 53.8 <span style="color: red;">[43.2]</span> ±21.3 <span style="color: red;">[16.1]</span>; included 67 <span style="color: red;">[66] </span>% males and 33 <span style="color: red;">[34] </span>% females. The mean age in Group III was 58.2 <span style="color: red;">[55.0] </span>±20.9 <span style="color: red;">[14.0]</span>; included 68 <span style="color: red;">[74]</span> % males and 32 <span style="color: red;">[26]</span> % females. The mean age in Group IV was 59.6 <span style="color: red;">[54.2]</span> ±18.2 <span style="color: red;">[13.7]</span>; included 74 <span style="color: red;">[73]</span> % males and 26 <span style="color: red;">[27]</span>% females. The mean age in Group V was 57.6 <span style="color: red;">[48.8]</span> ±18.4 <span style="color: red;">[9.2]</span>; included 75 % males and 25% females. The mean age in Group VI was 56.8 <span style="color: red;">[54.4]</span>±18.2 <span style="color: red;">[8.8]</span>; included 72 % males and 28% females. There was no statistical significance variation between groups regarding mean age or sex distribution (p-value >0.05).</div></div></div></blockquote><div class="MsoNormal" style="line-height: 24px;"><br /></div><div class="MsoNormal" style="line-height: 24px;">The sex of one of the participants in group V was coded in the Excel sheet as "A" (cell C449), rather than "M" or "F". The preprint made no mention of any patients identifying as anything other than Male or Female. In order for the numbers of patients of each sex in group V to match the numbers reported in the preprint, I counted "A" as "M".</div><div class="MsoNormal" style="line-height: 24px;"><br /></div><div class="MsoNormal" style="line-height: 24px;"><div>Second paragraph:</div></div><blockquote style="border: none; margin: 0px 0px 0px 40px; padding: 0px;"><div class="MsoNormal" style="line-height: 24px;"><div class="MsoNormal" style="line-height: 24px; text-align: left;">Co morbid conditions distributed between different studied groups showed that DM was present in 15 <span style="color: red;">[4]</span>% of Group I patients, 14 <span style="color: red;">[16]</span>% of Group II patients, 18% of Group III patients, 21 <span style="color: red;">[26]</span>% of Group IV patients 15% of group V and 19 % of group VI. HTN presented in 11 <span style="color: red;">[6]</span>% of Group I patients, 12 <span style="color: red;">[13]</span> % of Group II patients, 14% of Group III patients, 18 <span style="color: red;">[32]</span>% of Group IV patients ,15 <span style="color: red;">[14]</span> % of group V patients and 14 <span style="color: red;">[13]</span>%of group VI patients . 2 <span style="color: red;">[1]</span>% of Group I patients had IHD versus 6 <span style="color: red;">[7]</span>% in Group II, 5% in group III; 12 <span style="color: red;">[5]</span>% in group IV;1% in group V and 3 <span style="color: red;">[4]</span> % in group VI respectively with statistically significant prevalence of ischemic heart disease as severity increase (p-value < 0.03).. Bronchial asthma presented in 5 <span style="color: red;">[3]</span>% of Group I patients, 6 % of Group II patients, and 14% of Group III patients, in 12 <span style="color: red;">[0]</span>% of Group IV patients; 5% of group V and 4% of group VI patients.</div></div></blockquote><div class="MsoNormal" style="line-height: 24px;"><br /></div><div class="MsoNormal" style="line-height: 24px;">I assume that the authors' calculation of "prevalence of ischemic heart disease as severity increase" involved grouping the patients into three pairs of groups by severity (groups I and II, groups III and IV, and groups V and VI). Here are the results of that operation using, first, their IHD prevalence numbers, and second, my calculated numbers. The authors reported a p value of "< 0.03".</div><div class="MsoNormal" style="line-height: 24px;"><br /></div><div class="MsoNormal" style="line-height: 24px;"><div class="MsoNormal" style="line-height: 24px;"><span style="font-family: courier;">> chisq.test(c(8, 15, 4))</span></div><div class="MsoNormal" style="line-height: 24px;"><span style="font-family: courier;">X-squared = 6.8889, df = 2, p-value = 0.03192</span></div><div class="MsoNormal" style="line-height: 24px;"><span style="font-family: courier;"><br /></span></div><div class="MsoNormal" style="line-height: 24px;"><span style="font-family: courier;">> chisq.test(c(8, 10, 5))</span></div><div class="MsoNormal" style="line-height: 24px;"><span style="font-family: courier;">X-squared = 1.6522, df = 2, p-value = 0.4378</span></div><div><br /></div><div><div>Third paragraph:</div></div></div><blockquote style="border: none; margin: 0px 0px 0px 40px; padding: 0px;"><div class="MsoNormal" style="line-height: 24px;"><div class="MsoNormal" style="line-height: 24px; text-align: left;">Clinically there was a highly statistically significant difference between groups of diseased patients regarding fatigue, dyspnea, and respiratory failure (p-value <0.001), as most of group III & IV, showed fatigue and dyspnea (86 <span style="color: red;">[86, 85]</span>% and 88 <span style="color: red;">[85, 84]</span>%, respectively), compared to (36 <span style="color: red;">[28]</span>%, 38 <span style="color: red;">[47]</span>% ; 54 <span style="color: red;">[34]</span>% and 52 <span style="color: red;">[49]</span>%, respectively), in group I & II. Respiratory failure had been detected in 38% and 40% in group III& IV respectively while no patients in group I& II developed respiratory failure. No skin manifestation had been detected in any group.</div></div></blockquote><div class="MsoNormal" style="line-height: 24px;"><br /></div><div class="MsoNormal" style="line-height: 24px;">The authors' reporting of fatigue and dyspnea is unclear here, as for groups III and IV they report only two percentages. I have assumed that their claim was that these were identical for each of the two groups (i.e., fatigue, 86 for both groups; dyspnea, 88 for both groups), whereas I found three out of four numbers to be different. I was unable to calculate the percentage of respiratory failure as this was not apparently reported in the data file, although "sore throat" was. Nor could I find anything in the data file corresponding to "skin manifestation". Regarding the "highly statistically significant difference between groups of diseased patients", the <i>p</i> values are less than 0.001 (indeed, less that 1E-9) with the authors' numbers or mine.</div><div class="MsoNormal" style="line-height: 24px;"><h3 style="line-height: 24px;"><br /></h3><h3 style="line-height: 24px;">Table results that do not match the preprint</h3><div>I wrote some R code that attempts to reproduce the authors' tables, as far as possible. This is available <a href="http://nickbrown.fr/blog/elgazzar" target="_blank">here</a> so that readers can judge (a) whether the small number of decisions that I needed to make in order to adapt the Excel sheet for analysis are reasonable, and (b) whether I have programmed the subsequent calculations correctly.</div><div><br /></div><div>Here are the results that I obtained. (Full-resolution images are available at the same link as the code.) Readers are invited to compare these results with the tables in Elgazzar et al.'s preprint. I think it is fair to say that there is a substantial degree of divergence.</div><div><br /></div><div><u>Table 1</u></div><div><u><br /></u></div><div><u><div class="separator" style="clear: both; text-align: center;"><a href="https://lh3.googleusercontent.com/-Lp_IA8Po5MU/YOjftnWDCfI/AAAAAAAAGHA/yhU3dxTfNAggCK3dBOVq7D1xerYKK9A5gCLcBGAsYHQ/image.png" style="margin-left: 1em; margin-right: 1em;"><img alt="" data-original-height="457" data-original-width="705" height="414" src="https://lh3.googleusercontent.com/-Lp_IA8Po5MU/YOjftnWDCfI/AAAAAAAAGHA/yhU3dxTfNAggCK3dBOVq7D1xerYKK9A5gCLcBGAsYHQ/w640-h414/image.png" width="640" /></a></div><br /></u><div class="separator" style="clear: both; text-align: center;"><a href="https://lh3.googleusercontent.com/-WF-T_xA_pG0/YOnFezoJ0-I/AAAAAAAAGHM/UfHsvBVU0U8g13XvH3SEq7joDVrQhe5AACLcBGAsYHQ/image.png" style="margin-left: 1em; margin-right: 1em;"><img alt="" data-original-height="190" data-original-width="844" height="144" src="https://lh3.googleusercontent.com/-WF-T_xA_pG0/YOnFezoJ0-I/AAAAAAAAGHM/UfHsvBVU0U8g13XvH3SEq7joDVrQhe5AACLcBGAsYHQ/w640-h144/image.png" width="640" /></a></div><br />The D-dimer results from the Table 1 in preprint could not be reproduced because no data for that measure exist in the Excel file.</div><div><br /></div><div><u>Tables 2 and 3</u></div><div><u><br /></u></div><div><u><div class="separator" style="clear: both; text-align: center;"><a href="https://lh3.googleusercontent.com/-F0gxgHh3zUo/YOjfcv3kidI/AAAAAAAAGGw/TTzb4a_taMYDF3ZlBUBKBzOXWFTfH1e1wCLcBGAsYHQ/image.png" style="margin-left: 1em; margin-right: 1em;"><img alt="" data-original-height="318" data-original-width="933" height="218" src="https://lh3.googleusercontent.com/-F0gxgHh3zUo/YOjfcv3kidI/AAAAAAAAGGw/TTzb4a_taMYDF3ZlBUBKBzOXWFTfH1e1wCLcBGAsYHQ/w640-h218/image.png" width="640" /></a></div><div class="separator" style="clear: both; text-align: center;"><a href="https://lh3.googleusercontent.com/-4h0unHMJJe8/YOjfkal3taI/AAAAAAAAGG4/7PkDnv9fB80_TN9LGEvRAQsUkHpOizUXgCLcBGAsYHQ/image.png" style="margin-left: 1em; margin-right: 1em;"><img alt="" data-original-height="353" data-original-width="819" height="276" src="https://lh3.googleusercontent.com/-4h0unHMJJe8/YOjfkal3taI/AAAAAAAAGG4/7PkDnv9fB80_TN9LGEvRAQsUkHpOizUXgCLcBGAsYHQ/w640-h276/image.png" width="640" /></a></div><br /></u></div><div><div class="separator" style="clear: both; text-align: center;"><div class="separator" style="clear: both; text-align: center;"><a href="https://lh3.googleusercontent.com/-ph1cdj5Nj6Q/YOjYAJf1knI/AAAAAAAAGGk/4lCYPyhKYwAwM3rIiM9vxZV8EnRahUSXQCLcBGAsYHQ/Table2%25263-repro.png" style="margin-left: 1em; margin-right: 1em;"><img alt="" data-original-height="214" data-original-width="607" height="226" src="https://lh3.googleusercontent.com/-ph1cdj5Nj6Q/YOjYAJf1knI/AAAAAAAAGGk/4lCYPyhKYwAwM3rIiM9vxZV8EnRahUSXQCLcBGAsYHQ/w640-h226/Table2%25263-repro.png" width="640" /></a></div><br /></div>Only three of the elements from the preprint's Tables 2 and 3 correspond to values in the Excel file after one week of treatment. Longitudinal data for HGB, TLC, Lymphocyte % are all missing, and the Excel file contains no data for D-dimer at any time point.</div><div><br /></div><div>The time difference between the first and last RT-PCR tests, can be calculated in two ways: either using the authors' provided field (column Y) or by subtracting the date of the first PCR test (column W) from the date of the final PCR test (column X) and then (as was apparently done by the authors, at least for group I) adding one.<br /><br /></div><div><u>Table 4</u></div><div><u><br /></u></div><div><u><div class="separator" style="clear: both; text-align: center;"><a href="https://lh3.googleusercontent.com/-ZAWv_mFyQgE/YOjfS995SkI/AAAAAAAAGGs/qvpi0OQplycuNk5XEJ1OhJBwI_yROf2NACLcBGAsYHQ/image.png" style="margin-left: 1em; margin-right: 1em;"><img alt="" data-original-height="452" data-original-width="840" height="344" src="https://lh3.googleusercontent.com/-ZAWv_mFyQgE/YOjfS995SkI/AAAAAAAAGGs/qvpi0OQplycuNk5XEJ1OhJBwI_yROf2NACLcBGAsYHQ/w640-h344/image.png" width="640" /></a></div><br /></u></div><div><div class="separator" style="clear: both; text-align: center;"><a href="https://lh3.googleusercontent.com/-z8jK5obUiSg/YOjV_iXMvrI/AAAAAAAAGGU/BW6IoPnpA-g5uQSK3suWvAUHL4oClGw3gCLcBGAsYHQ/image.png" style="margin-left: 1em; margin-right: 1em;"><img alt="" data-original-height="147" data-original-width="688" height="136" src="https://lh3.googleusercontent.com/-z8jK5obUiSg/YOjV_iXMvrI/AAAAAAAAGGU/BW6IoPnpA-g5uQSK3suWvAUHL4oClGw3gCLcBGAsYHQ/w640-h136/image.png" width="640" /></a></div>The first half of Table 4 ("Prognosis") cannot be reproduced, as this variable only exists in the Excel file for groups V and VI. As for Tables 2 and 3, there are two ways to calculate the time of stay in the hospital. The per-group ranges associated with the version labelled "RecordedStay", corresponding to column Y in the Excel sheet, are not too far from the ones reported in the preprint, with 6 out of 8 numbers (minimum or maximum) being identical; would seems to suggest that the reproduction is on the right track.</div><div><br /></div><div>Also noteworthy here are the extremely small standard deviations of the stays in group I (both as recorded in column Y, and as calculated from columns W and X) and group III (as calculated from columns W and X; in this last case I find myself wondering why there is such a difference between the SDs of the recorded and calculated stays).</div><div><br /></div><div>Relatedly, in the preprint, the standard deviations for the hospital stay are remarkably different between the groups. The large SD for group IV (8, with a mean of 18 and a range of 9–25) implies that about 40% of the patients stayed 9 days and 60% stayed for 25, with almost no room for any other lengths of stay, as shown by SPRITE (Heathers et al., 2018, https://peerj.com/preprints/26968/; https://shiny.ieis.tue.nl/sprite/).</div><div><br /></div><div><div class="separator" style="clear: both; text-align: center;"><a href="https://lh3.googleusercontent.com/-v1QdKfo5nEk/YOoYmZfzRdI/AAAAAAAAGHk/_fBzRbJDxCg-4XXt4l8pvkiXntZTPMxywCLcBGAsYHQ/image.png" style="margin-left: 1em; margin-right: 1em;"><img alt="" data-original-height="870" data-original-width="1587" height="350" src="https://lh3.googleusercontent.com/-v1QdKfo5nEk/YOoYmZfzRdI/AAAAAAAAGHk/_fBzRbJDxCg-4XXt4l8pvkiXntZTPMxywCLcBGAsYHQ/w640-h350/image.png" width="640" /></a></div><div style="text-align: center;">SPRITE analysis of the possible distribution of the recovery times claimed by Elgazzar et al. for patients in group IV.</div><br /></div><div>[[Begin update 2021-07-16 12:46 UTC]]</div><div>Alert reader Anatoly Lubarsky <a href="https://twitter.com/anatoly/status/1415827806865563649?s=20" target="_blank">pointed out on Twitter</a> that there are combinations of stay lengths that do not involve quite as many values at the limits, 9 and 25, as the chart above. He is correct. By specifying one decimal place when generating the above chart I had, in effect, told SPRITE to look for SD values in the range 7.95–8.05, whereas the authors reported only integers and so their SD could have been anywhere from 7.50001 through 8.49999. It's a bit ironic that I missed that since I previously wrote <a href="https://steamtraen.blogspot.com/2018/07/this-researcher-compared-two-identical.html">this post</a> on pretty much exactly this topic.</div><div><br /></div><div>rSPRITE doesn't currently work with zero decimal places, but Anatoly also <a href="https://docs.google.com/spreadsheets/d/1Hd-UdfnAAB7Oq_ecYg99z2EPCKc2CY1FcLHRqXCna3s/edit#gid=1851759707" target="_blank">provided an example</a> that he had constructed to show what seems to me to be the most favourable (i.e., the least extreme) result from the point of view of the authors. Here is the resulting chart from that example. I do not think that this greatly alters the idea that this pattern of days spent in hospital is unlikely to be a reflection of real-world data.</div><div><br /></div><div><div class="separator" style="clear: both; text-align: center;"><a href="https://lh3.googleusercontent.com/-ZrJdwjQ-LEk/YPF94RUHeZI/AAAAAAAAGIk/LBWUWjRZ9uUVlm5eOSPACbQtyDZE2sMNgCLcBGAsYHQ/image.png" style="margin-left: 1em; margin-right: 1em;"><img alt="" data-original-height="413" data-original-width="874" height="302" src="https://lh3.googleusercontent.com/-ZrJdwjQ-LEk/YPF94RUHeZI/AAAAAAAAGIk/LBWUWjRZ9uUVlm5eOSPACbQtyDZE2sMNgCLcBGAsYHQ/w640-h302/image.png" width="640" /></a></div><div style="text-align: center;">Chart showing 100 values with minimum=9, maximum=25, mean=17.86 (which rounds to 18 if no decimal places are included), and SD=7.505 (rounds to 8), cf. Elgazzar et al.'s Table 4, bottom row, group IV.</div></div><div><div style="text-align: center;">See the file "<span style="font-family: courier;">SD-simulation.xls</span>" in "Resources", below.</div><br /></div><div><div>[[End update 2021-07-16 12:35 UTC]]</div></div><div><br /></div><div><br /></div><div>It is unclear why the authors claim to have performed a chi-squared test (χ2=87.6, p<0.001) on the value of "Recovery time &Hospital stay", as it is clearly not a categorical variable. It is tempting to imagine that this result was copied and pasted from the first half of Table 4 (with the test statistic being altered by subtracting 1 from each digit) by someone who did not understand what they were doing and did not realise that a chi-squared test is meaningless here.</div><div><br /></div><div><u>Table 5</u></div><div>The data that would be needed to reproduce Table 5 do not seem to be available in the Excel file.</div><div><br /></div><div><u>Table 6</u></div><div>As with Table 4, the first part of this table cannot be reproduced, as the "Prognosis" variable is not available in the Excel file.</div><div><br /></div><div>I have not reproduced the second part of Table 6 as it appears to be redundant or to use unavailable data. The RT-PCR results correspond to the last lines of Tables 2 and 3. Interestingly, the "Hospital Stay" variable appears to be different from "RT-PCR" here, although as I hope to have demonstrated earlier, the variable marked "Hospital Stay" in column Y of the data file has a very close relationship with the difference between columns W ("symptoms date&+ ve PCR") and X ("recovery date & -ve PCR"). It seems that the authors are unsure whether the difference in days between the first positive and last negative PCR test (columns W and X) corresponds to the hospital stay or not, with or without an adjustment of one day for the "fencepost" issue mentioned earlier.</div><div><br /></div><h3 style="text-align: left;">Other issues</h3><div><br /></div><h4 style="line-height: 24px; text-align: left;">The age distribution</h4><div class="MsoNormal" style="line-height: 24px;">The distribution of patient ages is very strange. There are 34 patients aged 48 and 31 aged 58, but only 3 aged 50 and 4 aged 53. Furthermore, of the 600 patients, 410 have an age that is an even number of years while only 190 have an age that is an odd number of years.</div><div class="MsoNormal" style="line-height: 24px;"><br /></div><div class="MsoNormal" style="line-height: 24px;">It is difficult to see how any of this could have arisen by chance. (The R function <span style="font-family: courier;">pbinom()</span> reports that the binomial probability of 399 out of 600 ages being even is 1.11E-16; the chance of there being 400 or more even ages out of 600 is too small for R to calculate.)</div><div class="MsoNormal" style="line-height: 24px;"><div class="separator" style="clear: both; text-align: center;"><br /></div><div class="separator" style="clear: both; text-align: center;"><a href="https://lh3.googleusercontent.com/-Cq31P_iNwWE/YOdB21XVEJI/AAAAAAAAGFg/-ogsUEiY7sUcZ-MNViQ6WWmLH0pcuS2_QCLcBGAsYHQ/AgeHist.png" style="margin-left: 1em; margin-right: 1em;"><img alt="" data-original-height="621" data-original-width="846" height="470" src="https://lh3.googleusercontent.com/-Cq31P_iNwWE/YOdB21XVEJI/AAAAAAAAGFg/-ogsUEiY7sUcZ-MNViQ6WWmLH0pcuS2_QCLcBGAsYHQ/w640-h470/AgeHist.png" width="640" /></a></div><div class="separator" style="clear: both; text-align: left;"><br /></div><h4 style="text-align: left;">Trailing digits of numerical variables</h4><div><a href="https://twitter.com/K_Sheldrick" target="_blank">Kyle Sheldrick</a> discovered that of the 400 values of the variable "serum ferritin before treatment", only three end in the digit 3. I looked at the other numerical results in the preprint and found that in almost all cases the distribution of the trailing digits is extremely unusual, and in contrast with what one would expect from <a href="https://en.wikipedia.org/wiki/Benford%27s_law" target="_blank">Benford's Law</a>, which—although it is perhaps best known for its predictions about the first digits of data corresponding to natural phenomena anchored at zero—shows that for the digits of a random variable apart from the first, the expected distribution is approximately uniform. (In November 2020 I used the predictions of the distributions of trailing digits from Benford's Law to <a href="https://twitter.com/sTeamTraen/status/1331985658139860998?s=20" target="_blank">demonstrate</a> that the official Covid-19 statistics from the Turkish Ministry of Health were <a href="https://t24.com.tr/haber/veri-bilimci-dr-nick-brown-saglik-bakanligi-verilerinin-dogru-olma-ihtimali-milyarda-bir-yani-neredeyse-imkansiz,918347" target="_blank">probably fabricated</a>.)</div><div><br /></div><div>In cases where the dominant trailing digit is zero we might allow for the possibility that different people collected data to different degrees of precision, thus leading to numbers being rounded and, consequently, a trailing digit of zero more often than might be expected by chance. But this cannot explain why, for example, 82% of the numbers for HGB end in the digits 2–5, or why 17.5% of the numbers for TLC end in 8 whereas <i>none</i> end in 2. The large chi-square statistics and their associated homeopathic <i>p</i> values in the tests of the trailing digits from Elgazzar et al.'s data file suggest that none of these patterns are the result of a natural process. They are, however, highly compatible with the idea that the numbers in the Excel table have either been copied and pasted in bulk, or invented out of whole cloth by someone who was trying (and failing) to simulate random numbers—an activity that <a href="https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0041531" target="_blank">humans are not very good at</a>.</div><div><br /></div><div><div class="separator" style="clear: both; text-align: center;"><a href="https://lh3.googleusercontent.com/-YSCv2W10Ge4/YO8R7nIU89I/AAAAAAAAGIA/TwRghFurCDsYL9aFXL7ioxOYsKyfz3KRQCLcBGAsYHQ/image.png" style="margin-left: 1em; margin-right: 1em;"><img alt="" data-original-height="610" data-original-width="336" height="640" src="https://lh3.googleusercontent.com/-YSCv2W10Ge4/YO8R7nIU89I/AAAAAAAAGIA/TwRghFurCDsYL9aFXL7ioxOYsKyfz3KRQCLcBGAsYHQ/w352-h640/image.png" width="352" /></a></div><div style="text-align: center;">Counts of the trailing digits (0–9) of various numeric variables in Elgazzar et al.'s data file, and the chi-square statistics for the test against the null hypothesis that their distribution is uniform.</div></div><div style="text-align: center;"><br /></div><div style="text-align: center;"><br /></div><h4 style="text-align: left;">Study entry and exit dates</h4></div><div class="MsoNormal" style="line-height: 24px;">The preprint states (p. 3) that "The study was carried out from 8th June to 15th September 2020". This seems to conflict with <a href="https://clinicaltrials.gov/ct2/show/results/NCT04668469?view=results" target="_blank">the study's registration on ClinicalTrials.gov</a>, which states that the "Actual Study Completion Date"—defined as "The date [of] the last participant's last visit"—was 30 October 2020. We cannot, perhaps, infer much from the fact that the last recorded entry (positive PCR) and exit (negative PCR) dates in the Excel file are 18 August 2020 and 21 August 2020, respectively, as we do not have date information for the outpatients (groups V and VI). However, we can see that there are 120 patients (71 in group II, 3 in group III, and 47 in group IV) with an entry date prior to 8 June 2020, with the earliest being 12 May 2020. Similarly, there are 49 patients (31 in group II, 1 in group III, and 17 in group IV) with an exit date prior to 8 June 2020, with the earliest being 23 May 2020.</div><div class="MsoNormal" style="line-height: 24px;"><br /></div><h4 style="line-height: 24px; text-align: left;">SPSS</h4><div class="MsoNormal" style="line-height: 24px;">Another strange feature of this story is that, although the authors claim to have performed their analyses using SPSS (p. 4 of the preprint), they did not share the SPSS data file (in .SAV or .CSV format), although this would have been a much better way to allow readers to reproduce their analyses. Instead they shared what they called the "study data master sheet". As I have shown here, these data (a) contain numerous signs of manipulation and (b) once cleaned up and analysed with the same statistical tests that authors used, mostly—but, perhaps significantly, not entirely—fail to produce the results reported by the authors in their Results section text and tables.</div><div class="MsoNormal" style="line-height: 24px;"><div class="MsoNormal" style="line-height: 24px;"><br /></div><div class="MsoNormal" style="line-height: 24px;">There is another curious sentence in the preprint that makes me wonder whether the authors actually used SPSS at all, or indeed have ever done so. On p. 5 they wrote "After the calculation of each of the test statistics, the corresponding distribution tables were counseled to get the 'P' (probability value)". Assuming that "counseled" here is a typo for "consulted", it appears that the authors' claim is that they read the test statistics from the SPSS output and then looked up the corresponding <i>p</i> values in a table, such as <a href="https://www.itl.nist.gov/div898/handbook/eda/section3/eda3672.htm" target="_blank">the one on this page</a>. I wonder why anyone would do this, given that the SPSS output for all of the tests that the authors reported having run contains the <i>p</i> value right next to the test statistic. Looking up test statistics in a table to get the <i>p</i> value has been out of fashion since we stopped computing <i>t</i> statistics using pencil and paper, circa 1995 ("Ah, now I know why my desk calculator has a square root key").</div></div><div class="MsoNormal" style="line-height: 24px;"><br /></div></div><h3 style="line-height: 24px; text-align: left;">Conclusion</h3><div class="MsoNormal" style="line-height: 24px;">In view of the problems described in the preceding sections, most notably the repeated sequences of identical numbers corresponding to apparently "cloned" patients, it is difficult to avoid the conclusion that the Excel file provided by the authors does not faithfully represent the results of the study, and indeed has probably been extensively manipulated by hand.</div><div class="MsoNormal" style="line-height: 24px;"><br /></div><div class="MsoNormal" style="line-height: 24px;">In some cases where forensic researchers have discovered discrepancies between images or datasets and the results reported in a paper, the authors have attempted to claim that they "accidentally" provided a "training" version that had been made to calibrate their software. It does not seem possible that such a defence could be used in this case, however, since the Excel sheet provided by Elgazzar et al. cannot possibly have been used for this purpose, in view of the extensive amount of manual cleaning that would be required to make it useable for any purpose.</div><div class="MsoNormal" style="line-height: 24px;"><br /></div><div class="MsoNormal" style="line-height: 24px;">I urge the authors to make their SPSS data file publicly available without delay, in order that we can see the exact numbers on which their analyses were based—because, as demonstrated above, those numbers cannot be those in the Excel file. If the authors cannot provide their SPSS data file then I believe that either they or Research Square should consider retracting their preprint as a matter of urgency.</div><div class="MsoNormal" style="line-height: 24px;"><br /></div><div class="MsoNormal" style="line-height: 24px;"><h3 style="line-height: 24px;">Resources</h3><div>I have made the following files available <a href="http://nickbrown.fr/blog/elgazzar" target="_blank">here</a>:</div><div><ul style="text-align: left;"><li>The unlocked original data file, named "<span style="font-family: courier;">Copy of covid_19 final master sheet12-12-2020 (1).xlsx</span>". That is exactly the name that it would have if you were to download it (albeit still locked with a password at that point).</li><li>The password-protected version of the original data file, to the name of which I have prepended <span style="font-family: courier;">(Locked)</span>, so the full name is now "<span style="font-family: courier;">(Locked) Copy of covid_19 final master sheet12-12-2020 (1).xlsx</span>". The password for read-only access is 1234. If someone has the right tools to extract the second-level password that is needed to modify the contents of the file, I'd be curious to know what it is. [[ Update 2021-07-17 15:58 UTC: <a href="https://twitter.com/gsuberland/status/1416424606974492678?s=20" target="_blank">Graham Sutherland has discovered</a> that there are many passwords that unlock read-write access, the simplest of which appears to be 0001. ]]</li><li>An annotated version of that file, named "<span style="font-family: courier;">(Nick) Elgazzar data.xls</span>". I have highlighted the anomalous individual cells in yellow and the runs of duplicated cells in a variety of other colours.</li><li>Slightly higher resolution versions of the images from this post, named "<span style="font-family: courier;">Table1.png</span>" (etc), "<span style="font-family: courier;">SPRITE-Table4.png</span>", and "<span style="font-family: courier;">Trailing-digits.png</span>".</li><li>My analysis code, named "<span style="font-family: courier;">Elgazzar.R</span>".</li></ul><div><br /></div></div><h3 style="text-align: left;">Acknowledgements</h3><div>Thanks to Jack Lawrence (@JackMLawrence) for bringing the paper to my attention and cracking the password of the data file; Gideon Meyerowitz-Katz (@GidMK) for pointing out the Table 4 standard deviation problem, the issue with the patient dates relative to the reported study dates, and the claim that the <i>p</i> values were looked up in a table; and Kyle Sheldrick (@K_Sheldrick) for making the initial discovery of the lack of trailing 3s in the serum ferritin numbers.</div><div><br /></div><div><br /></div><div>(*) Things move fast in Covid world. Less than 24 hours before this blog post was due to be published, Research Square posted <a href="https://www.researchsquare.com/article/rs-100956/v4" target="_blank">"V4" of the preprint</a>, which is simply a placeholder that says "Research Square has withdrawn this preprint due to ethical concerns". To ensure that the preprint does not disappear, I have posted the PDFs of the first three versions in the same location as the other supporting files for this post (see "Resources", above).</div><div><br /></div><div><div class="separator" style="clear: both; text-align: center;"><a href="https://lh3.googleusercontent.com/-zK66BXTMcY8/YO9c4uTCDSI/AAAAAAAAGII/6rbro7DP80gDyXww0ilKfHmqF-JioCWrQCLcBGAsYHQ/image.png" style="margin-left: 1em; margin-right: 1em;"><img alt="" data-original-height="684" data-original-width="1566" height="280" src="https://lh3.googleusercontent.com/-zK66BXTMcY8/YO9c4uTCDSI/AAAAAAAAGII/6rbro7DP80gDyXww0ilKfHmqF-JioCWrQCLcBGAsYHQ/w640-h280/image.png" width="640" /></a></div><br /><div style="text-align: center;">2021-07-14: Research Square withdraws the preprint.</div></div><div><br /></div></div>Nick Brownhttp://www.blogger.com/profile/07481236547943428014noreply@blogger.com0tag:blogger.com,1999:blog-7890764972166411105.post-45201321300591934012021-05-06T20:02:00.007+02:002021-05-10T22:58:40.575+02:00My minor involvement in the investigation of some strange articles from marine ecology<div style="line-height: 24px;"><p>Today's topic is <a href="https://www.sciencemag.org/news/2021/05/does-ocean-acidification-alter-fish-behavior-fraud-allegations-create-sea-doubt">this report</a> in <i>Science</i> by Martin Enserink about possible scientific misconduct in a series of studies that investigate the relation between increasing CO<span style="font-size: xx-small;">2</span> levels (causing a decrease in the pH of the world's oceans) and the behaviour of fish. Martin's report gives most of the background that you will need to follow this post. While he was preparing it, he asked me to look at the dataset for a couple of articles from the research group whose work he was investigating. In particular, I found a lot of interesting things in this article:</p><div style="line-height: 24px; margin-left: 36pt; text-indent: -36pt;"><span style="mso-no-proof: yes;">Dixson, D. L., Abrego, D., & Hay, M. A. </span>(2014). <span style="text-indent: -36pt;">Chemically mediated behavior of </span><span style="text-indent: -36pt;">recruiting corals and fishes: A tipping </span><span style="text-indent: -36pt;">point that may limit reef recovery</span><span style="text-indent: -36pt;">. </span><i style="mso-bidi-font-style: normal; text-indent: -36pt;">Science, 345</i><span style="mso-bidi-font-style: normal; text-indent: -36pt;">(6192)</span><span style="text-indent: -36pt;">, </span><span style="text-indent: -36pt;">892–897. https://doi.org/10.1126/science.1255057</span></div><div style="line-height: 24px;">(You can find the PDF of the article on a ResearchGate page <a href="https://www.researchgate.net/publication/264986380_Chemically_mediated_behavior_of_recruiting_corals_and_fishes_A_tipping_point_that_may_limit_reef_recovery" target="_blank">here</a>; I'm not sure if <a href="https://www.researchgate.net/profile/Mark-Hay/publication/264986380_Chemically_mediated_behavior_of_recruiting_corals_and_fishes_A_tipping_point_that_may_limit_reef_recovery/links/56e8903008ae9bcb3e1cd68a/Chemically-mediated-behavior-of-recruiting-corals-and-fishes-A-tipping-point-that-may-limit-reef-recovery.pdf?_sg%5B0%5D=aHtbcokxLq6DNPAbDUiS7hgzVPzcWS4AwWAs53tzAffJC81LmmO1RCfNb_mv7MJSeZIb3Hy2dBvKSmrXcBBb7A.4GyGqEnrE_jjQzFCrro1eZjCSwVUSJkSRqsBeq2SKRAwUn7mHRXV3uQrq4l9dq9_sYOEBIvUTMtWOVD9FG9W0w&_sg%5B1%5D=BYmBFJwaWRBGsxyz29oXTUzkRoemG0Y-7bnI3o6bvpwgBN3UwzeF8Vh1xHTZS9z_Uq8TDKLsFCHFdEJW2uluo-Xta50deznEjrkxFGtiXadO.4GyGqEnrE_jjQzFCrro1eZjCSwVUSJkSRqsBeq2SKRAwUn7mHRXV3uQrq4l9dq9_sYOEBIvUTMtWOVD9FG9W0w&_iepl=">this direct link</a> to the file will work.)</div><p><span style="text-indent: -48px;">Most of my analyses of the article and its associated dataset are written up in a report that you can find </span><a href="http://www.meta-systems.eu/nickbrown/blog/dixson/20210510%20-%20Analysis%20of%20the%20Dixson%20et%20al.%20data%20set.pdf" style="text-indent: -48px;">here</a> [PDF]. In this short post I just want to mention one other point that isn't in that report, which is the whole question of why the dataset is in the form of an Excel file (which you can find <a href="http://www.meta-systems.eu/nickbrown/blog/dixson/Dixson%20et%20al.%202014%20Raw%20Data%20.xlsx">here</a> [XLSX]) in the first place.</p><p>As I note in the report, just for the observations of the behaviours of 15 different species of fish (the first 15 of the 19 worksheets in the Excel file) the researchers must have made 864,000 separate notes. That is, the real "Raw Data" (this phrase appears in the title of the Excel file, but those are not the "raw data") consist of 864,000 entries corresponding to the position, in one of two possible channels, of 20 examples of 15 species of fish captured from 6 locations being recorded in 10 samples of water over 2 sets of trials of 2 minutes each with 12 observations per minute (20 x 15 x 6 x 10 x 2 x 2 x 12 = 864,000). That's almost a million ones and zeroes, each labelled with a species, fish number, capture location, water type, trial number, and sequence number of the observation.</p><p>Somewhere there must exist, or at the very least have existed, a CSV file (or, perhaps, a file in some other proprietary format, such as SPSS or SAS or Stata; but as far as I know, all of those packages can export to CSV format) containing those raw numbers. Even if the 864,000 observations were initially made on a very very very large stack of paper, at some point they would have been entered into a computer in a format from which the analyses reported in the article could have been run. Importantly, the analyses could almost certainly not have been run directly from this Excel file, because of the inconsistencies that it contains. Indeed, when I wrote some code (available <a href="http://www.meta-systems.eu/nickbrown/blog/dixson/Dixson2014.R">here</a>) to try to extract some summary statistics from the dataset, I had to explicitly work around the errors in the data, such as the cases where there are 21 rather than 20 fish in a set of tests, or where data elements are in different positions from one sheet to the next. Had the original analyses been based on these Excel sheets, the authors would surely have noticed that these misalignments were causing strange results or even crashes, and fixed the dataset. (And if by some chance they didn't notice these problems, there would be some inconsistencies in the published results, whereas—as the group of investigators led by Timothy Clark has pointed out—the results in this paper, and indeed across multiple studies from this laboratory, are remarkably uniform.)</p><p><span style="text-indent: -48px;">As Martin's article mentions, several other datasets (all Excel files) from the same laboratory seem to have similar problems. There seems to be a consistent pattern of the researchers deciding that in order to share their data, rather than just making their CSV file available with a few notes to explain the purpose and/or labels of each variable, they needed to laboriously re-enter their data into an Excel sheet, with lots of needless formatting whose only effects are to (a) increase the chance of errors and (b) make it harder for anyone to replicate their analyses in software. Meanwhile, the original raw files from which the statistics and charts in the published articles were made remain mysteriously absent. It is difficult to understand why anybody would work this way, when simply sharing the actual raw data would represent less effort and be much more reliable.</span></p><p><span style="text-indent: -48px;">[2021-05-10 20:58 UTC: Updated link to my analysis report with a new version. ]</span></p><p><span style="text-indent: -48px;"><br /></span></p></div>Nick Brownhttp://www.blogger.com/profile/07481236547943428014noreply@blogger.com4tag:blogger.com,1999:blog-7890764972166411105.post-13164777162673064482021-01-27T19:48:00.001+01:002021-01-27T19:48:16.869+01:00Why I blog about apparent problems in science<div class="MsoNormal" style="line-height: 24px;">In this post I want to discuss why I blog directly about what I see as errors or other problems in scientific articles. I had the idea to write this some time ago, and indeed some of the sentences below have been sitting in my drafts folder for quite a while, but the discussions on Twitter about my <a href="https://steamtraen.blogspot.com/2021/01/some-problems-in-high-profile-study-of.html" target="_blank">most recent post</a> have prodded me to finally write this up. (However, I don't go into that post or those Twitter discussions further here.)</div><div class="MsoNormal" style="line-height: 24px;"><br /></div><div class="MsoNormal" style="line-height: 24px;">I have seen criticism of the "blog first" approach because it "drops stuff on the authors out of the blue" or "doesn't give them a right to defend themselves". People have suggested that it would be better to approach the authors first and discuss the problems. That seems obvious, and it was how I used to approach things too, but over time I have <span style="background-color: white;">changed my mind, for a <span>couple </span>of</span> reasons.</div><div class="MsoNormal" style="line-height: 24px;"><br /></div><div class="MsoNormal" style="line-height: 24px;">First, I believe that, in principle, science should be conducted with radical transparency. Subject only to the need to protect participants, all review should take place in public, with all code and data fully open. Currently only a few journals (e.g., <i><a href="https://open.lnu.se/index.php/metapsychology/about" target="_blank">Meta-Psychology</a></i>) offer this, but open review and commenting, at least, is part of the deal with most preprint servers. The whole reason for posting a preprint is to allow direct feedback on it, which anyone can take part in. In contrast, in order to comment on published articles in journals, with a few exceptions such as <i>PLOS One</i> (which allows <a href="https://journals.plos.org/plosone/article/comments?id=10.1371/journal.pone.0160565" target="_blank">informal comments to be posted on articles</a> as well as formal comments that go through peer review), the choices are blogging/tweeting, PubPeer, or trying to fit a letter to the editor into some bizarre word count limit. (Some journals refuse to entertain any discussion of their articles unless it arrives <a href="http://comments.amnat.org/2021/01/registering-complaints-or-concerns.html" target="_blank">via the manuscript submission system</a>.)</div><div class="MsoNormal" style="line-height: 24px;"><br /></div><div class="MsoNormal" style="line-height: 24px;">So if I publish an unannounced blog describing what I see as issues in an article or a body of work, what I'm doing, in effect, is bringing the rules of Preprint World™ to bear. That might seem unfair, given that the authors have already "run the gauntlet" of peer review. (In some cases they may even have actually published a preprint first, but (spoiler alert) unless your preprint is about a politically hot topic, it isn't going to get much feedback, because we're all too busy with our own stuff.) But peer review is utterly broken, especially in its present mostly-secret for<span style="background-color: white;">m, which allows bad stuff to get published (through various forms of cronyism, as well as the limitations of editors and reviewers) and keeps critical voices out. A few journals now publish reviews alongside accepted articles, which is a first step along the road, but this strikes me as hugely insufficient because we don't get to see what happened to the manuscripts that were rejected.</span></div><div class="MsoNormal" style="line-height: 24px;"><span style="background-color: white;"><br /></span></div><div class="MsoNormal" style="line-height: 24px;"><span style="background-color: white;">(I should also acknowledge here that I am in a very unusual and privileged position. I am retired and don't have to keep anyone happy; nor, unlike if I were an emeritus professor, do I have a large list of buddies going back to my time in grad school to whom I feel some kind of obligation of loyalty.) </span></div><div class="MsoNormal" style="line-height: 24px;"><br /></div><div class="MsoNormal" style="line-height: 24px;">The second reason is much more personal. If I write to an author and say "Umm, I think I've found these problems in your article", it feels to me as if I'm entering into a process of negotiation with them. I worry that maybe it feels to them like there is something they can say or do that will persuade me not to share what I've found. Maybe they feel blackmailed. Maybe they will try and negotiate: say, to address three problems if I will "let them off" the fourth ("We didn't feel we had a choice to collect the data any other way, the postdoc had left and the grant money had run out").</div><div class="MsoNormal" style="line-height: 24px;"><br /></div><div class="MsoNormal" style="line-height: 24px;">I really hate that feeling.</div><div class="MsoNormal" style="line-height: 24px;"><br /></div><div class="MsoNormal" style="line-height: 24px;">Some people seem to have no problem with that kind of implicit, low-level conflict, but it really doesn't sit well with me. Perhaps this is irrational, but it's how my personal sense of embarrassment works, and I don't think that's about to change. (I'm not usually a fan of psychoanalytic approaches, but for what it's worth, I'm pretty sure that my relationship with my late mother is indeed involved here.)</div><div class="MsoNormal" style="line-height: 24px;"><br /></div><div class="MsoNormal" style="line-height: 24px;">Of course, this approach has disadvantages. Sometimes it can lead to pointing out "problems" that aren't actual problems, because I didn't understand something. You also risk sounding like a crank, which James Heathers and I wrote <a href="https://medium.com/@jamesheathers/how-not-to-be-a-crank-819103800502" target="_blank">here</a> about trying to not to be. You can mitigate this, for example by checking your analyses with multiple other people, but in practice even your closest colleagues don't always have time to go through the boring detailed stuff (and some of it is really, really boring). On occasions it gets you a nastygram from the authors, who in one case complained to my dean that I was violating their human rights [<i>sic</i>] by citing an e-mail of theirs verbatim. When I replaced the verbatim text with a paraphrased version, they then complained that I had misrepresented what they had written. (Another small benefit of not having corresponded with the authors is that you avoid the question of how to cite them.)</div><div class="MsoNormal" style="line-height: 24px;"><br /></div><div class="MsoNormal" style="line-height: 24px;">I think this is a consequence of the unique nature of science as a human activity. In an ideal world, science would not be conducted by social beings at all. Mr Spock doesn't mind if you call out apparent errors in his regression tables (although presumably he doesn't make many errors). But we don't live in that world. We all have reputations to consider, and we all like to be evaluated positively. (Aside: I thoroughly recommend <i><a href="https://www.bloomsbury.com/uk/judged-9781350113169/" target="_blank">Judged: The Value of Being Misunderstood</a></i> by Ziyad Marar, although I wish the author hadn't used quite so much recently-discredited social psychology to support his arguments.) So any kind of scientific criticism is likely to be orthogonal to our usual ways of rubbing along in polite society.</div><div class="MsoNormal" style="line-height: 24px;"><br /></div><div class="MsoNormal" style="line-height: 24px;">In fact, to me it feels even more rude/disloyal/distasteful to blog about an issue if I have been discussing it cordially (up to a point) with the authors. If you're having a "Dear Jane", "Hi Nick", "Best regards" kind of exchange of e-mails, and at some point you realise that something is badly wrong, what do you do? In some legal systems a lawyer can <a href="https://www.citizensinformation.ie/en/justice/witnesses/hostile_witness.html" target="_blank">request the judge to allow them to treat a witness as "hostile"</a> based on their responses during a trial, but lawyers are trained (at least, I hope it doesn't come naturally) to disconnect their embarrassment, they get to walk away at the end of the case, and all of this is taking place in front of others who can see why they are asking.</div><div class="MsoNormal" style="line-height: 24px;"><br /></div><div class="MsoNormal" style="line-height: 24px;">This problem is even more awkward when you realise that in some cases you may be helping the author to construct an alibi ("Gosh, do 80% of our reaction times really end in 7? Ah yes, I remember now, we made some fake data to test the code, ha ha, yes, <a href="https://finance.yahoo.com/news/cycling-officials-world-championships-theyre-185439942.html" target="_blank">we must have used that by mistake</a>, but hey, I've looked and just found the real data here, I'll write up the correction this afternoon, thanks for your keen observational skills"). Indeed, just this week we have heard <a href="https://crystalprisonzone.blogspot.com/2021/01/i-tried-to-report-scientific-misconduct.html" target="_blank">this story from Joe Hilgard</a> of how, whenever he pointed out a specific problem in the implausibly prolific output of one particular researcher, the next (equally implausible) article from the same person didn't have that particular problem in it. If this happened in a court of law there would be someone in overall charge who could take into account that the story keeps changing, but in science it seems you can get a lot of do-overs. In one case with which I was involved, when it was pointed out to the authors—two of whom were PIs with multiple R01 grants between them—that a coding error in their dataset meant that all of their regressions were uninterpretable (this was in the supplement of our critique, as it was only about the fourth worst thing about the original article), they merely uploaded a corrected version of the data without issuing a correction or indeed telling anyone at all. This meant that anyone who tried to reproduce the problem that I had discovered would now be unable to, but it also meant that re-running their code substantially changed all of their results. I wrote to the journal and the federal office of research integrity, but nothing happened.</div><div class="MsoNormal" style="line-height: 24px;"><br /></div><div class="MsoNormal" style="line-height: 24px;">Now, I'm aware that a risk of this approach is that it could end up turning me into some kind of solipsistic critic of science, the kind of person who has "Independent researcher" in their bio(*) and has been writing about the same one or two issues, and little else, for the last 10 years. The haughty drive-by posts that dominate sites devoted to "skepticism" in those fields of science that attract a lot of, er, <i>enthusiastic amateur investigators</i> also often seem to be based on an attitude of "I don't care what the authors have to say, here's why they're wrong" (although the one time I came close to having my own work featured on such a site, the potential author first sent me what appeared to be a rather crude form of blackmail note; I ignored it and as far as I know he never wrote up whatever perceived hypocrisy on my part that he was threatening to expose to the world). And indeed it is not hard to see parallels between a hardcore insistence on "science should be about objective truth" and the more juvenile kinds of libertarianism. Science should indeed be dispassionate, but it's still possible to be a dick about it. I don't want to be one of those perpetually disagreeable relatives that most of us have who are proud of proclaiming that "I say what I think, and I'm entitled to my opinion".</div><div class="MsoNormal" style="line-height: 24px;"><br /></div><div class="MsoNormal" style="line-height: 24px;">So there are limits to how one can go about this process in a reasonable way. I try to keep my language restrained; as James Heather<span style="background-color: white;">s is <span><a href="https://medium.com/@jamesheathers/fraud-aint-the-game-aae4b8ca1784" target="_blank">fond of repeating</a></span>, we can usually only talk in terms of error, because determining intent is hard and ultimately requires knowledge of s</span>omeone's mental states, a method for which has so far—perhaps ironically—eluded psychologists. (The justice system sometimes has to do it, but even then it can result in strange effects; Ziyad Marar's book, mentioned above, has a nice section on this.) Before I go public with something I generally get several other people to take a look at it and see what they think, and if anyone has strong doubts I will often leave a post at the draft stage out of caution.</div><div class="MsoNormal" style="line-height: 24px;"><br /></div><div class="MsoNormal" style="line-height: 24px;">However, a complex study or body of work can produce a lot of what might look like smoke without the need for any kind of major fire, so this is always going to be an imperfect process. I'm happy to correct my posts in as transparent a way as possible when they are shown to have been based on faulty assumptions or other errors, although I acknowledge that such a correction is always inferior to not getting things wrong in the first place<span style="background-color: white;">. But I think it's important for the issues themselves to be discussed in public; I hope that it keeps everyone honest, me first of all.</span></div><div class="MsoNormal" style="line-height: 24px;"><br /></div><div class="MsoNormal" style="line-height: 24px;"><br /></div><div class="MsoNormal" style="line-height: 24px;"><br /></div><div class="MsoNormal" style="line-height: 24px;">(*) Shout-out to a Twitter pal whose bio used to contain the words "Independent researcher (yes, I know...)"</div><div class="MsoNormal" style="line-height: 24px;"><br /></div>Nick Brownhttp://www.blogger.com/profile/07481236547943428014noreply@blogger.com4tag:blogger.com,1999:blog-7890764972166411105.post-33786269750447915522021-01-21T18:03:00.008+01:002021-01-27T14:44:58.906+01:00Some apparent problems in a high-profile study of ultra-processed vs unprocessed diets<div class="MsoNormal" style="line-height: 24px;"><div class="MsoNormal" style="line-height: 24px;"><div class="MsoNormal" style="line-height: 24px;"><div class="MsoNormal" style="line-height: 24px;"><div class="MsoNormal" style="line-height: 24px;">Update 2021-01-27 13:11 UTC: Added the word "apparent" to the post title. That should have been there before.</div><div class="MsoNormal" style="line-height: 24px;"><br /></div></div><div class="MsoNormal" style="line-height: 24px;">Update 2021-01-26 14:28 UTC: With the permission of Kevin Hall I have reproduced his responses in line below, in italics and bracketed by [KH... / ...KH]. I believe that these responses adequately address all of the points that I made in the original post.</div><div class="MsoNormal" style="line-height: 24px;"><br /></div></div><div class="MsoNormal" style="line-height: 24px;">Update 2021-01-22 23:21 UTC: I have received an extensive response from Kevin Hall, the lead author of the study under discussion here. This addresses the great majority of the points that I raised in this post. I will attempt to incorporate a version of those responses in a forthcoming update, but I wanted to get this acknowledgement of that response out there as soon as possible.</div><div class="MsoNormal" style="line-height: 24px;"><br /></div><div class="MsoNormal" style="line-height: 24px;">Update 2021-01-21 13:52 UTC: Clarified my use of the terms "processed" and "ultra-processed". In the first version I wrote "I will mostly follow the authors' use of the terms "processed" and "unprocessed" to distinguish between the two". This was sloppy on my part; the authors use "ultra-processed" consistently throughout the article. They mostly use "PROC" and "UNPROC" (or variations on those terms) in the data files, presumably for the easy visual contrast between the two, and that was what I wanted to convey. I also changed "processed" to "ultra-processed" in the title of the post.</div><div class="MsoNormal" style="line-height: 24px;"><br /></div><div class="MsoNormal" style="line-height: 24px;"><br /></div><div class="MsoNormal" style="line-height: 24px;">(Preamble: This post appears simultaneously with <a href="https://slimemoldtimemold.com/investigation-ultra-processed-diets-by-hall-et-al-2019" target="_blank">this post by Ethan and Sarah Ludwin-Peery</a>, who have some questions about patterns in the data associated with the article that is discussed. They got in touch with me via Ivan Oransky at Retraction Watch, who is also writing about this today. I recommend that you read their analysis first, not least because it provides a much more comprehensive introduction to the study. Here I discuss a variety of other apparent problems with the same article, which I found while Ethan and Sarah got on with sorting out the mystery of the daily weight changes. Although the contents of this post do not need a great deal of statistical training to follow, they are—as is often the case with post-publication data-based forensics—not exactly a thrill-a-minute ride, but I hope that I have made the implications of each set of points reasonably clear.)</div><div class="MsoNormal" style="line-height: 24px;"><br /></div></div><div class="MsoNormal" style="line-height: 24px;">This post looks at <a href="https://www.sciencedirect.com/science/article/pii/S1550413119302487" target="_blank">an article that first appeared in May 2019</a> describing a randomised controlled nutrition study. The authors claimed that people who were allowed to eat as much as they wished of a diet based on either "ultra-processed" or "unprocessed" food(*) consumed around 500 kcal/day more on the ultra-processed diet, and gained an average of 0.9 kg (2 lbs) in weight in two weeks, compared to people on the unprocessed diet, who lost an average of 0.9 kg in the same period. The same 20 participants ate both diets, in a randomised order. Importantly, the amount of macronutrients (protein, fat, and carbohydrates) provided in the meals was closely matched across diets, as was the number of calories offered (logically, since calories are a linear function of the macronutrients). That is, the principal claim of the study is that the mere fact that the food was ultra-processed, versus unprocessed, caused people to consume 500 kcal/day more and thus gain, rather than lose, weight in a controlled in-patient setting.</div></div><div class="MsoNormal" style="line-height: 24px;"><br /></div><div class="MsoNormal" style="line-height: 24px;">Perhaps not surprisingly, this study has attracted a lot of attention. It has already been cited <a href="https://scholar.google.com/scholar?bih=893&biw=1583&hl=en-GB&sxsrf=ALeKk00_aARpVMSkWkNdfZiuE7S4M-o71w:1610212948754&gs_lcp=CgZwc3ktYWIQAzIECAAQRzIECAAQRzIECAAQRzIECAAQRzIECAAQRzIECAAQRzIECAAQRzIECAAQR1DKgpsDWMqCmwNg74ebA2gAcAJ4AIABAIgBAJIBAJgBAKABAqABAaoBB2d3cy13aXrIAQjAAQE&uact=5&um=1&ie=UTF-8&lr&cites=1611177700195164799" target="_blank">more than 360 times</a> according to Google Scholar. The National Institutes of Health (NIH), which funded and conducted the study, put out <a href="https://www.nih.gov/news-events/news-releases/nih-study-finds-heavily-processed-foods-cause-overeating-weight-gain" target="_blank">an extensive news release</a> about it, and the story was covered by both <a href="https://www.sciencemag.org/news/2019/05/ultraprocessed-foods-may-make-you-eat-more-clinical-trial-suggests" target="_blank">Science</a> and <a href="https://www.nature.com/articles/d41586-019-01523-w" target="_blank">Nature</a>, as well as the <a href="https://www.bbc.com/news/health-48280772" target="_blank">BBC</a>, the <a href="https://www.theguardian.com/food/2020/feb/13/how-ultra-processed-food-took-over-your-shopping-basket-brazil-carlos-monteiro" target="_blank">Guardian</a>, the <a href="https://www.washingtonpost.com/lifestyle/wellness/its-trendy-to-scorn-processed-food-now-theres-research-to-back-up-that-attitude/2019/06/21/d19f54d8-929d-11e9-aadb-74e6b2b46f6a_story.html" target="_blank">Washington Post</a>, and many other major media outlets.</div><div class="MsoNormal" style="line-height: 24px;"><br /></div><div class="MsoNormal" style="line-height: 24px;">Here is the full APA-style reference of the article. For the first time since the appearance of the 7th edition of the APA Publication Manual (which says that we now have to list <a href="https://apastyle.apa.org/blog/more-than-20-authors">up to 20 authors’ names in a reference</a>) I'm actually going to need an ellipsis to omit some of the 25 authors:</div><div class="MsoNormal" style="line-height: 24px;"><br /></div><div class="MsoNormal" style="line-height: 24px; margin-left: 36pt; text-indent: -36pt;"><span style="mso-no-proof: yes;">H</span><span style="text-indent: 0px;">all, K. D., Ayuketah, A., Brychta, R., Cai, H., Cassimatis, T., Chen, K. Y., Chung, S. T., Costa, E., Courville, A., Darcey, V., Fletcher, L. A., Forde, C. G., Gharib, A. M., Guo, J., Howard, R., Joseph, P. V., McGehee, S., Ouwerkerk, R., Raisinger, K., ... Zhou1, M. (2019). Ultra-processed diets cause excess calorie intake and weight gain: An inpatient randomized controlled trial of ad libitum food intake. </span><i style="text-indent: 0px;">Cell Metabolism</i><span style="text-indent: 0px;">, </span><i style="text-indent: 0px;">30</i><span style="text-indent: 0px;">(1), 67–77. https://doi.org/10.1016/j.cmet.2019.05.008</span></div><div class="MsoNormal" style="line-height: 24px;"><br /></div><div class="MsoNormal" style="line-height: 24px;">The article is published on an Open Access basis; you can find the full text <a href="https://www.cell.com/action/showPdf?pii=S1550-4131%2819%2930248-7" target="_blank">here</a> (PDF, 2 MB) or a fuller version, including the Supplemental Information, <a href="https://www.cell.com/cell-metabolism/pdfExtended/S1550-4131(19)30248-7" target="_blank">here</a> (PDF, 23 MB). A small <a href="https://pubmed.ncbi.nlm.nih.gov/31269427/" target="_blank">erratum</a>, correcting a number of minor issues, was published on August 6, 2019; all of the issues mentioned in that erratum are already corrected in the PDF files, so you don't need to keep that to hand while reading the article.</div><div class="MsoNormal" style="line-height: 24px;"><br /></div><div class="MsoNormal" style="line-height: 24px;"><div class="MsoNormal" style="line-height: 24px;">[KH... <i>Importantly, another erratum was published in October 2020 and is available <a href="https://www.cell.com/cell-metabolism/fulltext/S1550-4131(20)30427-7" target="_blank">here</a>. The correction relates one of the questions raised below and we realize that the updated data and code were not yet deposited on the OSF website. We will do so.</i> ...KH]</div><div><br /></div></div><div class="MsoNormal" style="line-height: 24px;">This study has already been the subject of <a href="https://pubpeer.com/publications/757F8327D31C15B45D432F4B04ABEA" target="_blank">a comment on PubPeer</a> by Edward Archer, who is, I think it is fair to say, a prolific critic of the way that much nutritional research is carried out. I am not a nutrition scientist, so this blog post will mostly concentrate on the data and statistics of the study. I do have one or two small methodological questions too, but these are based only on my 60 years of experience of consuming food and 40 or so of preparing it, rather than any understanding of how nutrition studies are run.</div><div class="MsoNormal" style="line-height: 24px;"><p><b>The study</b></p><p>The authors recruited 20 volunteers, 10 male and 10 female, and kept them under observation for 28 days in an in-patient environment at the NIH Clinical Center in Bethesda, Maryland. The data show that between one and four people were in the facility at any point between the first admission on April 17, 2018 and the last recorded data collection on November 19, 2018.</p><p>Participants spent 14 days on each of two diets, one described as "ultra-processed" and the other as "unprocessed". The diets were presented on a 7-day rotation, so each participant ate the same meal twice, 7 days apart. Although the purpose of the study was to examine the effect of an "ultra-processed" diet, and that term tends to be used in nutrition science with a specific meaning that is different from "processed" (<a href="https://www.researchgate.net/publication/331462765_Ultra-Processed_Foods_Definitions_and_Policy_Issues" target="_blank">it's complicated</a>), I will use the terms "processed" and "unprocessed" to distinguish between the two, which I hope will avoid any confusion that might be caused by the fact that "ultra-processed" and "unprocessed" both start with the same letter. The participants were randomised to receive the processed diet first (N=10, 6 male, 4 female) or the unprocessed diet first (N=10, 4 male, 6 female); after 14 days on one diet they immediately switched to the other, as shown here.</p><div class="separator" style="clear: both; text-align: center;"><div class="separator" style="clear: both; text-align: center;"><a href="https://lh3.googleusercontent.com/-ltQWYW6OuQI/YASd-q4pGCI/AAAAAAAAFtM/2IzozIarhdQ2LMAMNQ55n61Y1ms_dPNSgCLcBGAsYHQ/image.png" style="margin-left: 1em; margin-right: 1em;"><img alt="" data-original-height="283" data-original-width="567" src="https://lh3.googleusercontent.com/-ltQWYW6OuQI/YASd-q4pGCI/AAAAAAAAFtM/2IzozIarhdQ2LMAMNQ55n61Y1ms_dPNSgCLcBGAsYHQ/s16000/image.png" /></a></div></div><div style="text-align: center;">Timeline of participants in the study. Reproduced from Figure 1 of Hall et al.'s article.</div><div style="text-align: left;"><br /></div><div style="text-align: left;"><br /></div><div style="text-align: left;">This study seems to have been a substantial undertaking. The participants spent 28 days in a highly controlled environment. The study was invasive, with subcutaneous sensors to monitor glucose levels as well as multiple finger stick blood testing operations daily. I like to imagine that the participants were handsomely compensated for taking a month out of their lives in the name of science; certainly the budget for the study must have run well into six figures.</div><p></p><p><b>The study code and data</b></p><p>The authors have made their data and SAS analysis code available in an OSF repository <a href="https://osf.io/rx6vm/" target="_blank">here</a>. There are two datasets, named ADLDataSAScode and ADLDataSAScode1, each in its own ZIP file. The only difference between these seems to be that ADLDataSAScode1, which was uploaded on August 20, 2019 (three months after the article was first published online, which was on May 16, 2019), contains one extra data file, and the code has been extended with a few lines to produce a table from that file (more on this later). All of the analyses in this post refer to the ADLDataSAScode1 dataset.</p><p></p><div class="separator" style="clear: both; text-align: center;"><a href="https://lh3.googleusercontent.com/-nyU9mfDssg8/YASotmrgp_I/AAAAAAAAFtc/e_8h9na9nkA4ZJN57btjS-gO2pWZqS7SQCLcBGAsYHQ/image.png" style="margin-left: 1em; margin-right: 1em;"><img alt="" data-original-height="864" data-original-width="1633" height="338" src="https://lh3.googleusercontent.com/-nyU9mfDssg8/YASotmrgp_I/AAAAAAAAFtc/e_8h9na9nkA4ZJN57btjS-gO2pWZqS7SQCLcBGAsYHQ/w640-h338/image.png" width="640" /></a></div><div style="text-align: center;">Screenshot of the timestamps of the OSF repository for the study. A full-size version of this image is available as part of the supporting files for this post (see "My code and data", below).</div><br /><p></p><p>The SAS code is not, as one might have hoped, a run-once script that generates all of the tables and figures from the article. Indeed, as supplied, the main script file (ADLDocumentation1.sas) produces two runtime errors at line 61 because the variables created within the SAS data file DLW at lines 42 and 43 are lost when this file is overwritten twice at lines 45 and 46. It seems that the code is best regarded as a collection of "building blocks" of code that can be run individually, possibly with minor modifications to use different subsets of the data. However, for completeness, I patched up the code so that it would run without error messages, and also to include both the original and adjusted analyses of the figures from Table 3D (see "The adjusted weight data", below), and ran it in SAS University edition. I have made the resulting code ("Nick-ADLDocumentation1.sas") and output ("(Annotated) Results_Nick-ADLDocumentation1.pdf") files available online (see "My code and data", below).</p><div class="MsoNormal" style="line-height: 24px;"><h4>The exact length of the study</h4></div><div class="MsoNormal" style="line-height: 24px;"><div class="MsoNormal" style="line-height: 24px;"><div class="MsoNormal" style="line-height: 24px;"><div class="MsoNormal" style="line-height: 24px;"><p>An issue that stands out immediately when one looks at any of the data files containing daily records is that there seems to be a <a href="https://en.wikipedia.org/wiki/Off-by-one_error#Fencepost_error" target="_blank">fencepost error</a>. Participants spent 14 days on each of two diets, with no break in between; their weight at the start of day 1 was the baseline for the first phase (processed or unprocessed diet, assigned at random), and their weight at the start of day 15 was the baseline for the second phase, when they received the other diet. It would seem, therefore, that they should have been weighed 29 times—once at the very start of the study, and then 28 more times after eating a day's worth of meals each time—but there are only 28 daily weight records for each participant. That is, we apparently do not know the effect on their weight of the last (14th) day of the second diet, because the last measurement of their weight on that second diet was apparently the one made on the morning of the 14th day (their 28th in the study), <i>before</i> they proceeded to eat their food and undergo whatever other measurements were performed on that day. This seems to make little sense, from the standpoint of either study design or ethics. Why feed your participants the controlled diet on the last day if you are not going to collect weight data from them relating to that day?</p><p>In fact this problem seems to be exacerbated because, as the data files <span style="font-family: courier;">deltabc</span> and <span style="font-family: courier;">deltabw</span> show, the difference in weight retained for each participant on <i>both</i> diets was the difference between their weights at the start of the first and 14th day on that diet. That is, even for the first diet that each person followed, their final weight was the weight at the start of the 14th day in the study, not that at the start of the 15th day; and the effect of the meals that they consumed on the 14th day of the study is also essentially disregarded.</p><p></p><div class="separator" style="clear: both; text-align: center;"><a href="https://lh3.googleusercontent.com/-w7X88xJ8Cl0/YAmGDvGUtQI/AAAAAAAAFu4/1dqifQh9Fd4P_Pbs90zBOgfWqZFdTCkFACLcBGAsYHQ/image.png" style="margin-left: 1em; margin-right: 1em;"><img alt="" data-original-height="470" data-original-width="479" src="https://lh3.googleusercontent.com/-w7X88xJ8Cl0/YAmGDvGUtQI/AAAAAAAAFu4/1dqifQh9Fd4P_Pbs90zBOgfWqZFdTCkFACLcBGAsYHQ/s16000/image.png" /></a></div><div class="separator" style="clear: both; text-align: center;"><br /></div><div class="separator" style="clear: both; text-align: center;"><a href="https://lh3.googleusercontent.com/-5G7jV0rttQU/YAmGHjy7qbI/AAAAAAAAFu8/rJnWS1EYdS4xweE-Ax4rif_ugqq2_tSYQCLcBGAsYHQ/image.png" style="margin-left: 1em; margin-right: 1em;"><img alt="" data-original-height="124" data-original-width="660" src="https://lh3.googleusercontent.com/-5G7jV0rttQU/YAmGHjy7qbI/AAAAAAAAFu8/rJnWS1EYdS4xweE-Ax4rif_ugqq2_tSYQCLcBGAsYHQ/s16000/image.png" /></a></div><div style="text-align: center;">Participant delta weights according to the data files <span style="font-family: courier;">deltabw</span> (top; an Excel filter has been applied to show only values near to the start and end of each diet period) and <span style="font-family: courier;">deltabc</span> (bottom). It can be seen that the retained weight change for each participant and each diet is the difference between their weight at the start of the 14th day on that diet and the start of the first day on that diet, apparently representing making a span of 13 rather than 14 days. The same pattern holds for every participant.</div><br /><p></p></div><div class="MsoNormal" style="line-height: 24px;">[KH... <i>Participants were admitted the afternoon before the study began. An overnight fasted body weight measurement was collected the next morning (day 1) which served as the fiducial point for the weight change calculations during the next 14 days on the first diet. On the morning of day 15, subjects were weighed which served as the fiducial point for weight change calculations on the alternate diet that was provided after an oral glucose tolerance test (OGTT). Fasted body weight measurements were then collected each morning including day 29 when the final OGTT was performed after which the subject was discharged. Thus, there were 29 fasted body weight measurements for each subject corresponding to the fiducial markers on days 1 and 15 prior to delivery of each diet and 14 days thereafter. However, the reported body weight changes in the manuscript correspond to days 1-14 of the first diet period and days 15-28 of the second diet period as shown in Figure 3A of the manuscript as described as the weight changes on each respective day on the diet. It would have been possible to report body weight changes corresponding to days 1-15 of the first diet period and days 15-29 of the second diet period, but we thought this would have been confusing to readers.</i> ...KH]</div></div></div></div><h4 style="text-align: left;">Which days did participants spend in the respiratory chamber?</h4><p>Participants spent one day per week in a respiratory chamber to enable their energy expenditure to be studied in detail. The article states that "On the chamber days, subjects were presented with identical meals within each diet period, and those meals were not offered on non-chamber days" (p. 72), which makes sense from an experimental control point of view, in that all participants would have consumed the same food on that day. The article's <a href="https://www.cell.com/cms/10.1016/j.cmet.2019.05.008/attachment/f7d43756-3f67-4557-8322-59a9d143d63c/mmc1.pdf" target="_blank">Supplemental Information</a> [PDF, 21MB] further states (on pp. 15, 16, 17, 37, 38, and 39) that the chamber day was day 5 of each weekly meal rotation, corresponding to days 5 and 12 of each participant's time on each diet.</p><p>However, the great majority of the records in the data file <span style="font-family: courier;">chamber</span> appear to contradict this. I looked for precise matches between the recorded energy intake on the chamber days and the records for each participant in the <span style="font-family: courier;">dailyintake</span> file, and found exactly one match for each participant and chamber day. Support for the idea that these matches are not coincidental is provided by the fact that the calendar dates of each record of the matched pairs (one in <span style="font-family: courier;">chamber</span> and one in <span style="font-family: courier;">dailyintake</span>) are identical. The matched records imply that of the 80 chamber days (20 participants x 2 diets x 2 chamber days per diet), only 7 took place on day 5 of the weekly meal rotation (whereas 2 were on day 1, 24 on day 3, 3 on day 4, 31 on day 6, and 13 on day 7). Furthermore, of the 40 pairs of chamber days within the same diet, 15 were on different meal rotation days within the pair (e.g., for participant ADL002 on the unprocessed diet, the chamber days were 3 and 8, corresponding to the third and first days of the meal rotation, respectively), meaning that the participant would have eaten different meals on their two chamber days for a given diet in 37.5% of cases. It is difficult to reconcile these records with the claims in the article and supplemental information.</p><p>[KH... <i>The article and supplement do not claim that “participants did indeed all spend days 5 and 12 of each diet in the chamber”. Rather, the main manuscript describes that participants spent one day each week in the respiratory chambers but does not specify the days of the week. The Supplementary Materials provide information about the rotating 7-day menu of meals provided on each diet and the chamber days were listed as occurring on day 5 of each week. This was not intended to indicate that the chamber days only occurred on day 5 but rather that the meals provided during the chamber days were prespecified and did not vary between subjects on the same diet no matter what day the chamber days occurred. The clinical protocol (available on the OSF website) indicates in Appendix A that the proposed schedule (page 34) had chamber days planned for days 3 and 10 on each diet. However, the protocol also notes on pages 13-14 that “Every effort will be made to adhere to the proposed timelines, but some flexibility is required for scheduling of other studies, unanticipated equipment maintenance, etc. Scheduling variations will not be reported.” Thus, while chamber days varied to accommodate such scheduling challenges, the meals provided on the chamber days were constant within each diet.</i> ...KH]</p><div class="MsoNormal" style="line-height: 24px;"><h4>Counting the calories</h4></div></div><div class="MsoNormal" style="line-height: 24px;"><div class="MsoNormal" style="line-height: 24px;"><p>The data file <span style="font-family: courier;">dailyintake</span> contains information about the amount of calories and individual nutrients consumed by the participants on each day. The total number of calories consumed is reported to two decimal places, but the individual readings of calories for protein, fat, and carbohydrates that sum to that total are reported to six decimal places, which on visual inspection do not appear to contain any regular patterns (which might correspond to, say, recurring decimals).</p><p></p><div class="separator" style="clear: both; text-align: center;"><div class="separator" style="clear: both; text-align: center;"><div class="separator" style="clear: both; text-align: center;"><div class="separator" style="clear: both; text-align: center;"><div class="separator" style="clear: both; text-align: center;"><div class="separator" style="clear: both; text-align: center;"><a href="https://lh3.googleusercontent.com/-mL1lQqs0Wjk/YASbibzUqXI/AAAAAAAAFsQ/dhJN9ZwbIG4tTl2DaJ3yYA8CTKJ1hVG8ACLcBGAsYHQ/image.png" style="margin-left: 1em; margin-right: 1em;"><img alt="" data-original-height="446" data-original-width="611" src="https://lh3.googleusercontent.com/-mL1lQqs0Wjk/YASbibzUqXI/AAAAAAAAFsQ/dhJN9ZwbIG4tTl2DaJ3yYA8CTKJ1hVG8ACLcBGAsYHQ/s16000/image.png" /></a></div></div></div></div></div></div><div style="text-align: center;">Extract from <span style="font-family: courier;">dailyintake</span> file, showing six digits of precision for macronutrient calorie counts. Some columns have been reduced to zero width to enable the image to fit on this web page.</div><div style="text-align: center;"><br /></div><p></p><p>It is not clear how such numbers could have been generated, however, as the process for calculating the amount of calories consumed presumably ought to have been a fairly simple multiplicative one, based on estimates of the numbers of grams of protein, fat, and carbohydrates in the uneaten portions of each food that was offered, after deducting an estimate of the amount of water. (<a href="https://www.blogger.com/blog/post/edit/7890764972166411105/3378626975044791552#" target="_blank">Edward Archer's comment on PubPeer</a> mentions this issue, and suggests that using a bomb calorimeter might have been a better way to measure energy intake, although this doesn't seem to address the split into macronutrient types.) The authors report that the diets were designed and analyzed using ProNutra software, made by Viocare of Princeton, NJ. I wrote to Viocare to ask how this software calculate calories from macronutrients—for example, whether it uses the <a href="https://en.wikipedia.org/wiki/Atwater_system">Atwater values</a> of 4.0 kcal/g for protein and carbohydrates and 9.0 kcal/g for fat, and whether it typically generates long mantissas in its output. Its founder and president, Rick Weiss, sent me this reply: </p></div></div><blockquote style="border: none; margin: 0px 0px 0px 40px; padding: 0px;"><div class="MsoNormal" style="line-height: 24px;"><div class="MsoNormal" style="line-height: 24px;"><p style="text-align: left;">ProNutra’s standard nutritional database is from USDA which we load into ProNutra with the resolution as USDA provides. Typically a research group using ProNutra would round off to the decimal place that they need. So I agree, seeing a value to the 6th decimal doesn’t make sense. The analysis of calories from macronutrients does use Atwater values.</p></div></div></blockquote><div class="MsoNormal" style="line-height: 24px;"><div class="MsoNormal" style="line-height: 24px;"><p>[KH... <i>More specifically, ProNutra uses specific Atwater factors which can deviate from the general values of 4.0 kcal/g for protein and carbohydrates and 9.0 kcal/g for fat. Therefore, the assumption immediately below is invalid.</i> ...KH]</p><p>But if the calories per gram are always integers, the presence of six decimal places of precision in the macronutrient information of every meal would seem to imply that the authors calculated the amount of food that was (a) served and (b) remained uneaten to the nearest microgram, which would require rather a lot of effort.</p><p>[KH... <i>The six decimal points for the macronutrient kcals in the data files are easily explained. The data for the total energy consumed and the percentage from each macronutrient were provided to 2 decimal places. For example, 15.68% of energy consumed as protein and a total energy intake of 2003.47 kcal. Therefore, the kcal provided from protein was calculated to six decimal places in the data file as follows: 2003.47*0.1568 = 314.144096 kcal from protein. </i>...KH]</p><p>I also wonder what was done in the case of processed snacks, where one would expect the authors to have simply used the nutrition information provided by the manufacturers.</p><p>[KH... <i>The assumption that we used manufacturer provided nutrition information is not correct. As indicated in the manuscript, nutrient information was obtained from the USDA standard reference databases or if an item was not found in that database, we pulled from the Food and Nutrition Database for Dietary Studies, (also through the USDA).</i> ...KH]</p><p>For example, on four days of the processed diet, three participants (ADL006 on days 3 and 4, ADL007 on day 8, and ADL015 on day 9) are recorded in the data file <span style="font-family: courier;">intakebymeal</span> as having consumed 403.14 kcal in snacks, with 42.007956, 202.218222, and 158.933010 kcal coming from protein, fat, and carbohydrates respectively (these amounts are precisely identical on all four days). The chances that three people left exactly the same amount of snack food unfinished on a total of four occasions would seem to be negligible, so this duplication presumably corresponds to these participants having completely finished the contents of the same combination of snack packages on each day. But the nutrition information for each of these packaged snacks reports the amount of macronutrients with a precision of 1 g, so the calories from each of these macronutrients ought also to be an integer (a multiple of 4 or 8), unless the authors perhaps contacted the manufacturers and obtained analyses down to the microgram level.</p><p></p><div class="separator" style="clear: both; text-align: center;"><div class="separator" style="clear: both; text-align: center;"><a href="https://lh3.googleusercontent.com/-4_M08UarYdc/YASb6W4matI/AAAAAAAAFsY/Zm6SV_jUeVobia3rPxF0D6ltKtrEEP35wCLcBGAsYHQ/image.png" style="margin-left: 1em; margin-right: 1em;"><img alt="" data-original-height="185" data-original-width="743" src="https://lh3.googleusercontent.com/-4_M08UarYdc/YASb6W4matI/AAAAAAAAFsY/Zm6SV_jUeVobia3rPxF0D6ltKtrEEP35wCLcBGAsYHQ/s16000/image.png" /></a></div></div><div style="text-align: center;">Three different participants, four different days, identical snack consumption.</div><div style="text-align: left;"><br /></div><div style="text-align: left;">[KH... <i>Indeed, ADL006 consumed the same snack items on days 3 and 4 as did ADL007 on day 8 and ADL015 day 9. From the mass consumed (grams), the subjects did finish the entire package of the snacks (28g, 39 g and 113 g for peanuts, cheese & peanut butter crackers, and applesauce, respectively). As explained above, we did not use manufacturer provided nutrition information, but rather nutrition information from the USDA database. Specific Atwater factors were used for the applesauce and the peanuts, whereas general Atwater factors were used for the cheese & peanut butter crackers. As also explained above, the six decimal points in the reported macronutrient kcals resulted from multiplying the macronutrient percentages by the total energy consumed.</i> ...KH]</div><p></p><p>A further problem here is that these records show that the three participants in question consumed more calories in the form of fat versus carbohydrates from their snacking on these four days, but substantially fewer calories from protein versus carbohydrates. The only processed snack in the image on p. 24 of the Supplemental Information that has more calories from fat than from carbohydrates is the 28 g package of Planters salted peanuts (see my file <span style="font-family: courier;">snacks.xls</span>), but this also has more calories from protein than from carbohydrates. I have not been able to identify any combination of packaged snacks that would get even close to the proportions of calories from protein, fat, and carbohydrates that is reported for these four participants, especially given the presumed constraint of counting only entire packages.</p><p>[KH... <i>The combination of foods that result in these proportions of calories from protein, fat, and carbohydrates was indicated above: 28g, 39 g and 113 g for peanuts, cheese & peanut butter crackers, and applesauce, respectively. </i></p><p><i></i></p><div class="separator" style="clear: both; text-align: center;"><i><a href="https://lh3.googleusercontent.com/-ZwzoTfgV5tE/YBAkHrcDwGI/AAAAAAAAFv4/y-6gUc9fz_w8WtVq6SBEHjIpiYeQvRFJQCLcBGAsYHQ/image.png" style="margin-left: 1em; margin-right: 1em;"><img alt="" data-original-height="245" data-original-width="2500" height="63" src="https://lh3.googleusercontent.com/-ZwzoTfgV5tE/YBAkHrcDwGI/AAAAAAAAFv4/y-6gUc9fz_w8WtVq6SBEHjIpiYeQvRFJQCLcBGAsYHQ/w640-h63/image.png" width="640" /></a></i></div><i><br /></i><i>As an approximate calculation using general Atwater factors, we have:</i><p></p><p></p><ul style="text-align: left;"><li><i>Peanuts 28 g providing 163.8 kcal, 6.63 g protein, 13.9 g fat, 6.02 g carbohydrates</i></li><li><i>Cheese & Peanut butter crackers 39 g providing 191.88 kcal, 4.21 g protein, 9.55 g fat, 23.01 g carbohydrates</i></li><li><i>Applesauce 113 g providing 47.46 kcal, 0.19 g protein, 0.11 g fat, 12.74 g carbohydrates</i></li></ul><p></p><p><i>When summed, these snacks provide 403.1 kcal, 11.03 g protein (44.12 kcal using general Atwater factor), 23.56 g fat (212.04 kcal using general Atwater factor), 41.77 gm carbohydrates (167.08 kcal using general Atwater factor). Thus, most of the total calories come from fat, followed by carbs, and then protein.</i> ...KH]</p><div>[Nick: Aaaarggggghhh. When preparing the spreadsheet that I used to try and determine a possible combination of snacks, I somehow entered 2 kcal/g instead of 9 kcal/g for fat. <homer_Doh!.gif > When I correct this, even using the manufacturers' approximate nutrition information, the combination leading to 403 pops right out at me. Apologies for my incompetence on this point. ]</div><p></p><div class="separator" style="clear: both; text-align: center;"><a href="https://lh3.googleusercontent.com/-s8MUXiXC93g/X_-LfeuHoFI/AAAAAAAAFpw/kAH2akb7RvoND8ptx1xWXr-ZhpGkcapwgCLcBGAsYHQ/image.png" style="margin-left: 1em; margin-right: 1em;"><img alt="" data-original-height="629" data-original-width="321" src="https://lh3.googleusercontent.com/-s8MUXiXC93g/X_-LfeuHoFI/AAAAAAAAFpw/kAH2akb7RvoND8ptx1xWXr-ZhpGkcapwgCLcBGAsYHQ/s16000/image.png" /></a></div><div style="text-align: center;">Nutrition information for Planters salted peanuts snack package (<a href="https://www.kraftheinz-foodservice.com/product/10029000076928/PLANTERS-Salted-Peanuts-1-oz-Single-Serve-Bags-Pack-of-144?categoryid=0000070262" target="_blank">source</a>), showing total grams of protein, fat, and carbohydrate. The corresponding calorie amounts would be protein, 7 x 4 = 28 kcal; fat, 14 x 9 = 126 kcal; carbohydrates, 5 x 4 = 20 kcal.</div><div style="text-align: left;"><br /></div><p></p><h4>The participants</h4><p>Participants are identified in the data by sequentially numbered labels from ADL001 through ADL021. That represents a span of 21 unique values, but there are no records with the label ADL011. Whether this is due to an error in assigning a label or a participant dropping out is not clear; however, there is no mention in the article of anyone dropping out of the study.</p><p>[KH... <i>ADL011 declined to participate in the study after their successful screening visit when they were assigned their subject number. No participants dropped out or were withdrawn from the study after admission.</i> ...KH]</p><p>Participant ADL006 (male) had a baseline BMI of 18.050 kg/m², which is below the minimum specified in the inclusion criteria on pp. e1–e2 of the article (18.5 kg/m²). That is, on the authors' own terms it seems that he ought to have been excluded from the study.</p><p>[KH... <i>This participant met inclusion criteria at their screening visit, but their starting BMI was lower once admitted for the study.</i> ...KH]</p><p>Participant ADL020 (female) had a baseline BMI of 26.853. During her 14 days on the unprocessed diet she consumed an average of just 836 kcal/day and lost a total of 4.3 kg (9.4 lbs) in <span style="background-color: white;">weig<span>ht, accounting on her own for nearly a quarter (23.7%) of the total weight loss of the sample on the unprocessed diet. </span></span>On day 12 of the same diet she obtained 22% of her calories (128 kcal out of 578 kcal total) from carbohydrates, which was the lowest daily percentage of any participant on any day on either diet in the entire study, whereas on the next day, day 13, she obtained 62% of her energy intake (602 kcal out of 962 kcal total) calories from carbohydrates, which was the <i>highest</i> daily percentage of any participant on any day on either diet in the entire study. This combination of extraordinary weight loss, very low levels of energy intake, and highly variable eating patterns make me wonder how much we can generalise from this participant to a broader understanding of the effects of different types of diet on the wider population. It seems to me that some kind of <a href="https://en.wikipedia.org/wiki/Hawthorne_effect" target="_blank">Hawthorne-type effect</a> may have been present here.</p><p>[KH... <i>The limitations of our study regarding generalizability were discussed in the manuscript. It is well-known in human nutrition research that individual subjects have large day-to-day diet variability and that there is large individual variability in weight loss.</i> ...KH]</p><h4>Errors in the data for individual participants</h4><div style="text-align: left;"><u>ADL002</u></div><div style="text-align: left;"><u><br /></u></div><div>The data file <span style="font-family: courier;">intakebymeal</span> contains one record for every meal consumed by participants during the study (breakfast, lunch, dinner, and one record for all of the snacks that they took) containing an assortment of nutritional information about that meal, including the type of diet that the participant was following on that day (and, hence, at each meal). For participant ADL002, however, something strange seems to have happened. The three meals (but not the snacks) that he consumed on days when he was on the processed diet are marked with the "unprocessed" diet flag, and vice versa, for all 14 days of each diet.</div><div><br /></div><div><div class="separator" style="clear: both; text-align: center;"><a href="https://lh3.googleusercontent.com/-FiAzgK6jbE4/YAS9nv5O-KI/AAAAAAAAFto/C7v3cwOy2poCiLvM2X0v-61BY9Sjx6C_gCLcBGAsYHQ/image.png" style="margin-left: 1em; margin-right: 1em;"><img alt="" data-original-height="251" data-original-width="500" src="https://lh3.googleusercontent.com/-FiAzgK6jbE4/YAS9nv5O-KI/AAAAAAAAFto/C7v3cwOy2poCiLvM2X0v-61BY9Sjx6C_gCLcBGAsYHQ/s16000/image.png" /></a></div><div style="text-align: center;">Extract from data file <span style="font-family: courier;">intakebymeal</span> showing that participant ADL002 apparently consumed unprocessed meals and processed snacks on the same day. <span style="text-align: center;">Some columns have been reduced to zero width to enable the image to fit on this web page.</span></div><br /></div><div><br /></div><div>It is not at all clear how this could have happened, because one would expect the data to have been recorded directly at the end of the day in question (either in a spreadsheet or directly into the ProNutra software) such that the type of diet would either have been completed automatically by the system, or obvious based on the records from the preceding day. Certainly one would expect the snacks for any given day to have the same diet code as the three meals. (I believe that the three meals have the wrong diet code and the snacks have the right one, rather than the reverse, based on the fact that the <span style="font-family: courier;">dailybw</span> and <span style="font-family: courier;">dailyintake</span> files both show ADL002 being on the processed diet for the first 14 days of the study and the unprocessed diet for the last 14 days, whereas <span style="font-family: courier;">intakebymeal</span> shows "unprocessed" as the diet for the breakfast, lunch, and dinner records for the first 14 days, and "processed" for the last 14 days.)</div><div><br /></div><div>[KH... <i>This error was previously discovered and an erratum was published in October of 2020 that corrected this error and is available <a href="https://www.cell.com/cell-metabolism/fulltext/S1550-4131(20)30427-7" target="_blank">here</a>. We realize that we have yet to update the files in the OSF website to correct this previously identified error and apologize for the delay.</i> ...KH]</div></div></div><div class="MsoNormal" style="line-height: 24px;"><div class="MsoNormal" style="line-height: 24px;"><div><br /></div><div style="text-align: left;"><u>ADL010</u></div><div><p></p><p>Participant ADL010 has a baseline (day 1, unprocessed diet) weight of 91.97 kg in the data file <span style="font-family: courier;">deltabw</span> but 93.17 kg in the data file <span style="font-family: courier;">baseline</span>. This affects, at least, the results shown in Table S1. If 91.97 kg is the correct weight then the Total mean for weight is correct but the Male mean (79.2 reported, 79.0 actual) and Male SE (6.6 reported, 6.5 actual) are not. If 93.17 kg is correct then the Male mean and SE are correct, but the Total mean (78.2 reported, 78.3 actual) isn't. I have not evaluated the effect of this discrepancy on the headline results of the study, but given that the total weight loss of all 20 participants on the unprocessed diet was 18.07 kg, a difference of 1.40 kg would seem to be potentially quite important.</p><p>ADL010's weight on day 2 (versus day 1) is recorded as 93.17 kg in <span style="font-family: courier;">deltabw</span>, so one possibility is that for this participant only, the copying process that generated the <span style="font-family: courier;">baseline</span> table somehow picked up the day 2 value rather than the day 1 value. Interestingly, according to that same file, this participant's weight fell back again to exactly 91.97 kg on day 3, which seems like quite a strong yo-yo effect.</p><div class="separator" style="clear: both; text-align: center;"><div class="separator" style="clear: both; text-align: center;"><div class="separator" style="clear: both; text-align: center;"><a href="https://lh3.googleusercontent.com/-CHQJ_jdRijc/YAlhVUb5O9I/AAAAAAAAFug/MORCZ5d_UsMLtrMKxIPZyBrf6w2aR_n-QCLcBGAsYHQ/image.png" style="margin-left: 1em; margin-right: 1em;"><img alt="" data-original-height="406" data-original-width="556" src="https://lh3.googleusercontent.com/-CHQJ_jdRijc/YAlhVUb5O9I/AAAAAAAAFug/MORCZ5d_UsMLtrMKxIPZyBrf6w2aR_n-QCLcBGAsYHQ/s16000/image.png" /></a></div><br /><div class="separator" style="clear: both; text-align: center;"><a href="https://lh3.googleusercontent.com/-dZID8ipYnVg/YAlhYFVDLTI/AAAAAAAAFuo/w0a1n8oU6pAldStinecq5i3KlvpZNHEnwCLcBGAsYHQ/image.png" style="margin-left: 1em; margin-right: 1em;"><img alt="" data-original-height="193" data-original-width="492" src="https://lh3.googleusercontent.com/-dZID8ipYnVg/YAlhYFVDLTI/AAAAAAAAFuo/w0a1n8oU6pAldStinecq5i3KlvpZNHEnwCLcBGAsYHQ/s16000/image.png" /></a></div></div></div><div style="text-align: center;">Weight of participant ADL010 in the data files <span style="font-family: courier;">baseline</span> (top) and <span style="font-family: courier;">deltabw</span> (bottom).<span>.</span></div><div><br /></div><div>[KH... <i>The baseline information in Table S1 contains body composition measurements obtained by DXA. All of the subjects except ADL001 and ADL010 had their first DXA measurement on day 1, but ADL001 and ADL010 were measured on day 2. For ADL001, their body weight measurements were the same on days 1 and 2, but ADL010 had different weights on these days. Therefore, the body weight measurement on day 2 for ADL010 was included in the baseline information to correctly correspond to the day of the DXA measurement.</i> ...KH]</div><div><br /></div><div>As with several other issues raised in this blog post, it is not clear how this discrepancy could have arisen with any kind of systematic processing of the study data from raw observations. If values were copied manually across the various data files, one wonders how many other transcription errors might be lurking.</div><h4 style="text-align: left;">Other oddities in the data</h4><div>As mentioned above, the data file <span style="font-family: courier;">intakebymeal</span> contains a record for each meal (plus snacks), with information such as macronutrient and total calories, free water consumption, the total mass of the food consumed, etc. Meanwhile, the data file <span style="font-family: courier;">dailyintake</span> has a record for each day's consumption for each participant, broken down similarly. One would therefore expect the values in the four records in <span style="font-family: courier;">intakebymeal</span> to sum to the values in the corresponding record in <span style="font-family: courier;">dailyintake</span>. Curiously, however, this is not the case. Indeed, while the energy intake (EI) field in <span style="font-family: courier;">dailyintake</span> matches the sum of the per-meal EI values in <span style="font-family: courier;">intakebymeal</span> to within 0.05 kcal in every case (once the diet code error for participant ADL002, discussed above, has been corrected), the calories for protein, fat, and carbohydrates from the four meal records each day frequently sum to a total that is some way from the equivalent values in the daily record.</div><div><br /></div><div><div class="separator" style="clear: both; text-align: center;"><div class="separator" style="clear: both; text-align: center;"><div class="separator" style="clear: both; text-align: center;"><a href="https://lh3.googleusercontent.com/-btUUSz2gRFg/YAScJub51dI/AAAAAAAAFsc/IPeVc05GZ3Eyi3uPUYSxAKUQWplt4pMywCLcBGAsYHQ/image.png" style="margin-left: 1em; margin-right: 1em;"><img alt="" data-original-height="148" data-original-width="695" src="https://lh3.googleusercontent.com/-btUUSz2gRFg/YAScJub51dI/AAAAAAAAFsc/IPeVc05GZ3Eyi3uPUYSxAKUQWplt4pMywCLcBGAsYHQ/s16000/image.png" /></a></div></div></div><br /></div><div><div class="separator" style="clear: both; text-align: center;"><div class="separator" style="clear: both; text-align: center;"><a href="https://lh3.googleusercontent.com/-BOhWKZlt8MI/YAScXo66IRI/AAAAAAAAFsk/V_A9r-hgL84QKNLDnpZUULNjOu384MqlwCLcBGAsYHQ/image.png" style="margin-left: 1em; margin-right: 1em;"><img alt="" data-original-height="55" data-original-width="600" src="https://lh3.googleusercontent.com/-BOhWKZlt8MI/YAScXo66IRI/AAAAAAAAFsk/V_A9r-hgL84QKNLDnpZUULNjOu384MqlwCLcBGAsYHQ/s16000/image.png" /></a></div></div><div style="text-align: center;">Per-meal (top, with sum for all four meals under "Total") and per-day intake for participant ADL001 on the first day of the processed diet. Note that while the total energy intake ("EI") from the meals is identical to within 0.01 kcal, the total for each of the macronutrients (protein, fat, and carbohydrates) is different by between 9 and 26 kcal. Some columns have been reduced to zero width to enable the image to fit on this web page.</div></div><div><br /></div><div>[KH... <i>We have been able to reproduce this problem using the data from several subjects and it appears to be an issue with the ProNutra software. We have contacted the manufacturer to identify the reason for the problem but have yet to receive a reply. However, we agree with the blogger that the magnitude of the discrepancy is very small (tens of calories) and we note that it does not affect the primary study outcome of total energy intake. This issue may be related to the next problem below. </i>...KH]</div><div><br /></div><div>A related problem is that, within <span style="font-family: courier;">intakebymeal</span>, the three macronutrient calorie observations for a meal frequently do not sum to the overall energy intake from that same meal. A spectacular example of this is the dinner of participant ADL005 on day 3 of the unprocessed diet, where the macronutrient calories sum to 1235.54 kcal, but whose total energy content is shown as 1720.21 kcal—a net discrepancy of about 484.67 kcal.</div><div><br /></div><div><div class="separator" style="clear: both; text-align: center;"><a href="https://lh3.googleusercontent.com/--kLMxRFj2N0/YAmQ2mrXkiI/AAAAAAAAFvM/dCmdfOPErtICY8OvY18JxrcMISbm2_l4gCLcBGAsYHQ/image.png" style="margin-left: 1em; margin-right: 1em;"><img alt="" data-original-height="95" data-original-width="684" src="https://lh3.googleusercontent.com/--kLMxRFj2N0/YAmQ2mrXkiI/AAAAAAAAFvM/dCmdfOPErtICY8OvY18JxrcMISbm2_l4gCLcBGAsYHQ/s16000/image.png" /></a></div><div><div style="text-align: center;">Per-meal total and per-macronutrient calories for participant ADL005 on day 3 of the unprocessed diet. Some columns have been reduced to zero width to enable the image to fit on this web page.</div></div><div><br /></div><br /></div><div>A total of 639 of the 2,240 participant x day x meal records in <span style="font-family: courier;">intakebymeal</span> suffer from this problem, whereas none of the records in <span style="font-family: courier;">dailyintake</span> do. Put simply, a large number of the per-meal macronutrient values in the <span style="font-family: courier;">intakebymeal</span> data file appear to be incorrect. Interestingly, all of these discrepancies are on the positive side—that is, when the reported overall energy intake differs substantially from the total of the energy intake from the macronutrients, the former is always larger— suggesting that whatever process is responsible for these discrepancies might not be entirely random.</div><div><br /></div><div><div>[KH... <i>We noticed this problem with the meal data (and not the daily data) when preparing our correction published in Cell Metabolism in October of 2020. We identified that this was an error in the ProNutra software that listed the fraction of calories coming from all three macronutrients as 0% while correctly providing a value for the total calories for the following food items: </i></div><div><i>Garlic, raw</i></div><div><i>Lemon juice, fresh squeezed</i></div><div><i>NutriSource Fiber</i></div><div><i>OLD FOODS- Oil, olive (Nina)</i></div><div><i>Oil, olive</i></div><div><i>Oil, olive (Nina)</i></div><div><i>Oranges, raw</i></div><div><i>Pepper, black (Monarch)</i></div><div><i>Salsa (del Pasado)</i></div><div><i>Tomatoes, raw</i></div><div><i><br /></i></div><div><i>We contacted the manufacturer of ProNutra at the time, but we have yet to receive a satisfactory explanation for this error. Nevertheless, we corrected these data in the erratum published in Cell Metabolism in October of 2020. We realize that we have yet to update the files in the OSF website to correct this previously identified error and apologize for the delay. </i>...KH]</div></div><h4 style="text-align: left;">The adjusted weight data</h4><div>I mentioned earlier that the OSF repository for the project contains two ZIP files. The second of these, uploaded after the article was published, includes an extra data file name<span style="background-color: white;">d <span style="font-family: courier;">deltabcadj14</span>, a</span>nd the SAS code has been extended with a few lines that analyse this file. This code seems to be quite important as it claims to generate the results for figure 3D of the article, which presents what are arguably the headline findings of the study: a mean weight gain of 0.9 kg per participant on the processed diet and a mean weight loss of 0.9 kg on the unprocessed diet. The code file contains this comment:</div><div><br /></div></div></div></div><blockquote style="border: none; margin: 0px 0px 0px 40px; padding: 0px;"><div class="MsoNormal" style="line-height: 24px;"><div class="MsoNormal" style="line-height: 24px;"><div><div><div style="text-align: left;">Update1: Body composition changes presented in Figure 3D are adjusted for 14 days because the body compositions were not measured exactly 14 days apart. In the previous version of SAS code and data, such adjustment was not provided. Here we have updated the SAS code at the section "data for figure 3D" and added a dataset "DeltaBCadj14"; </div></div></div></div></div></blockquote><div class="MsoNormal" style="line-height: 24px;"><div class="MsoNormal" style="line-height: 24px;"><div><div><br /></div><div>It is not clear what adjustments were performed to make this new data file. The extra code provided merely re-runs the comparisons of before/after weight, fat mass, and fat-free mass for the two types of diet, using the adjusted data. When the new code is run, it produces results for the mean weight loss and gain that are around 20% different from the originals; had these numbers been available when the article was published, the authors would presumably have reported a mean gain of 0.8 kg (to one decimal place) on the processed diet and a mean loss of 1.1 kg on the unprocessed diet.</div><div><div class="separator" style="clear: both; text-align: center;"><br /><div class="separator" style="clear: both; text-align: center;"><div class="separator" style="clear: both; text-align: center;"><a href="https://lh3.googleusercontent.com/-J8sIUIjqiiE/X_4rpCPCCbI/AAAAAAAAFo8/32sf1Kj053w3301ddpSrfN79092FxmnKgCLcBGAsYHQ/image.png" style="margin-left: 1em; margin-right: 1em;"><img alt="" data-original-height="214" data-original-width="632" src="https://lh3.googleusercontent.com/-J8sIUIjqiiE/X_4rpCPCCbI/AAAAAAAAFo8/32sf1Kj053w3301ddpSrfN79092FxmnKgCLcBGAsYHQ/s16000/image.png" /></a></div><br /></div><div class="separator" style="clear: both; text-align: center;"><div class="separator" style="clear: both; text-align: center;"><a href="https://lh3.googleusercontent.com/-YqjBI-afHMk/X_4rVqKvzcI/AAAAAAAAFos/nnTBL76YMqYhsEgUNj48q_GorMe5wKrWACLcBGAsYHQ/image.png" style="margin-left: 1em; margin-right: 1em;"><img alt="" data-original-height="220" data-original-width="605" src="https://lh3.googleusercontent.com/-YqjBI-afHMk/X_4rVqKvzcI/AAAAAAAAFos/nnTBL76YMqYhsEgUNj48q_GorMe5wKrWACLcBGAsYHQ/s16000/image.png" /></a></div>Comparison of pre- and post-study weights (first two lines of each panel, for the processed and unprocessed diets, respectively) and fat/non-fat mass, using the original (top) and adjusted (bottom) data. The output from the original data file contains a descriptive label for each line, which I have removed here to allow the figures in the tables to appear in the same size font for both images.<br /><div style="text-align: left;"><br /></div><div style="text-align: left;"><br /></div></div><div style="text-align: left;"><span style="background-color: white;">Interestingly, the sample size for fat mass and fat-free mass on the unprocessed diet is higher with the adjusted data than the original data. The data file <span style="font-family: courier;">deltabc</span> is missing these values for participant ADL002, whereas <span style="font-family: courier;">deltabcadj14</span> is not. Thus, whatever the adjustment process was, it seems to have extrapolated or interpolated in some way whatever data relating to fat mass might have been missing for this participant, such that he could now be included. (I assume that fat-free mass is calculated as weight minus fat mass, so that only one missing value needs to have been inferred in this way.)</span></div><div style="text-align: left;"><br /></div><div style="text-align: left;">I wonder if this adjustment might be an attempt to compensate for the issue that I raised earlier under the heading "The exact length of the study". But if that is the case, it is not clear why it would be necessary to adjust the values for both diets for each participant. After all, the start of day 15 of the study—the day on which the participants changed to the other diet—ought to correspond to exactly 14 days after they were weighed on day 1. (See also my section "The exact length of the study", above.)</div><div style="text-align: left;"><br /></div><div style="text-align: left;">The article states (p. e2) that participants were weighed at 6am every day. If it turns out that they were weighed substantially later on day 1 (or earlier on the last day), the question then arises of whether they skipped one or more meals on that day, although there are records for every scheduled meal in <span style="font-family: courier;">intakebymeal</span>. On the other hand, if they were weighed only an hour or so late, the adjustment hardly seems necessary, especially since the Welch Allyn Scale-Tronix 5702 weighing scale that was used for the study has a precision of only 0.1 kg (a fact that I confirmed by e-mail correspondence with the manufacturer; see also Ethan and Sarah's post, which explores the consequences of this constraint in more detail). The adjusted values are reported to 10 or more decimal places, which—assuming that the adjustment was indeed a function of the difference between the actual elapsed time from the first to last measurement, and exactly 14 days—suggests that the time at which participants' weight and fat mass was measured must have been recorded to a very high degree of precision indeed.</div><div style="text-align: left;"><br /></div><div style="text-align: left;">[KH... <i>The question about the precision of the body weight measurements is addressed in our response to the blog post by Ethan and Sarah Ludwin-Peery. These apparent high-precision body weight measurements and the statistical anomalies noted by Ethan and Sarah Ludwin-Peery are explained by subtracting pre-weighed pajamas worn during the body weight measurements as described in the manuscript Method Details section.</i> ...KH]</div><div style="text-align: left;"><br /></div><div style="text-align: left;">Two questions arise from this operation:</div><div style="text-align: left;"><ul style="text-align: left;"><li>First, it would be interesting to know what the adjustment process was. It seems to have been quite powerful, because some of the differences between the original and adjusted values are substantial. For example, for participant ADL014, the loss in weight on the unprocessed diet has been adjusted from 0.10 kg to 0.95 kg, and for ADL005 the equivalent loss has gone from 0.26 kg to 1.79 kg; participant ADL019's gain of 0.30 kg on the unprocessed diet has been adjusted to a loss of 0.24 kg, while participant ADL021's loss of 0.30 kg on the processed diet has been adjusted to a gain of 0.16 kg. These changes appear to affect principally the fat-free mass rather than the fat mass, which in numerous cases (8 out of 20 on the processed diet, 2 out of 19 on the unprocessed diet) is identical to two decimal places after adjustment. For example, participant ADL010's original weight gain of 3.60 kg on the processed diet becomes 2.69 kg in the adjusted file, but his fat mass did not change at all.</li><li>Second, if the authors believe that these adjusted figures provide a better estimate of the effects of the diets, one might wonder why they have not submitted a correction, updating the claims about weight loss that featured in the abstract of their article, rather than allowing this important new information to languish in an OSF repository. Otherwise it is not clear what the point of performing these "adjusted" analyses was.</li></ul><div>[KH... <i>The results in the published manuscript correspond to the unadjusted data and code that was originally deposited on the OSF website. The adjustments in the second file on the OSF website were performed to address the fact that the DXA body composition measurements were not performed on exactly at the same time points for all subjects. Furthermore, subject ADL002 was missing one DXA measurement during the unprocessed diet period. The adjusted data attempt to estimate the mean changes in body composition that would have occurred had the DXA measurements been aligned on day 14. To do this, we calculated the slope of the best fit regression line to the fat mass measurements over each diet period to estimate the fat mass change on day 14. The DXA measurement at the end of the first diet period was also used as the fiducial measurement for the start of the second diet period and subject ADL002 contributed only 2 fat mass measurements during the unprocessed diet period. The corresponding body weight measurements on those days were used to calculate the fat-free mass estimates by subtracting the estimated fat masses on those aligned days. This explains the minor differences between mean results reported in the original file deposited in OSF (which correspond to the results published in the manuscript) and the first updated file. The mean results are not materially different between these analyses, and the adjusted data merely address the potential criticism that the DXA measurements were not all conducted on the same days in all subjects. The reported data in the manuscript are not in error.</i> ...KH]</div></div></div></div></div><div class="MsoNormal" style="line-height: 24px;"><div class="MsoNormal" style="line-height: 24px;"><h4 style="text-align: left;">Conclusion</h4><div>Hall et al.'s article seems to have had a substantial impact on the field of nutrition research. However, both Ethan & Sarah's post and this one raise a number of concerning questions about the reliability of this study. There seem to be problems with the design, the data collection process, and the analyses. I only looked at about half of the 23 data files, so there may be other problems lurking. I hope that the authors and the editors of <i>Cell Metabolism</i> will take another look at this study and perhaps consider issuing a correction of some kind.</div><div><br /></div><div><div>[KH... <i>A correction was published in Cell Metabolism in October of 2020 and is available <a href="https://www.cell.com/cell-metabolism/fulltext/S1550-4131(20)30427-7" target="_blank">here</a>.</i></div><div><i>This correction regards an error described by the blogger that we previously independently discovered. Many of the other questions raised above are the result of misinterpretations of the data and the study. We hope that we have now clarified these issues. One remaining question appears to involve the ProNutra software used to calculate the individual macronutrient amounts, but the discrepancies are very small and do not affect the primary study outcome.</i> ...KH]</div></div><div><h4 style="text-align: left;">My code and data</h4><div>I have made my R analysis code, which reproduces most of the results reported above, <a href="http://nickbrown.fr/blog/proc-unproc" target="_blank">here</a>. Some of my results can probably best be checked by examining the data files in a spreadsheet, so my code also includes a loop (which you need to enable, following what I hope are clear instructions) that will export the original SAS data files to CSV format. Also included at the same location is a spreadsheet file named <span style="font-family: courier;">snacks.xls</span> which summarises the nutrition information for the snacks that were served on the processed diet, plus the OSF screenshot and the SAS code and results files mentioned earlier.</div><h4 style="text-align: left;">Acknowledgements</h4></div><div>Thanks to Andrew Althouse and James Heathers for help with the analyses, and to Ethan and Sarah Ludwin-Peery for sharing their discoveries about the Hall et al. article and some very interesting discussions about what it all might mean.</div><h4 style="text-align: left;">Note on copyright</h4><div>I believe that the reproduction of two images in this post (<span style="text-align: center;">Figure 1 of Hall et al.'s article and the Planters nutrition information label) constitute </span>fair use.</div><h4 style="text-align: left;">Footnotes</h4><p>(*) I have put these terms in quote marks to emphasise that they have a specific technical meaning. I don't know if that it a good idea, though; perhaps it looks like I am putting Dr Evil-style air quotes around them. That isn't my intention.</p><p><br /></p></div></div></div></div>Nick Brownhttp://www.blogger.com/profile/07481236547943428014noreply@blogger.com3tag:blogger.com,1999:blog-7890764972166411105.post-54468895867122525062020-12-06T22:22:00.000+01:002020-12-06T22:22:19.138+01:00Treating COVID-19 patients in the ICU: A doctor's point of view<p> A few days ago I posted this tweet:</p>
<blockquote class="twitter-tweet"><p dir="ltr" lang="en">Germany's COVID-19 cases peaked about 18 days ago, but their deaths are still going up very fast. That seems like a longer lag than we see elsewhere. <a href="https://t.co/E6aUztI2No">pic.twitter.com/E6aUztI2No</a></p>— Nick Brown (@sTeamTraen) <a href="https://twitter.com/sTeamTraen/status/1334279102182936576?ref_src=twsrc%5Etfw">December 2, 2020</a></blockquote> <script async="" charset="utf-8" src="https://platform.twitter.com/widgets.js"></script>
<p>In reply to that, I was contacted privately by “Dominique” (not their real name), a Twitter user whose profile says that they are an intensive care doctor in a country that has been quite badly affected by COVID-19. I have no reason to doubt that claim—indeed, it’s the entire basis of this post—but I haven’t attempted to formally verify it. I hope that regular readers of this blog will trust my judgment on this point; “false flag” theorists are invited to look away now.</p><p>Dominique’s initial contact was in response to my tweet, to discuss possible reasons why Germany’s numbers do not seem to be following the same pattern as other countries. But the conversation developed into a much wider-ranging discussion of the treatment of COVID-19 patients in the ICU. I found this fascinating, and thought it might be of interest (and perhaps even of professional use) to other people. So, with Dominique’s permission, I am converting this exchange of messages into a sort of interview, paraphrasing their responses a little and arranging them for context under “questions” that are intended to function as headings (they may not correspond to actual prompts that I gave during the exchange).</p><p>Dominique has read this post and approved it as a description of their thoughts. Everything here is written from Dominique’s point of view, unless it starts with “<i>Nick:</i>”.</p><p><i>Nick: Looking at my tweet, why do you think that Germany’s lag between cases and deaths seems to be so large?</i></p><p>One possible cause for the lag is that they wait for longer until withdrawing support. Waiting is somewhat futile in some cases, but we tend to wait longer if demand for ICU beds is not especially intense. That might not explain why the lag is so long, but it is a trend that we’ve noticed. But there might be other causes we don’t know about.</p><p>It could also be that German patients are more aggressively treated in general (the German system has plenty of resources), which could make it more likely that the elderly last longer, because they’re more likely to receive invasive organ support. On the other hand, maybe Germany has middle-aged patients attached to machines for a long time. We also wait longer in the younger patients, although it is futile in some cases.</p><p><i>Nick: I’ve also been wondering why Germany’s death rate is (or has been) so much lower than almost all its neighbours. <a href="https://en.wikipedia.org/wiki/List_of_countries_by_hospital_beds" target="_blank">This Wikipedia page</a> suggests that Germany has 3-4 times more ICU beds than many comparable countries. Could that be a factor?</i></p><p>Germany did well from the beginning. I have the feeling that it has consistently shown the most robust response in Europe. They allocated resources according to the size of the problem. For example, they realized— which is not obvious to most physicians who don’t work in ICU—that proning patients [[i.e., lying them on their fronts in an induced coma]] clearly signals a likely stay of more than three weeks, which potentially leads to shortages of ICU beds. Apart from COVID cases, most people typically leave ICU much more quickly, one way or another</p><p>No other country reacted like this as quickly. Also, existing public health and hospital plans were typically designed for a bad flu epidemic, which requires different treatment patterns. The virus was quickly understood from an epidemiological point of view, but to design containment measures they needed to also understand the clinical aspects of the disease, which are what determine exactly what stresses it places on the healthcare system. It seems that Germany has understood this.</p><p><i>Nick: What are the conditions under which you stop treatment in the ICU?</i></p><p>There comes a point after which there will be no tissue function improvement, but we have to wait until the inflammation is over to determine that. It’s a bit like when Notre-Dame cathedral caught fire; they had to wait for it to stop burning to be able to evaluate whether what was left was enough to rebuild on. If what’s left in a patient is not enough to support life outside of the mechanical ventilation setup, one of us will mention this in the morning round and then we make a group assessment with several doctors involved; this might be “let’s wait a few more days”. Such decisions are not evaluated by a single person: there’s a stable group of physicians that receive the relevant information and coordinate such decisions for ICU patients. </p><p>The process of actually switching off support after a provisional assessment to do so has been made can take a few days more, maybe even a week or two if we are waiting for other processes that might be ongoing (such as a bacterial sepsis or gut ischemia) to have an outcome, which gives us time to prepare the family for what might happen. In fact many deaths are “semi-programmed”, a result of withdrawal of treatment due to ethical reasons (for example, via Do Not Resuscitate notices); the ICU doctors don’t necessarily walk solemnly over to the machine and switch it off at a particular time. But today, for example, I have a withdrawal of invasive support in a (non-COVID) patient for whom the decision was made 36 hours ago, so I will be present for that one. They will be discharged from the ICU and put on palliative care.</p><p>End of life in ICU hasn’t been a particularly pressing issue in our region. Maybe in XXX (a region of the country with three times as many cases) they could give you some examples of ethical dilemmas with COVID. Fortunately we have been able to manage them like any other ARDS [[see below]].</p><p><i>Nick: People from Sweden have told me that (also before COVID) almost nobody in that country gets admitted to ICU, as it’s considered unethical. I suppose it’s a coincidence that it’s also cheaper for the system.</i></p><p>Maybe they offered only basic or intermediate care to the elderly. And the healthier among the elderly will generally survive for three weeks even if the final outcome is death.</p><p>It’s sensible to be parsimonious with ICU care. We should probably reinforce intermediate-level care instead, which is a lot less resource-intensive and probably equally effective for many COVID cases. As it is, we are often operating at the limits of ethicality. There are too many decisions to take in a day to get everything right.</p><p>Something worth noting is that patients of all ages who die in the ICU share similarities in their disease progression. The older ones are somewhat more numerous, but otherwise their treatment, and the withdrawal of that treatment if it happens, is actually very similar to that of younger people. In fact we have very few patients over 75. So the idea that elderly people can somehow be blamed for taking up ICU capacity is rather unfair. An older person who gets infected is more likely to need the ICU than a younger one, but once they are there, they take up a similar amount of resources.</p><p>Another unusual thing with COVID ICU patients is that if they are going to improve, this often doesn’t happen until the third week. Before that their signs tend to oscillate randomly up and down. So we sometimes don’t know how the patient will do until the third week from the onset of severe symptoms, when acute inflammation declines noticeably.</p><p><i>Nick: Can you say something about the effects that all this has on your own mental wellbeing?</i></p><p>Part of the stress of COVID for healthcare workers, apart from the heavy workload, is the mental toll of not seeing what’s ahead in the mist any better than our patients can. It makes you feel like you are a passenger in the car, rather than the driver at the bedside. But once we recognised this pattern and came to accept it, it was easier mentally.</p><p>We were also lucky due to our region of the country having a relatively low level of the disease in the first wave. But because of the lockdown at that time, a lot of surgery was cancelled. So we spent part of that time doing planning and logistics, setting up decontamination routes, and customising our PPE. The sense of control that we got from that was good for dealing with our stress.</p><p>In the bad times we worked double shifts apart from the 24-hour ones, so there was one more of us on duty. It means we all worked more hours, but it took less of a mental toll, and it only lasted for a few weeks. Today we are over the worst of the second wave. We maintained a sense of control, and remained united like a small army, which really helped. It was also really great when people sent in gifts, especially PPE. I can’t imagine how hard it must have been in XXX.</p><p><i>Nick: What makes a case unrecoverable?</i></p><p>In this sense, COVID cases are easy to call. We sometimes have ethical cases in other pathologies for which we need to work with the support or advice of the medical ethics committee, but for COVID it’s quite clear-cut. After a month or so, there’s not going to be a major change in inflammation or further improvement to any tissue damage. So we test to see if the patient can live with that. The ones that can’t live quickly gasp and start to drown when you try to remove mechanical ventilation support. This relative uniformity in disease course, due to organ damage, can be seen from the fact that mortality waves lag weeks after the admission waves. Admission waves signal the onset of organ damage with loss of function.</p><p>There’s no way to escape permanent lung damage. Lungs will never heal in a way that recovers their previous function, and we don’t have the equivalent of a dialysis machine for them. People who manage to get out of ICU despite severe lung damage have a couple of likely outcomes, namely secondary right ventricular (heart) failure and tertiary liver dysfunction. It’s also unethical in any country to keep someone alive if that requires them to be sedated and connected to a machine. That’s worse than death. So the decision is straightforward.</p><p>Lung transplants are generally not an option. Only the best cases will be offered a transplant. Almost certainly nobody over 70, and probably not someone younger if they have other health issues. When a patients’ lungs (or heart, or kidneys...) fail, their relatives always ask “Maybe they can get a transplant?”, and we always give the same response: not unless they can survive without critical support. If they can’t make it to that stage, there’s no point. If a patient is not yet stable without invasive support then trying to transplant an organ is probably just a very nasty and expensive way to kill them. The patient would almost certainly experience multi-organ failure due to the extreme stress of transplant surgery, and we would have used a donor organ that could have saved another life. This is so obviously unethical that there are no dissenting opinions that we’re aware of. But I think transplant ethics might be worthy of some review. Transplant teams have been very stringent on some of these criteria, but probably for a good reason.</p><p>We only had an ethical issue in one patient when we tried to get her into transplant. We couldn’t, because she was too old, but we thought she deserved it. She made it out of ICU out of pure strength and effort. Maybe she could have had an opportunity, but now she will just have a couple of difficult years before her heart fails. Her kidneys and liver will follow. We hope that her case will be reviewed soon.</p><p>It’s a little like doing triage after a bad traffic accident. People with very extreme injuries are left to die if there are only limited resources available at the roadside and people with lesser injuries need those. That’s just how it is: Very severe patients are lower in the priority order sometimes.</p><p>But of course, our problems over a critical patient transplant seem trivial when the negligence of governments has caused thousands to die...</p><p><i>Nick: Can you say a bit more about ARDS (Acute Respiratory Distress Syndrome)?</i></p><p>ARDS is a symptom of many diseases, including COVID, but blood and plasma transfusions can cause ARDS as well, for example. Even giving people oxygen can cause a mild form of ARDS. The ARDS that we see in COVID and other SARS-type diseases has some specific features, but the lung management of all ARDS is more or less the same.</p><p>Of course we also have to treat the underlying condition that caused ARDS. For example, when we give corticoids to COVID patients, that’s aimed at treating the underlying infection. We don’t hope to improve ARDS directly with corticoids; several studies have reported that steroids are not consistently effective in ARDS. Corticoid treatment might prevent ARDS from getting worse, but we don’t know if it really improves the ARDS that has already been “programmed” by the course of COVID up to that point.</p><p>COVID ARDS is like a jail sentence, in that it requires the patient to go through the three week process described earlier. ARDS is much rarer in flu, and when it occurs it has much wider variation in the time spent in the ICU.</p><div><br /></div>Nick Brownhttp://www.blogger.com/profile/07481236547943428014noreply@blogger.com0tag:blogger.com,1999:blog-7890764972166411105.post-85676928299517874122020-07-24T00:24:00.008+02:002020-07-25T12:28:30.371+02:00How bad is self-plagiarism? A case studyOne of the recurring topics in this blog over the last couple of years has been self-plagiarism, also known as duplicate publication or text recycling. I've shown that a number of senior scholars appear to have used this method to <a href="https://steamtraen.blogspot.com/2017/03/some-instances-of-apparent-duplicate.html">boost their number of publications</a> without having to go to the effort of <a href="https://steamtraen.blogspot.com/2018/04/some-instances-of-apparent-duplicate.html">producing new research</a>, or <a href="https://steamtraen.blogspot.com/2019/08/some-instances-of-apparent-duplicate.html">rewriting existing knowledge</a> substantially for a new audience.<br />
<br />
However, there has been some discussion online suggesting that quite a few people do not consider that self-plagiarism is a problem at all. For example, <a href="https://www.rna.uzh.ch/en/aboutus/researchgroups/aguzzi.html">Dr Adriano Aguzzi of the University of Zürich</a> sees no harm in it:<br />
<blockquote class="twitter-tweet">
<div dir="ltr" lang="en">
Today's rant is about academic "self-plagiarism". As the Editor-in-Chief of a medical journal, I thought long and hard about this issue. I have concluded that self-plagiarism is a non-existing crime. There is no moral imperative why authors should not re-use their own words.</div>
— Adriano Aguzzi (@AdrianoAguzzi) <a href="https://twitter.com/AdrianoAguzzi/status/1285264599093317632?ref_src=twsrc%5Etfw">July 20, 2020</a></blockquote>
<script async="" charset="utf-8" src="https://platform.twitter.com/widgets.js"></script>That said, Dr Aguzzi does attach a couple of conditions to his support for authors recycling their own text:<br />
<blockquote class="twitter-tweet">
<div dir="ltr" lang="en">
4. My only constraint is that (1) republication be labelled as such, and (2) that copyright legislation be respected. But forcing people to paraphrase the exactly same concepts with different words, that may be a useful exercise for secondary school - but not for scientists.</div>
— Adriano Aguzzi (@AdrianoAguzzi) <a href="https://twitter.com/AdrianoAguzzi/status/1285264603132420100?ref_src=twsrc%5Etfw">July 20, 2020</a></blockquote>
[[ Update 2020-07-25 10:28 UTC: Dr Aguzzi seems to have deleted the above tweets, along with the rest of the thread in which they appeared. I had taken a screenshot of the first one, and @deadinsideg1 kindly hunted down the second.<div><br /></div><div><br /></div><div class="separator" style="clear: both; text-align: center;"><a href="https://1.bp.blogspot.com/-EXdNP0BN9Jo/XxrgS1amxOI/AAAAAAAAFQQ/MCYyysVgaIY0xLSu42H0M9rBwwnHwAWdACLcBGAsYHQ/s805/Untitled.png" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="805" data-original-width="617" height="320" src="https://1.bp.blogspot.com/-EXdNP0BN9Jo/XxrgS1amxOI/AAAAAAAAFQQ/MCYyysVgaIY0xLSu42H0M9rBwwnHwAWdACLcBGAsYHQ/s320/Untitled.png" /></a></div><div><br /></div><div class="separator" style="clear: both; text-align: center;"><a href="https://1.bp.blogspot.com/-TTTEaLDxiwg/XxwI5UOlTOI/AAAAAAAAFQs/fc3txX-n_D4b64GIOceJqMf3VA6jtuXCACLcBGAsYHQ/s1123/Aguzzi4.png" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="287" data-original-width="1123" height="129" src="https://1.bp.blogspot.com/-TTTEaLDxiwg/XxwI5UOlTOI/AAAAAAAAFQs/fc3txX-n_D4b64GIOceJqMf3VA6jtuXCACLcBGAsYHQ/w500-h129/Aguzzi4.png" width="500" /></a></div><div><br /></div><div>]]</div><div><br /></div><div>I opened <a href="https://scholar.google.co.uk/citations?hl=en&user=3cNp9mgAAAAJ" target="_blank">Dr Aguzzi's Google Scholar page</a> and looked for the most-cited article for which he was the lead author, which was this:</div><div>
<br />
Aguzzi, A., & Polymenidou, M. (2004). Mammalian prion biology: One century of evolving concepts. <i>Cell</i>, <i>116</i>(2), 313–327. https://doi.org/10.1016/S0092-8674(03)01031-6<br />
<br />
A little searching online revealed that this was the starting point for a succession of examples of self-plagiairism, or "republication", or what Dr Aguzzi prefers to call it. In fact, as we will see, Dr Aguzzi takes a very liberal approach to recycling—or perhaps, in the modern parlance, upcycling—his previous publications.<br />
<br />
Let's start with the 2004 Aguzzi & Polymenidou article:<div><br />
<div class="separator" style="clear: both; text-align: center;">
</div>
<div class="separator" style="clear: both; text-align: center;">
</div>
<div>
</div>
<div class="separator" style="clear: both; text-align: center;">
<a href="http://1.bp.blogspot.com/-LO1dtGw7k2M/XxoEcR6DgkI/AAAAAAAAFPU/2-xN8ONDwG8z2Ol-NZwkr0oZtVzmXBheQCK4BGAYYCw/s1600/%2528One-page%2529%2BAguzzi%252C%2BPolymenidou%2B-%2B2004%2B-%2BMammalian%2BPrion%2BBiology%2B-%2BOne%2BCentury%2Bof%2BEvolving%2BConcepts.png" style="margin-left: 1em; margin-right: 1em;"><img border="0" height="640" src="https://1.bp.blogspot.com/-LO1dtGw7k2M/XxoEcR6DgkI/AAAAAAAAFPU/2-xN8ONDwG8z2Ol-NZwkr0oZtVzmXBheQCK4BGAYYCw/s640/%2528One-page%2529%2BAguzzi%252C%2BPolymenidou%2B-%2B2004%2B-%2BMammalian%2BPrion%2BBiology%2B-%2BOne%2BCentury%2Bof%2BEvolving%2BConcepts.png" width="430" /></a></div>
<br />
The text highlighted in yellow here(*) appears to have been copied, verbatim and without attribution, to this 2006 article by the same author:<br />
<br />
Aguzzi, A. (2006). Prion diseases of humans and farm animals: Epidemiology, genetics, and pathogenesis. <i>Journal of Neurochemistry</i>, <i>97</i>(6), 1726–1739. https://doi.org/10.1111/j.1471-4159.2006.03909.x</div><div><br />
<div class="separator" style="clear: both; text-align: center;">
<a href="http://1.bp.blogspot.com/-k30WLtZPaf8/XxoFXZnnkeI/AAAAAAAAFPg/ow7AGP1W8WwUMnRtnVRSkJJlVnXgrVa4gCK4BGAYYCw/s1600/%2528One-page%2529%2BAguzzi%2B-%2B2006%2B-%2BPrion%2Bdiseases%2Bof%2Bhumans%2Band%2Bfarm%2Banimals.png" style="margin-left: 1em; margin-right: 1em;"><img border="0" height="640" src="https://1.bp.blogspot.com/-k30WLtZPaf8/XxoFXZnnkeI/AAAAAAAAFPg/ow7AGP1W8WwUMnRtnVRSkJJlVnXgrVa4gCK4BGAYYCw/s640/%2528One-page%2529%2BAguzzi%2B-%2B2006%2B-%2BPrion%2Bdiseases%2Bof%2Bhumans%2Band%2Bfarm%2Banimals.png" width="432" /></a></div>
<br />
In the second of Dr Aguzzi's tweets that I quoted earlier, he made a point that republication should be labelled, and copyright respected. Strangely, however, not only does the 2006 article not mention that parts of the text were previously published; the later article does not even include the 2004 article in its References section. (For completeness, in case the publication pipeline had gone slightly awry, I checked the References section of the 2004 article, but it didn't mention the 2006 article as being "in preparation" or anything like that.) Furthermore, the 2004 article is copyrighted by Cell Press, a division of Elsevier, while the 2006 article is copyrighted by the International Society for Neurochemistry (and the journal is published by Wiley). So it seems that neither of Dr Aguzzi's constraints are met here.<br />
<br />
Some parts of the text in the above image are highlighted in green. That brings us to another article, this time with Dr Aguzzi as second author:<br />
<br />
Weissmann, C., & Aguzzi, A. (2005). Approaches to therapy of prion diseases. <i>Annual Review of Medicine</i>, <i>56</i>, 321–344. https://doi.org/10.1146/annurev.med.56.062404.172936<br />
<div class="separator" style="clear: both; text-align: center;">
</div>
<br />
<div class="separator" style="clear: both; text-align: center;">
<a href="http://2.bp.blogspot.com/-UMt0Yu3YdBI/XxoFkXB8H5I/AAAAAAAAFPo/V-BkHrF4hpom-EIesvtpATD_CW7ZZnwMACK4BGAYYCw/s1600/%2528One-page%2529%2BWeissmann%252C%2BAgazzi%2B-%2B2005%2B-%2BApproaches%2Bto%2BTherapy%2Bof%2BPrion%2BDiseases.png" style="margin-left: 1em; margin-right: 1em;"><img border="0" height="640" src="https://2.bp.blogspot.com/-UMt0Yu3YdBI/XxoFkXB8H5I/AAAAAAAAFPo/V-BkHrF4hpom-EIesvtpATD_CW7ZZnwMACK4BGAYYCw/s640/%2528One-page%2529%2BWeissmann%252C%2BAgazzi%2B-%2B2005%2B-%2BApproaches%2Bto%2BTherapy%2Bof%2BPrion%2BDiseases.png" width="430" /></a></div>
<br />
Again, the text highlighted in yellow here appears to have been copied, verbatim and without attribution, from the 2004 article mentioned above. The text highlighted in green appears to have been copied, verbatim and without attribution, from the 2006 article. The 2005 article does not include the 2006 or the 2004 article in its References section, or vice versa. Copyright for the 2005 article is owned by Annual Reviews.<br />
<br />
Another new highlight colour (pink) appears here, especially in the second half. The more astute reader may be able to work out where we are going here. Adding the yellow, green, and pink text together, we arrive at close to 95% of the content of this book chapter:<br />
<br />
Aguzzi, A. (2007). Prions. In J. H. Growdon and M. N. Rossor (Eds.), <i>The Dementias 2</i> (pp. 250<span style="background-color: white; font-family: arial, sans-serif; font-size: 14px;">–</span>275). Butterworth-Heinemann.</div><div><br />
<div class="separator" style="clear: both; text-align: center;">
<a href="http://4.bp.blogspot.com/-9mL4zNAGTG4/XxoJVUHmkUI/AAAAAAAAFP4/IQBN9-58dx0YwH5FFihtWMFjrghRBlUhwCK4BGAYYCw/s1600/%2528One-page%2529%2BAguzzi%2B-%2B2007%2B-%2BPrions.png" style="margin-left: 1em; margin-right: 1em;"><img border="0" height="640" src="https://4.bp.blogspot.com/-9mL4zNAGTG4/XxoJVUHmkUI/AAAAAAAAFP4/IQBN9-58dx0YwH5FFihtWMFjrghRBlUhwCK4BGAYYCw/s640/%2528One-page%2529%2BAguzzi%2B-%2B2007%2B-%2BPrions.png" width="428" /></a></div><div><br /></div>
If you look hard at the third page in the top row, you can perhaps make out half a page of white text. With the exception of couple of sentences elsewhere in the chapter, that half-page represents the entire original content of this article. Again, this chapter does not cite any of the three articles from which the text has been apparently copied (all together now, 1, 2, 3) verbatim and without attribution. Still, on the bright side, the book's publisher is, like Cell Press, also a division of Elsevier, so presumably there is no risk of any legal issues with the text in yellow that was copied from the 2004 article.<br />
<br />
So, back to the question in the title. Is self-plagiarism a bad thing? Dr Aguzzi clearly doesn't think so, and one can only admire the consistency of his position relative to his actions (although he might want to consider addressing the issues around labelling and copyright). I happen to think that this sort of thing is extremely bad for science, but it appears to be sufficiently common that maybe we are going to have to just live with it, and accept that some people think that churning out the same material over and over is perfectly acceptable.</div><div><br /></div><div>[[ Update 2020-07-23 23:55 UTC: Thanks to Brendan O'Connor for <a href="https://www.psychologicalscience.org/publications/aps-editorial-policies" target="_blank">this link</a> to the Association for Psychological Science's policy on self-plagiarism. Spoiler: They are not too keen on it. ]]<br />
<br />
<br />
(*) I have made the full-sized images and the annotated PDF files available for download <a href="http://nickbrown.fr/blog/aguzzi" target="_blank">here</a>. I hope that this counts as fair use for the purposes of this blog post. If the owners of the copyright want to object to this, I hope that they will realise the irony that would be involved if they decided to enforce their rights now.<br />
<br />
<br />
<br /></div></div>Nick Brownhttp://www.blogger.com/profile/07481236547943428014noreply@blogger.com7tag:blogger.com,1999:blog-7890764972166411105.post-66273378501621007262020-07-16T17:33:00.003+02:002020-07-16T17:37:53.922+02:00An expression of concern about Expressions of Concern<div>In academic publishing, what is the purpose of a journal issuing an "Expression of Concern" (EoC)?</div><div><br /></div><div>When I first came across the concept, I was told that an EoC was a sort of preliminary step on the way to retraction. The journal acknowledges that it has received information that suggests that an article may not be reliable. This information seems, on the face of it, to be quite convincing. The journal is still investigating exactly what happened, but in the meantime, here is an early warning that people who are thinking of citing this article might want to think twice. We could see it as the equivalent of locking up someone who is accused of a serious crime: They have not yet been found guilty, their detention is only preventive (and often under better conditions than those who have been convicted), but the <i>prima facie</i> case is such that on balance, we probably don't want to have that person walking around unchecked.</div><div><br /></div><div>An example of this came in the Brian Wansink case. After <a href="https://retractionwatch.com/2017/10/20/retract-replace-retract-beleaguered-food-researcher-pulls-article-jama-journal/" target="_blank">retracting, republishing, and re-retracting</a> one of Wansink's articles, <i>JAMA</i> <a href="https://retractionwatch.com/2018/04/13/caught-our-notice-jama-warns-readers-about-all-of-brian-wansinks-papers-in-its-journals/" target="_blank">placed EoCs on six other articles</a> with Wansink as an author that had been published in its family of journals. A few months later, with no satisfactory response having being received to explain the problems in those articles, <a href="https://cornellsun.com/2018/09/20/wansink-sinks-lower-with-six-study-retractions-in-one-day/" target="_blank">all six were retracted</a>.</div><div><br /></div><div>However, it appears that many journals or editors are using the term "Expression of Concern" to mean something else. <a href="https://journals.sagepub.com/doi/abs/10.1177/0002764214537204">This article</a> has had an EoC on it for six years now. The editors of <i>Psychology of Music</i> just issued <a href="https://journals-sagepub-com.proxy.lnu.se/doi/full/10.1177/0305735620943366">this EoC</a>, but <a href="https://twitter.com/samuelmehr/status/1283661082444226563?s=20">according to Samuel Mehr</a> they have no plans to escalate to a retraction. The author of that last paper has also had <a href="https://www.rips-irsp.com/articles/10.5334/irsp.304/" target="_blank">five EoCs in place at another journal</a> for over a year.</div><div><br /></div><div>This type of EoC basically comes down to the following statement from the editors: "We have good reason to believe that this article is garbage, and you should not trust it. But we're not going to do anything about it that might hurt our impact factor, or embarrass us by getting us into Retraction Watch." It's like a restaurant menu with a small sticker saying "Pssst: The fish is terrible, please don't order it". (Plus, the sticker is permanent. It's inside the laminated cover of the menu.)</div><div><br /></div><div>It has <a href="http://www.nature.com/news/set-up-a-self-retraction-system-for-honest-errors-1.19619" target="_blank">been suggested</a>, <a href="https://www.biorxiv.org/content/early/2017/03/24/118356" target="_blank">more than once</a> (albeit with some <a href="https://absolutelymaybe.plos.org/2017/11/30/rebranding-retractions-and-the-honest-error-hypothesis/" target="_blank">pushback</a>) that we need different words for different types of retraction (say, "obvious fraud" versus "honest error"). It seems that we also need two different words to describe these two different usages of "Expression of Concern". One journal editor <a href="https://steamtraen.blogspot.com/2017/04/an-open-letter-to-dr-todd-shackelford.html" target="_blank">posted what he called an "Editorial Note" on a Wansink article</a>; while this was frustrating for those of us who wanted that article to be retracted, at least it was very clear from that "Editorial Note" that the editor was not remotely interested in doing anything else about the problem. Perhaps that's the way "forward", although it doesn't feel like progress. Correcting the scientific record continues to <a href="https://steamtraen.blogspot.com/2020/06/the-gueguen-saga-update-summer-2020.html">feel like pulling teeth</a>.</div><div><br /></div><div><br /></div><br />Nick Brownhttp://www.blogger.com/profile/07481236547943428014noreply@blogger.com3tag:blogger.com,1999:blog-7890764972166411105.post-41382093932127908462020-06-29T13:03:00.004+02:002022-11-23T01:10:09.986+01:00The Guéguen saga update, summer 2020 editionRegular readers of this blog may recall seeing a number of posts about the remarkable research of <a href="https://www.univ-ubs.fr/fr/annuaire/g/u/e/user-gueguenn-fr.html">Dr Nicolas Guéguen</a>. In 2017 I wrote <a href="https://steamtraen.blogspot.com/2017/11/some-problems-in-field-study-of-sexual.html">here</a> and <a href="https://steamtraen.blogspot.com/2017/12/more-problematic-sexual-attraction.html">here</a> (and <a href="https://twitter.com/jamesheathers">James Heathers</a> wrote <a href="https://medium.com/@jamesheathers/life-in-the-tinderbox-6b2e9760f3aa">here</a> and <a href="https://medium.com/@jamesheathers/long-hair-dont-care-5eeba266ec52">here</a>) about several articles with Dr Guéguen as sole author that seemed to have a number of problems, which we summarized <a href="https://steamtraen.blogspot.com/2017/12/a-review-of-research-of-dr-nicolas.html">here</a>. In May 2019 James and I posted an <a href="https://steamtraen.blogspot.com/2019/05/an-update-on-our-examination-of.html">update</a> in which we reported that the university had investigated and required Dr Guéguen to retract two articles, but that he had not yet done so by the deadline that he accepted. I wrote to the editors of the journals concerned, one of whom did not even acknowledge receipt of my e-mails until I wrote <a href="https://steamtraen.blogspot.com/2019/07/an-open-letter-to-dr-jerker-ronnberg.html">an open letter to him</a>.<br />
<br />
A year has gone by, and there have been a few developments. Modest developments, to be sure, but in the <a href="https://www.sciencemag.org/news/2018/02/meet-data-thugs-out-expose-shoddy-and-questionable-research">error detection business</a> you take whatever you can get...<br />
<br />
<h4>
Radio silence</h4>
<br />
Dr Guéguen appears to have almost entirely stopped publishing research articles. He has deleted his Google Scholar profile, but by searching with his name I was able to identify only a few articles that have appeared since 2017, and some of those appear to have been submitted some time before that (e.g., <a href="https://journals.sagepub.com/doi/abs/10.1177/1096348013515923">this one</a>, which was published online in 2013 but only assigned a definitive journal page number in 2017, in what seems --- unless there is a plausible alternative explanation --- to be rather dishonest behaviour by the journal, which appears to be using a known trick of garnering citations before attributing a final publication date in order to boost its impact factor). There is <a href="https://web.b.ebscohost.com/abstract?direct=true&profile=ehost&scope=site&authtype=crawler&jrnl=15277143&AN=139889396&h=5G6xMrxo9n9s%2f%2bDybpYYrYH0mDqgQUqQxks4v6CUir1Me3gRtt9%2bpd6XcQ8LBrjZjVtzMPpysAIu0CaYXgF6oA%3d%3d&crl=c&resultNs=AdminWebAuth&resultLocal=ErrCrlNotAuth&crlhashurl=login.aspx%3fdirect%3dtrue%26profile%3dehost%26scope%3dsite%26authtype%3dcrawler%26jrnl%3d15277143%26AN%3d139889396">one article from 2019</a> with Dr Guéguen listed as last author, which seems to follow a similar design to much of the rest of his research output, but <a href="http://najp.us/">the journal in that case</a> does not seem to report received/accepted dates on its articles, so this one could have been in the pipeline for some time. Apart from that, though, it seems that Dr Guéguen's previously prolific research output, with up to 20 single-authored publications in some years as well as numerous collaborations, seems to have suddenly ceased round about the time that we started raising questions about it. Presumably this is just a coincidence.<br />
<br />
<h4>
Antoine strikes again</h4>
<br />
<a href="https://twitter.com/samuelmehr">Samuel Mehr</a> has been in correspondence with the editors of <i>Psychology of Music</i> about this article:<br />
<br />
Guéguen, N., Meineri, S., & Fischer-Lokou, J. (2014). Men’s music ability and attractiveness to women in a real-life courtship context. <i>Psychology of Music</i>, <i>42</i>, 545–549. https://doi.org/10.1177/0305735613482025 (PDF available <a href="http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.987.1687&rep=rep1&type=pdf">here</a>)<br />
<br />
Readers who have read one or two of Dr Guéguen's sole-authored articles may wonder exactly what the other two authors contributed here, as this study is just like the others: A guy called Antoine is trying to pick up young women. In this study he was either carrying a guitar case, a sports bag, or nothing. He got the woman's phone number more often when carrying the guitar. The usual problems are apparent, notably the perfect response rate and the number of women who would have to have all decided to walk down this particular street on their own on a single Saturday afternoon.<br />
<br />
When I started drafting this post a couple of days ago, Samuel told me that his last correspondence with the journal had been <a href="https://twitter.com/samuelmehr/status/1277369610333851648?s=20">in January of this year</a>, when (from what I have seen) they appeared to suggest that some sort of action might be forthcoming quite quickly, after an investigation by the publisher's ethical committee. I initially wrote here "But since then, nothing has been forthcoming". However, today I have been in copy of an e-mail exchange in which the editors of the journal revealed that they are preparing an expression of concern for the article. (I plan to write a separate blog post about the whole question of expressions of concern.)<br />
<br />
<h4>
The awesome power of procrastination (1)</h4>
<br />
Last time, we reported that, as part of the investigation into his research conducted by the scientific integrity officials at his university, Dr Guéguen had agreed to retract two articles; however, the deadline to do this had passed, and neither article had been retracted. A few months after that blog post, in October 2019, <a href="https://link.springer.com/article/10.1007/s10508-019-01558-0#CR1">one of these articles</a> was <a href="https://retractionwatch.com/2019/10/12/in-2014-a-study-claimed-high-heels-made-women-more-attractive-now-its-been-retracted/">retracted by the journal</a> "at the request of the Université de Bretagne-Sud" (UBS), Dr Guéguen having apparently not honoured his commitment to do this himself. However, as of this writing, <a href="https://onlinelibrary.wiley.com/doi/abs/10.1002/col.20651">the second of these articles</a> still has not been retracted. I have been in contact with the UBS to ask why they apparently did not ask the journal for this second article to be retracted, and the scientific integrity officer there has told me that he will pass on my message to the Presidents of the two universities involved in the scientific investigation (UBS and Rennes-2). Perhaps something will come of this.<br />
<br />
<h4>
I fought the law, and the law went "meh"</h4>
<br />
One apparently positive development of the scientific investigation (see previous paragraph) is that the Presidents of UBS and Rennes-2 decided to launch a disciplinary investigation into Dr Guéguen. This took the better part of a year to convene, apparently because they had difficulty finding people to serve on the panel. In the end it was outsourced to the University of Angers.<br />
<br />
James Heathers and I gave evidence to this inquiry in October 2019. In my evidence I emphasised that Dr Guéguen's principal defence --- namely, that he had naïvely trusted his students to do good fieldwork, despite them having zero training --- did not hold up, because in many cases his articles would have required input from faculty members and the expenditure of budget money (whereas no funding ever seems to be reported). For example, studies where saliva samples are taken require analyses to be performed in a laboratory; even if this is on the university campus these assays will certainly need to be paid for, and even if there is some remarkable system whereby this is done for free in the name of research, there will exist traces of the request for the analyses to be performed.<br />
<br />
Since then we have heard nothing official. But I have been told by two people with inside knowledge that a report exists, and that it states that Dr Guéguen did nothing to violate scientific integrity.<br />
<br />
<h4>
The awesome power of procrastination (2)</h4>
<br />
Also last time, we noted that we had strong evidence that Dr Guéguen's article entitled "Women’s hairstyle and men’s behavior: A field experiment", published in the <i>Scandinavian Journal of Psychology</i>, was, as a minimum, stolen from the work of three undergraduates, with the added twist that these undergraduates might well have themselves fabricated the study.<br />
<br />
Following my open letter to Dr Jerker Rönnberg, the editor-in-chief of the journal, he agreed that he would look into the matter. I have written to him a couple of times since then, but there was no reply or acknowledgement of any kind until, in response to an e-mail that I sent on 29 May 2020, I received an out-of-office (from the e-mail address that is listed for contact on the journal's web site) stating that Dr Rönnberg now has the status of professor emeritus and "will only answering questions/e-mails occasionally". This didn't seem like a very satisfactory state of affairs, so I wrote to the editorial assistant of the journal. Dr Stefan Gustafson. He told me that the editors had discussed the matter with the publisher and then contacted the original reviewers, of whom one didn't respond and the other said they thought the paper was fine. No decision has yet been taken about "Women’s hairstyle and men’s behavior", and I got the impression from Dr Gustafson's e-mail that this is unlikely to happen before a new editor-in-chief is in place. <judge_judy_taps_wristwatch.gif><br />
<br />
<h4>
The past is a foreign country; they do science differently there</h4>
<br />
In September 2019, the editors of the <i>International Review of Social Psychology</i> (IRSP) received a report that they had commissioned from <a href="https://twitter.com/hansijzerman">Hans IJzerman</a> into the six articles by Dr Guéguen that were published in their journal. This report recommended that two of the articles be retracted immediately, two others be given an expression of concern, and two should be corrected. One of the people asked by Hans to verify the accuracy of the report wrote that "the evidence from the blog posts and statistical investigators supports the conclusions ... that research misconduct likely took place".<br />
<br />
Instead of issuing any retractions, however, the editors of IRSP <a href="https://www.rips-irsp.com/articles/10.5334/irsp.304/print/">issued five expressions of concern and accepted one correction</a>. As part of their reasoning for why no article should be retracted, they stated that, although "[t]he report concludes misconduct", "the standards for conducting and evaluating research have evolved since [these articles were published]". I will leave it up to the reader to judge whether what took place (or, perhaps more relevantly, did not take place) in these cases was reasonable by the standards of social psychology in the period 2002–2011. Perhaps <a href="https://www.apa.org/science/about/psa/2011/12/diederik-stapel">Diederik Stapel</a> will be getting his PhD back soon; after all, this was his most prolific period too.<br />
<br />
<h4>
That open goal looks so nice, it would be a shame to kick a ball at it</h4>
<br />
To my knowledge, the only other formal action by a journal in the past year, apart from the response by ISRP (see previous point), is the expression of concern that was issued on 16 March 2020 by the editors of <a href="https://lebs.hbesj.org/index.php/lebs/article/view/335">Letters on Evolutionary Behavioral Science</a> regarding this article, which we examined in <a href="https://www.dropbox.com/s/t61v1ypndryts63/NB-JH%20commentary%20on%20articles%20-%2016%20January%202018.pdf?dl=0">our original report</a> (in which we showed that the claimed pattern of behaviour by participants was highly unlikely in all conditions of the study):<br />
<br />
Guéguen, N. (2012). Risk taking and women’s menstrual cycle: Near ovulation, women avoid a doubtful man. <i>Letters on Evolutionary Behavioral Science</i>, <i>3</i>, 1–3. https://doi.org/10.5178/lebs.2012.17<br />
<br />
This expression of concern ends with the following paragraph, which will sound rather familiar to anyone who has been following this, or indeed almost any other recent story of <a href="https://econjwatch.org/articles/the-stewart-retractions-a-quantitative-and-qualitative-analysis">journals' responses to terrible articles</a>:<br />
<blockquote class="tr_bq">
Although the investigation committee concluded that there is no decisive evidence of scientific misconduct, they still share Brown and Heathers’s (2017) concerns. Moreover, the above errors in statistics severely discredit the scientific value of Guéguen (2012). In sum, we admit that we do not have decisive evidence to retract the publication of Guéguen (2012). However, we would like to advise readers of LEBS to exercise great caution in interpreting the reported results in Guéguen (2012).</blockquote>
The article is still sitting there as part of the scientific record in the journal, <a href="https://lebs.hbesj.org/index.php/lebs/article/view/lebs.2012.17">its web page</a> does not mention the expression of concern, and the PDF file has not been modified.<br />
<br />
<h4>
Conclusion</h4>
<br />
At the risk of letting my attempt at a mask of professionalism slip for a moment: FFS. This whole process is like pulling teeth. There has to be a better way to handle cases of obviously shoddy science than this. Four and a half years after James and I started looking at a huge number of studies that cannot possibly have taken place as described, we have a total of one retraction and seven (soon top be eight) expressions of concern (and it is unclear whether any of those represent a prelude to retraction). If the French academic establishment and the international publishing system can't be bothered to clean up a case as obviously terrible as this in a field with almost no conflicts of interest, what chance is there of anything being done about, say, <a href="https://retractionwatch.com/2020/04/12/elsevier-investigating-hydroxychloroquine-covid-19-paper/">ethical violations in COVID-19 research</a>?<br />
<br />Nick Brownhttp://www.blogger.com/profile/07481236547943428014noreply@blogger.com0tag:blogger.com,1999:blog-7890764972166411105.post-26948220796274571752020-05-24T19:14:00.002+02:002020-07-31T17:19:36.967+02:00The Silence of the RIOs<div class="tr_bq">
Just over a month ago, I published <a href="https://steamtraen.blogspot.com/2020/04/some-issues-in-recent-gaming-research.html">these</a> <a href="https://steamtraen.blogspot.com/2020/04/the-mystery-of-missing-authors.html">two</a> blog posts. After the first, Daniël Lakens tweeted this:</div>
<blockquote class="twitter-tweet">
<div dir="ltr" lang="en">
I hope you forwarded a link to your blog to the university ethics person at these 12 respective institutions? At least a benefit of having so many authors is that 1 uni might take some sort of action?</div>
— Daniël Lakens (@lakens) <a href="https://twitter.com/lakens/status/1252568370337189888?ref_src=twsrc%5Etfw">April 21, 2020</a></blockquote>
I thought that was a good idea, so I set out to find who the "university ethics person" might be for the 15 co-authors of the article in question. (I wrote directly and separately to the two PhD supervisors of the lead author, as it is he who appears to be <i>prima facie</i> responsible for most of its deficiencies; I also wrote to <i>Nature Scientific Reports</i> outlining my concerns about the article. In both cases I received a serious reply indicating that they were concerned about the situation.)<br />
<br />
It turns out that finding the address of the person to whom complaints about research integrity at a university or other institution is not always easy. There were only one or two cases where I was able to do this by following links from the institution's web site, as <a href="https://xkcd.com/773/">regular readers of xkcd might have been able to guess</a>. In a few cases I used Google with the site: option to find a person. But about half the time, I couldn't identify anyone. In those cases I looked for the e-mail address of someone who might be the dean or head of department of the author concerned. Hilariously, in one case, the author <i>was</i> the head of department and I ended up writing to the president of the university.<br />
<br />
Anyway, by 24 April 2020 I had what looked like a plausible address at all of the different institutions to which the co-authors were affiliated (which turned out to be nine in total, not 12), so I sent this e-mail.<br />
<blockquote>
From: Nicholas Brown <nicholas.brown@lnu.se><br />
Sent: 24 April 2020 16:04<br />
To: [9 people]<br />
Subject: Possible scientific misconduct in an article published in Nature Scientific Reports</blockquote>
<blockquote>
First, allow me to apologise if I have addressed this e-mail to any of you in error, and also if my use of the phrase "Research Integrity Officer" in the above salutation is not an accurate summary of your job title. I had some difficulty in establishing, from your institution's web site, who was the correct person to write to for questions of research integrity in many cases, including [list]. In those cases I attempted to identify somebody who appears to have a senior function in the relevant department. In the case of [institution], I only found a general contact address --- I am trying to reach someone who might have responsibility for the ethical conduct of "XXX" in the XXX Department.</blockquote>
<blockquote>
I am writing to bring your attention to these [<i>sic</i>; I started drafting the e-mail before I wrote the second post, and not everything about it evolved correctly after that] blog posts, which I published on April 21, 2020: https://steamtraen.blogspot.com/2020/04/some-issues-in-recent-gaming-research.html. </blockquote>
<blockquote>
At least one author of the scientific article that is the principal subject of that blog post (Etindele Sosso et al., 2020; https://doi.org/10.1038/s41598-020-58462-0, published on 2020-02-06 in Nature Scientific Reports) lists your institution as their affiliation. </blockquote>
<blockquote>
While my phrasing in that public blog post (and a follow-up, which is now linked from the first post) was necessarily conservative, I think it is clear to anyone with even a minimum of relevant scientific training who reads it that there is strong <i>prima facie</i> evidence that the results of the Etindele Sosso et al. (2020) article have been falsified, and perhaps even fabricated entirely. Yet, 15 other scholars, including at least one at your institution (in the absence of errors of interpretation on my part) signed up to be co-authors of this article.</blockquote>
<blockquote>
There would seem to be two possibilities in the case of each author.</blockquote>
<blockquote>
1. They knew, or should have known, that the reported results were essentially impossible. (Even the Abstract contains claims about the percentage of variance explained by the main independent variable that are utterly implausible on their face.)</blockquote>
<blockquote>
2. They did not read the manuscript at all before it was submitted to a Nature group journal, despite the fact that their name is listed as a co-author and included in the "Author contributions" section as having, at least, "contributed to the writing".<br />
It seems to me that either of these constitutes a form of academic misconduct. If these researchers knew that the results were impossible, they are culpable in the publication of falsified results. If they are not --- that is, their defence is that they did not read and understand the implications of the results, even in the Abstract --- then they have made inappropriate claims of authorship (in a journal whose own web site states that it is the 11th most highly cited in the world). Either of these would surely be likely to bring your institution into disrepute.</blockquote>
<blockquote>
For your information, I intend to make this e-mail public 30 days from today, accompanied by a one-sentence summary (without, as far as possible, revealing any details that might be damaging to the interests of anyone involved) of your respective institutions' responses until that point. I would hope that, despite the difficult circumstances under which we are all working at the moment, it ought to be able to at least give a commitment to thoroughly investigate a matter of this importance within a month. I mention this because in previous cases where I have made reports of this kind, the modal response from institutional research integrity officers has been no response at all.</blockquote>
<blockquote>
Of course, whatever subsequent action you might decide to take in this matter is entirely up to you.</blockquote>
<blockquote>
Kind regards,<br />
Nicholas J L Brown, PhD<br />
Linnaeus University</blockquote>
The last-but-one paragraph of that e-mail mentions that, 30 days from the date of the e-mail, I intended to make it public, along with a brief summary of the responses from each institution. The e-mail is above. Here is how each institution responded:<br />
<br />
Nottingham Trent University, Nottingham, UK: Stated that they would investigate, and gave me an approximate date by which they anticipated that their investigation would be complete.<br />
Central Queensland University, Rockhampton, Australia: Stated that they would investigate, but with no estimate of how long this would take.<br />
Autonomous University of Nuevo Leon, Monterrey, N.L., Mexico: No reply.<br />
Jamia Millia Islamia, New Delhi, India: No reply.<br />
University of L’Aquila, L’Aquila, Italy: No reply.<br />
Army Share Fund Hospital, Athens, Greece: No reply.<br />
Université de Montréal, Montréal, Québec, Canada: No reply.<br />
University of Limerick, Limerick, Ireland: No reply.<br />
Lero Irish Software Research Centre, Limerick, Ireland: No reply.<br />
<br />
By "No reply" here, I mean that I received nothing. No "Undeliverable" message. No out-of-office message. No quick reply saying "Sorry, COVID-19 happened, we're busy". Not "We'll look into it". Not "We won't look into it". Not even "Get lost, there is clearly no case to answer here". Nothing, nada, nichts, rien, zip, in reply to what I (and, apparently, the research integrity people at the two institutions that did reply) think is a polite, professional e-mail, with a subject line that I hope suggests that a couple of minutes of the recipient's time might be a worthwhile investment, in 7 out of 9 cases.<br />
<br />
I find this disappointing. I wish I could say that I found it remotely surprising. Maybe I should just be grateful that Daniël's estimate of one institution taking any sort of action was exceeded by 100%.<br />
<br />
<br />
<iframe allow="accelerometer; autoplay; encrypted-media; gyroscope; picture-in-picture" allowfullscreen="" frameborder="0" height="315" src="https://www.youtube.com/embed/TI0fk_gCdzE" width="560"></iframe>
<script async="" charset="utf-8" src="https://platform.twitter.com/widgets.js"></script>
Nick Brownhttp://www.blogger.com/profile/07481236547943428014noreply@blogger.com4tag:blogger.com,1999:blog-7890764972166411105.post-50638022568979246422020-05-16T17:41:00.002+02:002020-05-18T02:20:14.942+02:00The perils of improvising with linear regression: Stedman et al. (in press)This article has been getting a lot of coverage of various kinds in the last few days, including the <a href="https://www.manchestereveningnews.co.uk/news/greater-manchester-news/quarter-population-infected-coronavirus-testing-18256136">regional</a>, <a href="https://www.mirror.co.uk/science/coronavirus-over-25-uk-likely-22025278">UK</a> <a href="https://www.independent.co.uk/news/health/coronavirus-19-million-uk-infected-study-antibody-test-cases-a9515886.html">national</a>, and <a href="https://news-24.fr/coronavirus-uk-220k-en-angleterre-pourraient-etre-infectes-actuellement/">international</a> news media:<br />
<br />
Stedman, M., Davies, M., Lunt, M., Verma, A., Anderson, S. G., & Heald, A. H. (in press). A phased approach to unlocking during the COVID-19 pandemic – Lessons from trend analysis. <i>International Journal of Clinical Practice</i>. https://doi.org/10.1111/ijcp.13528<br />
<br />
There doesn't seem to be a typeset version from the journal yet, but you can read the final draft version online <a href="https://onlinelibrary.wiley.com/doi/epdf/10.1111/ijcp.13528">here</a> and also download it as a PDF file.<br />
<br />
The basic premise of the article is that, according to the authors' model, COVID-19 infections are far more widespread in the community in the United Kingdom(*) than anyone seems to think. Their reasoning works in three stages. First, they built a linear model of the spread of the disease, one of whose predictors was the currently reported number of total cases (i.e., the official number of people who have tested positive for COVID-19). Second, they extrapolated that model to a situation in which the entire population was infected, assuming the the spread continues to be entirely linear. Third, they used the slope of their line to estimate what the official reported number of cases would be at that point. They concluded their model shows that the true number of cases in the population is 150 times larger than the number of positive tests that have been carried out, so that on the day when their data were collected (24 April 2020) 26.8% of the population were already infected.<br />
<br />
The above figures are from the Results section of the paper, on p. 11 of the final draft PDF. However, the Abstract contains different numbers, which seem to be based on data from 19 April 2020. The Abstract asserts that the true number of cases in the population may be 237 (versus 150) times the reported number, and that the percentage of the population who had been infected might have been 29% (versus 26.8% several days later). Aside from the question of why the Abstract includes different principal results from the Results section, it would appear to be something of a problem for the authors' assumptions of (ongoing) linearity in the relation between the spread of the disease and the number of reported cases if the slope of their model changed by a factor of one-third over five days.<br />
<br />
But it seems to me that, apart from the rather tenuous assumptions of linearity and the validity of extrapolating a considerable way beyond the range of the data (which, to be fair, they mention in their Limitations paragraph), there is an even more fundamental problem with how the authors have used linear regression here. Their regression model contained at least nine covariates(**), and we are told that "The stepwise(***) regression of the local UTLA factors to R<sub>ADIR</sub> showed that only one factor total reported cases/1,000 population [<i>sic</i>] was significantly linked" (p. 11). I take this to mean that, if the authors had reported the regression output in a table, this predictor would be the only one whose absolute value was at least twice its standard error. (The article is remarkably short on numerical detail, with neither a table of regression output nor a table of descriptives and correlations of the variables. Indeed, there are many points in the article where a couple of minutes spent on attention to detail would have greatly improved its quality, even in the context of the understandable desire to communicate results rapidly during a pandemic.)<br />
<br />
Having established that only one predictor in this ten-predictor regression was statistically significant (in the sense of a 95-year-old throwaway remark by R. A. Fisher), the authors then proceeded to do something remarkable. Remember, they had built this model:<br />
<br />
Y = B<sub>0</sub> + B<sub>1</sub>X<sub>1</sub> + B<sub>2</sub>X<sub>2</sub> + B<sub>3</sub>X<sub>3</sub> + ... + B<sub>10</sub>X<sub>10</sub> + Error<br />
<br />
(with Error apparently representing 78% of the variance, cf. line 3 of p. 11). But they then dropped ten of those terms (nine regression coefficients multiplied by the values of the predictors, plus the error term) to come up with this model (p. 11):<br />
<br />
R<span style="font-size: 13.3333px;">ADIR</span> = 1.06 - 0.16 x Current Total Cases/1,000<br />
<br />
What seems to have happened here is that authors in effect decided to set the regression coefficients B<sub>2</sub> through B<sub>10</sub> to zero, apparently because their respective values were less than twice their standard errors and so "didn't count" somehow. However, they retained the intercept (B<sub>0</sub>) and the coefficient associated with their main variable of interest (B<sub>1</sub>) from the 10-variable regression, as if the presence of the nine covariates had had no effect on the calculation of these values. But of course, those covariates <i>had</i> had an effect on both the estimation of the intercept and the coefficient of the first variable. That was precisely what they were included in the regression for. If the authors had wanted to make a model with just one predictor (the number of current total cases), they could have done so quite simply a one-variable regression. You can't just run a multiple regression and keep the coefficients (with or without the intercept) that you think are important while throwing away the others.<br />
<br />
This seems to me to be a rather severe misinterpretation of what a regression model is and how it works. There are many other things that could be questioned about this study(****), and indeed <a href="https://pubpeer.com/publications/1AD2A3A8FA505CEA1ECF0E4F1949E9">several people are already doing precisely that</a>, but this seems to me to a very fundamental problem with the article, and something that the reviewers really ought to have picked up. The first two authors appear to be <a href="https://beta.companieshouse.gov.uk/company/04632525/officers">management consultants</a> whose qualifications to conduct this sort of analysis are unclear, but the <a href="https://www.research.manchester.ac.uk/portal/mark.lunt.html">third author's faculty page</a> suggests that he knows his stuff when it comes to statistics, so I'm not sure how this was allowed to happen.<br />
<br />
Stedman et al. end with this sentence: "The manuscript is an honest, accurate, and transparent account of the study being reported. No important aspects of the study have been omitted." This is admirable, and I take it to mean that the authors did not in fact run a one-predictor regression to estimate the effect of their main IV of interest on their DV before they decided to run a stepwise regression with nine covariates. However, I suggest that it might be useful if they were to run that one-predictor regression now, and report the results along with those of the multiple regression (cf. Simmons, Nelson, & Simonsohn, 2011, p. 1362, Table 2, point 6). When they do that, they might also consider incorporating the latest testing data and see if the slope of their regression has changed, because since 24 April the number of cases in the UK has more than doubled (240,161 at the moment I am writing this), suggesting that between 54% and 84% of the population has by now become infected, depending on whether we take the numbers from p. 11 of the article or those from the Abstract.<br />
<br />
[[ Update 2020-05-18 00:15 UTC: There is a preprint of this paper, available <a href="https://www.medrxiv.org/content/10.1101/2020.04.20.20072264v1">here</a>. It contains the same basic model, which is estimated using data from 11 days earlier than the accepted manuscript. In the preprint, the regression equation (p. 7) is:<br />
<br />
R<sub>ADIR</sub> = 1.20 - 0.26 x Current Total Cases/1,000<br />
<br />
In other words, between the submission date of the preprint and the preparation data of the final manuscript, the slope of the regression line --- which the model assumes would be constant until everyone was infected --- changed from -0.26 to -0.16. And yet the authors did not apparently think that this was sufficient reason to question the idea that the progress of the disease would continue to match their linear model, <i>despite direct evidence that it had failed to do so over the previous 11 days</i>. This is quite astonishing. ]]<br />
<br />
(*) Whether the authors claim that their model applies to the UK or just to England is not entirely clear, as both terms appears to be used more or less interchangeably. They use a figure of 60 million as the population of England, although the Office of National Statistics <a href="https://www.ons.gov.uk/peoplepopulationandcommunity/populationandmigration/populationestimates/bulletins/annualmidyearpopulationestimates/mid2019">reports</a> figures of 56.3 million for England and 66.8 million for the UK in mid 2019.<br />
<br />
(**) I wrote <a href="https://twitter.com/sTeamTraen/status/1261334472533979141?s=20">here</a> that there were nine, but one of them is ethnicity, which would typically have been coded as a series of categories, each of which would have functioned as a separate predictor in the regression. But maybe they used some kind of shorthand such as "Percentage who did/didn't identify in the 'White British' category", so I'll continue to assume that there were nine covariates and hence 10 predictors in total.<br />
<br />
(***) Stepwise regression is generally not considered a good idea these days. See for example <a href="https://towardsdatascience.com/stopping-stepwise-why-stepwise-selection-is-bad-and-what-you-should-use-instead-90818b3f52df">here</a>. Thanks to Stuart Ritchie for this tip and for reading through a draft of this post.<br />
<br />
(****) Looking at Figure 4, it occurs to me that a few data points at the top left might well be lifting the left-hand side of the regression line up to some extent, but that's all moot until we know more about the single-variable regression. Also, there is no confidence interval or any other measure of the uncertainty --- even assuming that the model is perfectly linear --- of the estimated reported case rate when the infection rate drops to zero.<br />
<br />
<br />Nick Brownhttp://www.blogger.com/profile/07481236547943428014noreply@blogger.com5tag:blogger.com,1999:blog-7890764972166411105.post-60215177684105751742020-04-23T19:02:00.000+02:002020-04-27T19:24:40.765+02:00The Mystery of the Missing Authors<div dir="ltr" style="text-align: left;" trbidi="on">
What do the following researchers have in common?<br />
<br />
<a href="https://www.researchgate.net/profile/Hito_G_Muller" target="_blank">Muller G. Hito</a>, Department of Psychology and Sports Science, Justus-Liebig University, Germany <br />
<a href="https://www.researchgate.net/profile/Okito_Nakamura" target="_blank">Okito Nakamura</a>, Global Research Department, Ritsumeikan University, Japan<br />
<a href="https://www.researchgate.net/profile/Mitsu_Nakamura" target="_blank">Mitsu Nakamura</a>, The Graduated [<i>sic</i>] University of Advanced Studies, Japan<br />
<a href="https://www.researchgate.net/scientific-contributions/2156415653_John_Okutemo">John Okutemo</a>, Usman University, Sokoto, Nigeria<br />
<a href="https://www.researchgate.net/scientific-contributions/2144215074_Eryn_Rekgai" target="_blank">Eryn Rekgai</a>, Department of Psychology and Sports Science, Justus-Liebig University, Germany<br />
<div>
<a href="https://www.researchgate.net/scientific-contributions/2159315000_Mbraga_Theophile">Mbraga Theophile</a>, Kinshasa University, Republic of Congo</div>
<div>
<a href="https://www.researchgate.net/profile/Bern_Schmidt">Bern S. Schmidt</a>, Department of Fundamental Neuroscience, University of Lausanne, Switzerland</div>
<div>
<br /></div>
Despite their varied national origins, it seems that the answer is "quite a bit":<br />
<br />
1. They seem to collaborate with each other, in various combinations, on short articles with limited empirical content, typically published with less than a week from submission to acceptance. (Some examples: <a href="https://www.researchgate.net/publication/326248708_Hypothesis_of_Memorisation_Process">1</a> <a href="https://www.researchgate.net/publication/326824989_A_novel_reflection_on_sleep_disorders_approach">2</a> <a href="https://www.researchgate.net/publication/332653067_Executive_Functions_Are_Better_than_IQ">3</a> <a href="https://www.researchgate.net/publication/326064125_MHPE_An_Introduction_to_a_Complete_Evaluation_Tool">4</a> <a href="https://www.researchgate.net/publication/334122108_Cognitive_Performance_Opinion">5</a>) The majority of these articles date from 2017, although there are some from 2018 and 2019 as well.<br />
<br />
2. Apart from each other, these people have published with almost nobody else, except that:<br />
<br />
(a) Four of them have published with Faustin
Armel Etindele Sosso (whom I will refer to from now on as FAES), the lead author of the article
that I discussed in <a href="https://steamtraen.blogspot.com/2020/04/some-issues-in-recent-gaming-research.html" target="_blank">this post</a>. (Examples: <a href="https://www.researchgate.net/publication/326248322_African_burden_of_Mental_Health_necessity_of_global_exchange_between_researchers">6</a> <a href="https://www.researchgate.net/publication/319699832_Basic_activity_of_Neurons_in_the_dark_during_Somnolence_induced_by_Anesthesia">7</a> <a href="https://www.blogger.com/Epidemiology%20of%20Alzheimer's%20Disease:%20Comparison%20between%20Africa%20and%20South%20America">8</a>) In <a href="https://www.researchgate.net/publication/326607791_Applications_of_the_Mental_Health_Profile_of_Etindele_Questionnaire">one case</a>, FAES is the corresponding author although he is not listed as an actual author of the article. I don't think I have ever seen that before in scholarly publishing.<br />
<br />
(b) Two of them have published with an author named Sana Raouafi --- see the specific paragraph on this person towards the end of this post.<br />
<br />
3. Whether FAES is a co-author or not, these researchers have a remarkable taste for citing his published work, which typically accounts for between 50% and 100% of the References section of any of their articles.<br />
<br />
4. When one of these researchers, rather than FAES, is the corresponding author of an article, they always use a Yahoo or Gmail address. So far I have identified "s.bern@yahoo.com", "mullerhito@yahoo.com", "mitsunaka216@gmail.com", and "okitonaka216@gmail.com". None of these researchers seems to use an institutional e-mail address for correspondence. Of course, this is not entirely illegitimate (for example, if one anticipates moving to another institution in the near future), but it seems quite unusual for none of them to have used their faculty address.<br />
<br />
[[ Update 2020-04-27 17:22 UTC: I have identified that "Erin Regai", who I think is the same person as "Eryn Rekgai" but with slightly different spelling, has the e-mail address "eregai216@gmail.com". That makes three people with Gmail addresses ending in 216. It would be interesting to discover whether anybody involved in these authors' publication projects has a birthday on 21/6 (21 June) or 2/16 (February 16). ]]<br />
<br />
5. None of these people seems to have an entry in the online staff directory of their respective institutions. (The links under their names at the start of this post all go to their respective ResearchGate profiles, or if they don't have one, RG's collection of their "scientific contributions".) Of course, one can never prove a negative, and some people just prefer a quiet life. So as part of this blog post I am issuing a public appeal: If you know (or, even better, if you are) any of these people, please get in touch with me.<br />
<br />
I don't have time to go into all of these individuals in detail, but here are some highlights of what I found in a couple of cases. (For the two authors named Nakamura, I am awaiting a response to inquiries that I sent to their respective institutions; I hope that readers will forgive me for publishing this post before waiting for a reply to those inquiries, given the current working situation at many universities around the world.)<br />
<br />
[[ Update 2020-04-24 21:24 UTC: Ms. Mariko Kajii of the Office of Global Planning and Partnerships at The Ritsumeikan Trust has confirmed to me that nobody named "Okito Nakamura" is known to that institution. ]]<br />
<br />
[[ Update 2020-04-24 23:37 UTC: Mitsu Nakamura's <a href="https://www.researchgate.net/profile/Mitsu_Nakamura">ResearchGate page</a> claims that Okito Naakmura is a member of Mitsu's lab at "The graduated [<i>sic</i>] University of Advanced studies". It seems strange that someone would be affiliated with one university (even if that university denied any knowledge of them, cf. my previous update) while working in a lab at another. Meanwhile, Mitsu Nakamura's name does not appear in Japan's national database of researchers. ]]<br />
<br />
<u>Muller G. Hito</u><br />
<u><br /></u>
For this researcher --- who does not seem to be quite sure how their own name is structured(*), as they sometimes appear at the top of an article as "Hito G. Muller" --- we have quite extensive contact information, for example in <a href="https://translational-neuroscience.imedpub.com/a-novel-reflection-on-stigmatization.php?aid=21321" target="_blank">this article</a> (which cites 18 references, 12 of them authored by FAES).<br />
<br />
<div class="separator" style="clear: both; text-align: center;">
<a href="http://3.bp.blogspot.com/-ZSRkeAJnzdQ/XqHInGJPG9I/AAAAAAAAFGg/aiETPPtIPvgFHar84PtCpfADnMrGCdrqQCK4BGAYYCw/s1600/Hito.png" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" height="323" src="https://3.bp.blogspot.com/-ZSRkeAJnzdQ/XqHInGJPG9I/AAAAAAAAFGg/aiETPPtIPvgFHar84PtCpfADnMrGCdrqQCK4BGAYYCw/s640/Hito.png" width="640" /></a></div>
I looked up that phone number and found that <a href="https://www.uni-giessen.de/fbz/fb06/sport/arbe/trawis/mitarb/mueller" target="_blank">it does indeed belong to someone</a> in the Department of Psychology and Sports Science at Justus-Liebig University, namely Prof. Dr. Hermann Müller. For a moment I thought that maybe Prof. Dr. Müller likes to call himself "Hito", and maybe he got his first and last names mixed up when correcting the proofs of his article. But as my colleague Malte Elson points out, no German person named "Müller" would ever allow their name to be spelled "Muller" without the umlaut. (In situations where the umlaut is not available, for example in an e-mail address, it is compensated for by adding an e to the vowel, e.g., in this case, "Mueller".)<br />
<br />
In any case, Malte contacted Prof. Dr. Müller, who assured him that he is not "Hito D. Muller" or "Muller D. Hito". Nor has Dr. Müller ever heard of anyone with that name, or anyone with a name like "Eryn Rekgai", in the department where he works.<br />
<br />
<u>Bern S. Schmidt</u><br />
<br />
Bern Schmidt is another author who likes to permute the components of their name. They have published articles as "Bern S. Schmidt", "Bern Schmidt S.", "Bern, SS", and perhaps other combinations. Their <a href="https://www.imedpub.com/editor-profile/Dr_Bern_S_Schmidt/">bio on their author page</a> on the web site of Insight Medical Publishing, which publishes a number of the journals that contain the articles that are linked to throughout this post, says:<br />
<blockquote class="tr_bq">
Dr Bern S. Schmidt is a neuroscientist and clinical tenure track [<i>sic</i>] of the CHUV, working in the area of fundamental neuroscience and psychobiological factors influencing appearance of central nervous disorders and neurodegenerative disorders such as Alzheimer and Dementia. He holds a medical degree at the University of Victoria, follow by a residency program at The Shiga University of Medicine and a postdoctoral internship at the Waseda University.</blockquote>
I assume that "CHUV" here refers to "Centre Hospitalier Universitaire Vaudois", the teaching hospital of the University of Lausanne where Dr. Schmidt claims to be affiliated in the Department of Fundamental Neuroscience. But a search of the university's web site did not find any researcher with this name. I asked somebody who has access to a directory of all past and present staff members of the University of Lausanne if they could find anyone with a name that corresponds even partially to this name, and they reported that they found nothing. Meanwhile, The University of Victoria has no medical degree programme, and their neuroscience programme has no trace of anyone with this name.<br />
<br />
[[ Update 2020-04-27 17:24 UTC: A representative of the University of Lausanne has confirmed to me that they can find no trace of anybody named "Bern Schmidt" at their institution. ]]<br />
<br />
(A minor detail, but one that underscores how big a rabbit hole this story is: Dr. Schmidt seems to have an unusual telephone number. <a href="https://www.researchgate.net/publication/325968975_Stress_Two_Sides_of_a_Coin">This article</a> lists it as "516-851-8564", which looks more like a North American number than a Swiss one. Indeed, it is identical to the number given in <a href="https://translational-neuroscience.imedpub.com/zincfinger-proteins-in-brain-development-and-mental-illness.php?aid=22525">this apparently unrelated article in the same journal</a> for the corresponding author Hong Li of the Department of Neuroscience at Yale University School of Medicine. Dr. Hong Li's doubtless vital --- after all, she is at Yale --- contribution to neuroscience research was accepted within 6 days of being submitted, presumably having been pronounced flawless by the prominent scholars who performed the rigorous peer review process for the prestigious <i>Journal of Translational Neurosciences</i>. It is, however, slightly disappointing that typesetting standard at this paragon of scientific publishing do not extend to removing one author's phone number when typesetting the next one to be published on the same day. If anyone knows where Dr. Bern Schmidt is, perhaps they could mention this to them, so that this important detail can be corrected. We wouldn't want Dr. Hong Li's valuable Yale neuroscientist time to be wasted answering calls intended for Dr. Schmidt.<br />
<br />
<u>These authors' recent dataset</u><br />
<br />
The only activity that I have been able to identify from any of these authors in the last few months is the publication of <a href="https://data.mendeley.com/datasets/c53rh2h435/1">this dataset</a>, which was uploaded to Mendeley on March 22, 2020. As well as FAES, the authors are listed as HG Muller, E. Regai [<i>sic</i>], and O. Nakamura. From the "Related links" on that page, it appears that this dataset is a subset (total <i>N</i>=750) of the 10,566 cases that make up the sample described in the Etindele Sosso et al. article in <i>Nature Scientific Reports</i> that was the subject of my previous blog post.<br />
<br />
However, a few things about these data are not entirely consistent with that article. For example, while the per-country means for the variables "Age", "Mean Hours of Gaming/week", and "Mean months of gaming/gamer" correspond to the published numbers in Table 2 of the article in five out of six cases (for "Mean months of gaming/gamer" in the sample from Gabon the mean is 15.77, whereas in the article the integer-rounded value reported was 15), all of the standard deviations in the dataset are considerably higher than those that were published, by factors ranging from 1.3 to 5.1.<br />
<br />
Furthermore, there are some patterns in the distribution of scores in the four outcome variables (ISI, EDS, HADS-A subscale, and HADS-D subscale) that are difficult to explain as being the results of natural processes. For all four of these measures in the <i>N</i>=301 sample from Tunisia, and three of them (excluding the EDS) in the <i>N</i>=449 sample from Gabon, between 77% and 92% of the individual participants' total scores on each of these the subscales are even numbers. For the EDS in the sample from Gabon, 78% of the scores are odd numbers. In the Gabon sample, it is also noticeable that the ISI score for every participant is exactly 2 higher than their HADS-A score and exactly 3 higher than their EDS score; the HADS-A score is also 2 higher than the HADS-D for 404 out of 449 participants.<br />
<br />
It is not clear to me why Hito G. Muller, Eryn Re[k]gai, and Okito Nakamura might be involved with the publication of this dataset, when their names were not listed as authors of the published article. But perhaps they have very high ethical standards and did not feel that their contribution to the curation of the data, whatever that might have been, merited a claim of authorship in <a href="https://www.nature.com/srep/about">the 11th most highly cited scientific journal in the world</a>.<br />
<div>
<br /></div>
<u>The other author who does seem to exist</u><br />
<br />
There is one co-author on a few of the articles mentioned above who does actually appear to exist. This is Sana Raouafi, who reports an affiliation with the Department of Biomedical Engineering at the <a href="https://en.wikipedia.org/wiki/Polytechnique_Montr%C3%A9al">Polytechnique Montréal</a>. The records office of that institution informed me that she was awarded her PhD on January 27, 2020. I have no other contact information for her, nor do I know whether she genuinely took part in the authorship of these strange articles, or what her relationship with FAES (or, if they exist, any of the other co-authors) might be.<br />
<br />
<u>Supporting file</u><br />
<u><br /></u>
There is one supporting file for this post <a href="http://nickbrown.fr/blog/mystery-authors">here</a>:<br />
- <span style="text-indent: -24px;"><span style="font-family: "times new roman"; font-size: 7pt; font-stretch: normal; line-height: normal;"> </span></span><span style="font-family: "courier new"; font-size: 11pt; line-height: 22px; text-indent: -24px;">Muller-dataset-with-pivots.xls</span><span style="text-indent: -24px;">: An </span>Excel file containing my analyses of the Muller et al. dataset, mentioned above in the section "These authors' recent dataset". The basic worksheets from the published dataset have been enhanced with two sheets of pivot tables, illustrating the issues with the outcome measures that I described.<br />
<br />
<u>Acknowledgements</u><br />
<u><br /></u>
Thanks to Elisabeth Bik, Malte Elson, Danny Garside, Steve Lindsay, Stuart Ritchie, and Yannick Rochat for their help in attempting to track down these elusive researchers. Perhaps others will have more luck than us.<br />
<br />
<br />
(*) I am aware that different customs exist in different countries regarding the order in which "given" and "family" names are written. For example, in several East Asian countries, but also in Hungary, it is common to write the family name first. Interestingly, there is often some ambiguity about this among speakers of French. But as far as I know, German speakers, like English speakers, always use put their given name first and their family name last, unless there is a requirement to invert this order for alphabetisation purposes. And of course, in some parts of the world, the whole idea of "family names" is much more complicated than in Western countries. It's a fascinating subject that, alas, I do not have time to explore here.<br />
<br />
<br />
<br /></div>
Nick Brownhttp://www.blogger.com/profile/18266307287741345798noreply@blogger.com10tag:blogger.com,1999:blog-7890764972166411105.post-45750985920777638692020-04-21T13:20:00.000+02:002020-06-04T19:39:23.215+02:00Some issues in a recent gaming research article: Etindele Sosso et al. (2020)<div dir="ltr" style="text-align: left;" trbidi="on">
<br />
<div class="MsoNormal" style="line-height: 150%;">
Research into the possibly
problematic aspects of gaming is a hot topic. But most studies in this area have
focused on gamers in Europe and North America. So a recent article in <i style="mso-bidi-font-style: normal;">Nature Scientific Reports</i>, featuring
data from over 10,000 African gamers, would seem to be an important landmark
for this field. However, even though I am an outsider to gaming research, it
seems to my inexpert eye that this article may have a few wrinkles that need
ironing out.<br />
<br />
Let’s start with the article
reference. It has 16 authors, and the new edition of the APA Publication Manual
says that we now have to list <a href="https://apastyle.apa.org/blog/more-than-20-authors">up to 20 authors’ names in a reference</a>,
so let’s take a deep breath:</div>
<div class="MsoNormal" style="line-height: 150%;">
<br /></div>
<div class="MsoNormal" style="line-height: 150%; margin-left: 36.0pt; text-indent: -36.0pt;">
<span style="mso-no-proof: yes;">Etindele Sosso, F. A., Kuss, D. J.,
Vandelanotte, C., Jasso-Medrano, J. L., Husain, M. E., Curcio, G., Papadopoulos,
D., Aseem, A., Bhati, P., Lopez-Rosales, F., Ramon Becerra, J., D’Aurizio, G., Mansouri,
H., Khoury, T., Campbell, M., & Toth, A. J.</span> (2020). Insomnia,
sleepiness, anxiety and depression among different types of gamers in African
countries. <i style="mso-bidi-font-style: normal;">Nature Scientific Reports</i>,
<i style="mso-bidi-font-style: normal;">10</i>, 1937. <a href="https://doi.org/10.1038/s41598-020-58462-0">https://doi.org/10.1038/s41598-020-58462-0</a></div>
<div class="MsoNormal" style="line-height: 150%;">
(The good news is that it is an open
access article, so you can just follow the DOI link and download the PDF file.)</div>
<div class="MsoNormal" style="line-height: 150%;">
<br /></div>
<div class="MsoNormal" style="line-height: 150%;">
Etindele Sosso et al. (2020)
investigated the association between gaming and the four health outcomes
mentioned in the title. According to the abstract, the results showed that “problematic
and addicted gamers show poorer health outcomes compared with non-problematic
gamers”, which sounds very reasonable to me as an outsider to the field. A
survey that took about 20 minutes to complete was e-mailed to 53,634
participants, with a 23.64% response rate. After eliminating duplicates and
incomplete forms, a total of 10,566 gamers were used in the analyses. The “type
of gamer” of each participant was classified as “non-problematic”, “engaged”,
“problematic”, or “addicted”, depending on their scores on a measure of gaming
addiction, and the relations between this variable, other demographic
information, and four health outcomes were examined.</div>
<div class="MsoNormal" style="line-height: 150%;">
<br /></div>
<div class="MsoNormal" style="line-height: 150%;">
The 16 authors of the Etindele
Sosso et al. (2020) article report affiliations at 12 different institutions in
8 different countries. According to the “Author contributions” section, the
first three authors “contributed equally to this work” (I presume that this
means that they did the majority of it); 12 others (all except Papadopoulos, it
seems) “contributed to the writing”; the first three authors plus Papadopoulos
“contributed to the analyses”; and five (the first three authors, plus Campbell
and <span style="mso-no-proof: yes;">Toth)</span> “write [<i style="mso-bidi-font-style: normal;">sic</i>] the final form of the manuscript”. So this is a very
impressive international collaboration, with the majority of the work apparently
being split between Canada, the UK, and Australia, and it ought to represent a
substantial advance in our understanding of how gaming affects mental and
physical health in Africa.</div>
<div class="MsoNormal" style="line-height: 150%;">
<br /></div>
<div class="MsoNormal" style="line-height: 150%;">
<u>Funding<o:p></o:p></u></div>
<div class="MsoNormal" style="line-height: 150%;">
Given the impressive set of authors
and the large scale of this international project (data collection alone took
19 or 20 months, from November 2015 to June 2017), it is somewhat surprising that
Etindele Sosso et al.’s (2020) article reports no source of funding. Perhaps
everyone involved contributed their time and other resources for free, but there
is not even a statement that no external funding was involved. (I am quite
surprised that this last element is apparently not mandatory for articles in
the <i style="mso-bidi-font-style: normal;">Nature</i> family of journals.) The administrative
arrangements for the study, involving for example contacting the admissions
offices of universities in nine countries and arranging for their e-mail lists
to be made available, with appropriate guarantees that each university’s and
country’s standards of research ethics would be respected, must have been
considerable. The participants completed an online questionnaire, which might
well have involved some monetary cost, whether directly paid to a survey
hosting company or using up some part of a university’s agreed quota with such
a company. Just publishing an Open Access article in <i style="mso-bidi-font-style: normal;">Nature Scientific Reports</i> costs, according to the journal’s web
site, $1,870 plus applicable taxes.</div>
<div class="MsoNormal" style="line-height: 150%;">
<br /></div>
<div class="MsoNormal" style="line-height: 150%;">
<u>Ethical approval<o:p></o:p></u></div>
<div class="MsoNormal" style="line-height: 150%;">
One possible explanation for the absence
of funding information—although this would still constitute rather sloppy
reporting, since as noted in the previous paragraph funding typically doesn’t
just pay for data collection—might be if the data had already been collected as
part of another study. No explicit statement to this effect is made in the Etindele
Sosso et al. (2020) article, but at the start of the Methods section, we find
“This is a secondary analysis of data collected during the project MHPE
approved by the Faculty of Arts and Science of the University of Montreal
(CERAS-2015-16-194-D)”. So I set out to look for any information about the
primary analysis of these data.</div>
<div class="MsoNormal" style="line-height: 150%;">
<br /></div>
<div class="MsoNormal" style="line-height: 150%;">
I searched online to see if
“project MHPE” might perhaps be a large data collection initiative from the
University of Montreal, but found nothing. However, in the lead author’s
Master’s thesis, submitted in March 2018 (full text PDF file available <a href="https://papyrus.bib.umontreal.ca/xmlui/handle/1866/20345">here</a>—note
that, apart from the Abstract, the entire document is written in French, but
fortunately I am fluent in that language), we find that “MHPE” stands for
“Mental Health profile [<i style="mso-bidi-font-style: normal;">sic</i>] of
Etindele” (p. 5), and that the research in that thesis was covered by a
certificate from the ethical board of the university that carries exactly the
same reference number. I will therefore tentatively conclude that this is the
“project MHPE” referred to in the Etindele Sosso et al. (2020) article.</div>
<div class="MsoNormal" style="line-height: 150%;">
<br /></div>
<div class="MsoNormal" style="line-height: 150%;">
However, the Master’s thesis
describes how data were collected from a sample (prospective size,
12,000–13,000; final size 1,344) of members of the University of Montreal
community, collected between November 2015 and December 2016. The two studies—i.e.,
the one reported in the Master’s thesis and the one reported by Etindele et. al
(2020)—each used five measures, of which only two—the Insomnia Severity Index
(ISI) and the Hospital Anxiety and Depression Scale (HADS)—were common to both.
The questionnaires administered to the participants in the Montreal study
included measures of cognitive decline and suicide risk, and it appears from p.
27, line 14 of the Master’s thesis that participants were also interviewed
(although no details are provided of the interview procedure). All in all, the
ethical issues involved in this study would seem to be rather different to
those involved in asking people by e-mail about their gaming habits. Yet it
seems that the ethics board gave its approval, on a single certificate, for the
collection of two sets of data from two distinct groups of people in two very
different studies: (a) a sample of around 12,000 people from the lead author’s
local university community, using repeated questionnaires across a four-month
period as well as interviews; and (b) a sample of 50,000 people spread across
the continent of Africa, using e-mail solicitation and an online questionnaire.
This would seem to be somewhat unusual.</div>
<div class="MsoNormal" style="line-height: 150%;">
<br /></div>
<div class="MsoNormal" style="line-height: 150%;">
Meanwhile, we are still no nearer
to finding out who funded the collection of data in Africa and the time taken
by the other authors to make their (presumably extensive, in the case of the
second and third authors) personal contributions to the project. On p. 3 of his
Master’s thesis, the author thanks (translation by me) “The Department of
Biological Sciences and the Centre for Research in Neuropsychology and
Cognition of the University of Montreal, which provided logistical and
financial support to the success of this work”, but it is not clear that “this
work” can be extrapolated beyond the collection of data in
Montreal to include the African project. Nor do we have any more idea about why
Etindele Sosso et al. (2020) described their use of the African data as a "secondary analysis", when it seems, as far as I have been able to establish,
that there has been no previously published (primary) analysis of this data set.</div>
<div class="MsoNormal" style="line-height: 150%;">
<br /></div>
<div class="MsoNormal" style="line-height: 150%;">
<u>Results<o:p></o:p></u></div>
<div class="MsoNormal" style="line-height: 150%;">
Further questions arise when we
look at the principal numerical results
of Etindele Sosso et al.’s (2020) article. On p. 4, the authors report that “4
multiple linear regression analyses were performed (with normal gaming as
reference category) to compare the odds for having these conditions [i.e.,
insomnia, sleepiness, anxiety, and depression] (which are dependent variables)
for different levels of gaming.” I’m not sure why the authors would perform
linear, as opposed to logistic, regressions to compare the odds of someone in a
given category having a specific condition relative to someone in a reference
category, but that’s by no means the biggest problem here.</div>
<div class="MsoNormal" style="line-height: 150%;">
<br /></div>
<div class="MsoNormal" style="line-height: 150%;">
Etindele Sosso et al.’s (2020) Table 3 lists, for each of the four
health outcome variables, the regression coefficients and associated test
statistics for each of the predictors in their study. Before
we come to these numbers for individual variables, however, it is worth looking
at the R-squared numbers for each model, which range from .76 for depression to
.89 for insomnia. Although these are actually labelled as “ΔR<sup>2</sup>”, I
assume that they represent the total variance explained by the whole model,
rather than a change in R-squared when “type of gamer” is added to the model
that contains only the covariates. (That said, however, the sentence “Gaming
significantly contributed to 86.9% of the variance in insomnia, 82.7% of the
variance in daytime sleepiness and 82.3% of the variance in anxiety [p <
0.001]” in the Abstract does not make anything much clearer.) But whether these
numbers represent the variance explained by the whole model or just by the
“type of gamer” variable, they constitute remarkable results by any standard. I
wonder if anything in the prior sleep literature has ever predicted 89% of the
variance explained by a measure of insomnia, apart perhaps from another measure
of insomnia.</div>
<div class="MsoNormal" style="line-height: 150%;">
<br /></div>
<div class="MsoNormal" style="line-height: 150%;">
Now let’s look at the details of
Table 3. In principle there are seven variables (“Type of Gamers [<i style="mso-bidi-font-style: normal;">sic</i>]” being the main one of interest, plus
the demographic covariates Age, Sex, Education, Income, Marital status, and Employment
status), but because all of these are categorical, each of the levels except
the reference category will have been a separate predictor in the regression,
giving a total of 17 predictors. Thus, across the four models, there are 68
lines in total reporting regression coefficients and other associated statistics.
The labels of the columns seem to be what one would expect from reports of
multiple regression analyses: B (unstandardized regression coefficient), SE
(standard error, presumably of B), β (standardized regression coefficient), t
(the ratio between B and SE), Sig (the <i style="mso-bidi-font-style: normal;">p</i>
value associated with t), and the upper and lower bounds of the 95% confidence
interval (again, presumably of B).</div>
<div class="MsoNormal" style="line-height: 150%;">
<br /></div>
<div class="MsoNormal" style="line-height: 150%;">
The problem is that none of the actual
numbers in the table seem to obey the relations that one would expect. In fact
I cannot find a way in which any of them make any sense at all. Here are the
problems that I identified:</div>
<div class="MsoNormal" style="line-height: 150%; margin-left: 36.0pt; mso-list: l0 level1 lfo1; tab-stops: list 36.0pt; text-indent: -18.0pt;">
<!--[if !supportLists]--><span style="mso-list: Ignore;">-<span style="font: 7.0pt "Times New Roman";"> </span></span>When I compute the ratio B/SE, and compare it to column
t (which should give the same ratio), the two don’t even get close to being
equal in any of the 68 lines. Dividing the B/SE ratio by column t gives results
that vary from 0.0218 (Model 2, Age, 30–36) to 44.1269 (Model 1, Type of Gamers,
Engaged), with the closest to 1.0 being 0.7936 (Model 4, Age, 30–36) and 1.3334
(Model 3, Type of Gamers, Engaged).</div>
<div class="MsoNormal" style="line-height: 150%; margin-left: 36.0pt; mso-list: l0 level1 lfo1; tab-stops: list 36.0pt; text-indent: -18.0pt;">
<!--[if !supportLists]--><span style="mso-list: Ignore;">-<span style="font: 7.0pt "Times New Roman";"> </span></span>Perhaps SE refers to the standard error of the <i style="mso-bidi-font-style: normal;">standardized</i> regression coefficient (β),
even though the column SE appears to the left of the column β? Let’s divide β
by SE and see how the <i style="mso-bidi-font-style: normal;">t</i> ratio
compares. Here, we get results that vary from 0.0022 (Model 2, Age, 30–36) to
11.7973 (Model 1, Type of Gamers, Engaged). The closest we get to 1.0 is with
values of 0.7474 (Model 3, Marital Status, Engaged) and 1.0604 (Model 3,
Marital Status, Married). So here again, none of the β/SE calculations comes
close to matching column t.</div>
<div class="MsoNormal" style="line-height: 150%; margin-left: 36.0pt; mso-list: l0 level1 lfo1; tab-stops: list 36.0pt; text-indent: -18.0pt;">
<!--[if !supportLists]--><span style="mso-list: Ignore;">-<span style="font: 7.0pt "Times New Roman";"> </span></span>The <i style="mso-bidi-font-style: normal;">p</i> values
do not match the corresponding <i style="mso-bidi-font-style: normal;">t</i>
statistics. In most cases this can be seen by simple inspection. For example,
on the first line of Table 3, it should be clear that a <i style="mso-bidi-font-style: normal;">t</i> statistic of 9.748 would have a very small <i style="mso-bidi-font-style: normal;">p</i> value indeed (in fact, about 1E−22) rather than .523. In many
cases, even the conventional statistical significance status (either side of <i style="mso-bidi-font-style: normal;">p</i> = .05) of the <i style="mso-bidi-font-style: normal;">t</i> value doesn’t match the <i style="mso-bidi-font-style: normal;">p</i> value. To get an idea of this, I made
the simplifying assumption (which is not actually true for the categories “Age:
36–42”, “Education: Doctorate”, and “Marital status: Married”, but individual
inspection of these shows that my assumption doesn’t change much) that all
degrees of freedom were at least 100, so that any <i style="mso-bidi-font-style: normal;">t</i> value with a magnitude greater than 1.96 would be statistically
significant at the .05 level. I then looked to see if <i style="mso-bidi-font-style: normal;">t</i> and <i style="mso-bidi-font-style: normal;">p</i> were the same
side of the significance threshold; they were not in 29 out of 68 cases.</div>
<div class="MsoNormal" style="line-height: 150%; margin-left: 36.0pt; mso-list: l0 level1 lfo1; tab-stops: list 36.0pt; text-indent: -18.0pt;">
<!--[if !supportLists]--><span style="mso-list: Ignore;">-<span style="font: 7.0pt "Times New Roman";"> </span></span>The regression coefficients are not always contained
within their corresponding confidence intervals. This is the case for 29 out of
68 of the B (unstandardized) values. I don’t think that the confidence
intervals are meant to refer to the standardized coefficients (β), but just for
completeness, 63 out of 68 of these fall outside the reported 95% CI.</div>
<div class="MsoNormal" style="line-height: 150%; margin-left: 36.0pt; mso-list: l0 level1 lfo1; tab-stops: list 36.0pt; text-indent: -18.0pt;">
<!--[if !supportLists]--><span style="mso-list: Ignore;">-<span style="font: 7.0pt "Times New Roman";"> </span></span>Whether the regression coefficient falls inside the 95%
CI does not correspond with whether the <i style="mso-bidi-font-style: normal;">p</i>
value is below .05.<span style="mso-spacerun: yes;"> </span>For both the
unstandardized coefficients (B) and the standardized coefficients (β)—which,
again, the CI probably doesn’t correspond to, but it’s quick and cheap to look
at the possibility anyway—this test fails in 41 out of 68 cases.</div>
<div class="MsoNormal" style="line-height: 150%;">
<br /></div>
<div class="MsoNormal" style="line-height: 150%;">
There are some further concerns
with Table 3:</div>
<div class="MsoNormal" style="line-height: 150%; margin-left: 36.0pt; mso-list: l0 level1 lfo1; tab-stops: list 36.0pt; text-indent: -18.0pt;">
<!--[if !supportLists]--><span style="mso-list: Ignore;">-<span style="font: 7.0pt "Times New Roman";"> </span></span>In the third line (Model 1, “Type of Gamers”,
“Problematic”) the value for β is 1.8. Now it is actually possible to have a
standardized regression coefficient with a magnitude above 1.0, but its
existence usually means that you have big multicollinearity problems, and it’s
typically very hard to interpret such a coefficient. It’s the kind of thing
that at least one of the four authors who reported in the "Author contributions" section of the article that they "contributed to the analyses" would normally be expected to pick up on and discuss, but no such discussion is
to be found.</div>
<div class="MsoNormal" style="line-height: 150%; margin-left: 36.0pt; mso-list: l0 level1 lfo1; tab-stops: list 36.0pt; text-indent: -18.0pt;">
<!--[if !supportLists]--><span style="mso-list: Ignore;">-<span style="font: 7.0pt "Times New Roman";"> </span></span>From Table 1, we can see that there were zero
participants in the “Age” category 42–48, and zero participants in the
“Education” category “Postdoctorate”. Yet, in Table 3, for all four models,
these categories have non-zero regression coefficients and other statistics. It
is not clear to me how one can obtain a regression coefficient or standard
error from a categorical variable that corresponds to zero cases (and, hence,
when coded has a mean and standard deviation of 0).</div>
<div class="MsoNormal" style="line-height: 150%; margin-left: 36.0pt; mso-list: l0 level1 lfo1; tab-stops: list 36.0pt; text-indent: -18.0pt;">
<!--[if !supportLists]--><span style="mso-list: Ignore;">-<span style="font: 7.0pt "Times New Roman";"> </span></span>There is a surprisingly high number of repetitions of exactly
the same value, typically to 3 decimal places, within the same variable,
category, and absolute value of the statistic from one model to another. For example, the reported
value in the column t for the variable “Age” and category “24–30” is 29.741 in
both Models 1 and 3. For the variable “Employment status” and category
“Employed”, the upper bound of the 95% confidence interval is the same (2.978)
in all four models. This seems quite unlikely to be the result of chance, given
the relatively large sample sizes that are involved for most of the categories
(cf. Brown & Heathers, 2019), so it is not clear how these duplicates could
have arisen.</div>
<div class="MsoNormal" style="line-height: 150%;">
<br /></div>
<div class="MsoNormal" style="line-height: 150%;">
<!--[if gte vml 1]><v:shapetype
id="_x0000_t75" coordsize="21600,21600" o:spt="75" o:preferrelative="t"
path="m@4@5l@4@11@9@11@9@5xe" filled="f" stroked="f">
<v:stroke joinstyle="miter"/>
<v:formulas>
<v:f eqn="if lineDrawn pixelLineWidth 0"/>
<v:f eqn="sum @0 1 0"/>
<v:f eqn="sum 0 0 @1"/>
<v:f eqn="prod @2 1 2"/>
<v:f eqn="prod @3 21600 pixelWidth"/>
<v:f eqn="prod @3 21600 pixelHeight"/>
<v:f eqn="sum @0 0 1"/>
<v:f eqn="prod @6 1 2"/>
<v:f eqn="prod @7 21600 pixelWidth"/>
<v:f eqn="sum @8 21600 0"/>
<v:f eqn="prod @7 21600 pixelHeight"/>
<v:f eqn="sum @10 21600 0"/>
</v:formulas>
<v:path o:extrusionok="f" gradientshapeok="t" o:connecttype="rect"/>
<o:lock v:ext="edit" aspectratio="t"/>
</v:shapetype><v:shape id="_x0000_i1025" type="#_x0000_t75" style='width:252.75pt;
height:697.5pt'>
<v:imagedata src="file:///C:\Users\Nick\AppData\Local\Temp\msohtml1\01\clip_image001.png"
o:title=""/>
</v:shape><![endif]--><!--[if !vml]--><!--[endif]--></div>
<div class="separator" style="clear: both; text-align: center;">
<a href="http://4.bp.blogspot.com/-RYBKCfAypjg/Xp7WIgH1GoI/AAAAAAAAFFo/BzISXBAdeXQ_J_MJnoJoxDefSSUdGORoACK4BGAYYCw/s1600/Etindele-et-al-Table3-duplicates.png" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" height="640" src="https://4.bp.blogspot.com/-RYBKCfAypjg/Xp7WIgH1GoI/AAAAAAAAFFo/BzISXBAdeXQ_J_MJnoJoxDefSSUdGORoACK4BGAYYCw/s640/Etindele-et-al-Table3-duplicates.png" width="232" /></a></div>
<div class="MsoNormal" style="line-height: 150%;">
<br /></div>
<div class="MsoNormal" style="line-height: 150%;">
<span style="font-size: 10.0pt; line-height: 150%;">Table 3 from Etindele et al. (2020), with duplicated values (considering
the same variable and category across models) highlighted with a different colour for each set of duplicates. Two pairs are included where the
sign changed but the digits remained identical; however, <i style="mso-bidi-font-style: normal;">p</i> values that were reported as 0.000 are ignored. To
find a duplicate, first identify a cell that is outlined in a particular
colour, then look up or down the table for one or more other cells with the same
outline colour in the analogous position for one or more other models.<o:p></o:p></span></div>
<div class="MsoNormal" style="line-height: 150%;">
<br /></div>
<div class="MsoNormal" style="line-height: 150%;">
<u>The preprint <o:p></o:p></u></div>
<div class="MsoNormal" style="line-height: 150%;">
It is interesting to compare
Etindele Sosso et al.’s (2020) article with a preprint entitled “Insomnia and
problematic gaming: A study in 9 low- and middle-income countries” by Faustin
Armel Etindele Sosso and Daria J. Kuss (who also appears to be the second
author of the published article), which is available <a href="https://www.biorxiv.org/content/10.1101/451724v1">here</a>.
That preprint reports a longitudinal study, with data collected at multiple time
points—presumably four, including baseline, although only “after one months,
six months, and 12 months” (p. 8) is mentioned—from a sample of people (initial
size 120,460) from nine African countries. This must therefore be an entirely
different study from the one reported in the published article, which did not
use a longitudinal design and had a prospective sample size of 53,634. Yet, by
an astonishing coincidence, the final sample retained for analysis in the
preprint consisted of 10,566 participants, which is exactly the same as the
published article. The number of men (9,366) and women (1,200) was also
identical in the two samples. However, the mean and standard deviation of their
ages was different (M=22.33 years, SD=2.0 in the preprint; M=24.0, SD=2.3 in
the published article). The number of participants in each of the nine
countries (Table 2 of both the preprint and the published article) is also substantially
different for each country between the two papers, and with two exceptions—the
ISI and the well-known Hospital Anxiety and Depression Scale (HADS)—different
measures of symptoms and gaming were used in each case.</div>
<div class="MsoNormal" style="line-height: 150%;">
<br /></div>
<div class="MsoNormal" style="line-height: 150%;">
Another remarkable coincidence
between the preprint and Etindele Sosso et al.’s (2020) published article,
given that we are dealing with two distinct samples, occurs in the
description of the results obtained from the sample of African gamers on the
Insomnia Severity Index. On p. 3 of the published article, in the
paragraph describing the respondents’ scores on the ISI, we find: “The internal
consistency of the ISI was excellent (Cronbach’s α = 0.92), and each individual
item showed adequate discriminative capacity (<i style="mso-bidi-font-style: normal;">r</i> = 0.65–0.84). The area under the receiver operator characteristic
curve was 0.87 and suggested that a cut-off score of 14 was optimal (82.4%
sensitivity, 82.1% specificity, and 82.2% agreement) for detecting clinical
insomnia”. These two sentences are identical, in every word and number, to the
equivalent sentences on p. 5 of the preprint.</div>
<div class="MsoNormal" style="line-height: 150%;">
<br /></div>
<div class="MsoNormal" style="line-height: 150%;">
Naturally enough, because the
preprint and Etindele Sosso et al.’s (2020) published article describe entirely
different studies with different designs, and different sample sizes in each
country, there is little in common between the Results sections of the two
papers. The results in the preprint are based on repeated-measures analyses and
include some interesting full-colour figures (the depiction of correlations in
Figure 1, on p. 10, is particularly visually attractive), whereas the
results of the published article consist mostly of a fairly straightforward
summary, in sentences, of the results from the tables, which describe the
outputs of linear regressions.</div>
<div class="MsoNormal" style="line-height: 150%;">
<o:p><br /></o:p></div>
<div class="separator" style="clear: both; text-align: center;">
<a href="http://3.bp.blogspot.com/--oHbEH8DPc4/Xp7ViN0S4gI/AAAAAAAAFFM/l4sPRWZMJmYle9FnadyWC3nSBkb8fAKKgCK4BGAYYCw/s1600/Etindele%2BSosso-Kuss-Preprint-Figure1.png" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" height="528" src="https://3.bp.blogspot.com/--oHbEH8DPc4/Xp7ViN0S4gI/AAAAAAAAFFM/l4sPRWZMJmYle9FnadyWC3nSBkb8fAKKgCK4BGAYYCw/s640/Etindele%2BSosso-Kuss-Preprint-Figure1.png" width="640" /></a></div>
<div class="MsoNormal" style="line-height: 150%; text-align: left;">
<span style="font-size: 10.0pt; line-height: 150%;"><br /></span></div>
<div class="MsoNormal" style="line-height: 150%; text-align: left;">
<span style="font-size: 10.0pt; line-height: 150%;">Figure 1 from the preprint by
Etindele Sosso and Kuss (2018, p. 10). This appears to use an innovative technique to illustrate the correlation between two variables.</span></div>
<div align="center" class="MsoNormal" style="line-height: 150%; text-align: center;">
<br /></div>
<div class="MsoNormal" style="line-height: 150%;">
However, approximately 80% of the
sentences in the introduction of the published article, and 50% of the
sentences in the Discussion section, appear (with only a few cosmetic changes)
in the preprint. This is interesting, not only because it would be quite
unusual for a preprint of one study to be repurposed to describe en entirely
different one, but also because it suggests that the addition of 14 authors between the publication of the preprint and the Etindele Sosso et al. (2020) article resulted
in the addition of only about 1,000 words to these two parts of the manuscript.</div>
<div class="separator" style="clear: both; text-align: center;">
<a href="http://1.bp.blogspot.com/-UoAmqbO3D2o/Xp7VHYKDSBI/AAAAAAAAFE0/bkhDVTXEh0kLYEd5iIQ59uWFxS2rfQKewCK4BGAYYCw/s1600/Preprint-article-intro-side-by-side.png" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" height="640" src="https://1.bp.blogspot.com/-UoAmqbO3D2o/Xp7VHYKDSBI/AAAAAAAAFE0/bkhDVTXEh0kLYEd5iIQ59uWFxS2rfQKewCK4BGAYYCw/s640/Preprint-article-intro-side-by-side.png" width="430" /></a></div>
<div class="MsoNormal" style="line-height: 150%;">
<span style="font-size: 10pt;">The Introduction section of the Etindele and Kuss (2018) preprint (left) and the Etindele et al. (2020) published
article (right). Sentences highlighted in yellow are common to both papers.</span></div>
<div class="MsoNormal" style="line-height: 150%;">
<o:p><br /></o:p></div>
<div class="separator" style="clear: both; text-align: center;">
<a href="http://1.bp.blogspot.com/-YKk-kss_p5o/Xp7VP2aBXVI/AAAAAAAAFE8/dOyWCWC3jqMt91CUkHQMpUrq_SGyOmgEACK4BGAYYCw/s1600/Preprint-article-discussion-side-by-side.png" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" height="640" src="https://1.bp.blogspot.com/-YKk-kss_p5o/Xp7VP2aBXVI/AAAAAAAAFE8/dOyWCWC3jqMt91CUkHQMpUrq_SGyOmgEACK4BGAYYCw/s640/Preprint-article-discussion-side-by-side.png" width="242" /></a></div>
<div class="MsoNormal" style="line-height: 150%;">
<o:p><br /></o:p></div>
<div class="MsoNormal" style="line-height: 150%;">
<o:p><br /></o:p></div>
<div class="MsoNormal" style="line-height: 150%;">
<!--[if gte vml 1]><v:shape id="_x0000_i1028"
type="#_x0000_t75" style='width:358.5pt;height:945.75pt'>
<v:imagedata src="file:///C:\Users\Nick\AppData\Local\Temp\msohtml1\01\clip_image007.png"
o:title="discussion-side-by-side"/>
</v:shape><![endif]--><!--[if !vml]--><!--[endif]--></div>
<div class="MsoNormal" style="line-height: 150%;">
<span style="font-size: 10.0pt; line-height: 150%;">The Discussion section of <span style="font-size: 13.3333px;">the Etindele and Kuss (2018) preprint (left) and the Etindele et al. (2020) article (right)</span>. Sentences highlighted in yellow are common to both papers.<o:p></o:p></span></div>
<div class="MsoNormal" style="line-height: 150%;">
<br /></div>
<div class="MsoNormal" style="line-height: 150%;">
<u>Another (apparently unrelated) preprint
contains the same insomnia results<o:p></o:p></u></div>
<div class="MsoNormal" style="line-height: 150%;">
It is also perhaps worth noting that
the summary of the participants’ results on the ISI measure—which, as we saw above,
was identical in every word and number between the preprint and Etindele Sosso
et al. (2020)’s published article—also appears, again identical in every word
and number, on pp. 5–6 of a 2019 preprint by the lead author, entitled
“Insomnia, excessive daytime sleepiness, anxiety, depression and socioeconomic
status among customer service employees in Canada”, which is available <a href="https://www.medrxiv.org/content/medrxiv/early/2019/08/12/19003194.full.pdf">here</a> [PDF].
This second preprint describes a study of yet another different sample, namely 1,200
Canadian customer service workers. If this is not just another remarkable coincidence,
it would suggest that the author may have discovered some fundamental invariant
property of humans with regard to insomnia. If so, one would hope that both
preprints could be peer reviewed most expeditiously, to bring this important
discovery to the wider attention of the scientific community.</div>
<div class="MsoNormal" style="line-height: 150%;">
<br /></div>
<div class="MsoNormal" style="line-height: 150%;">
<u>Other reporting issues from the
same laboratory<o:p></o:p></u></div>
<div class="MsoNormal" style="line-height: 150%;">
The lead author of the Etindele
Sosso et al. (2020) article has published even more studies with substantial
numbers of participants. Here are two such articles, which have 41 and 35
citations, respectively, according to Google Scholar:</div>
<div class="MsoNormal" style="line-height: 150%; margin-left: 36.0pt; text-indent: -36.0pt;">
<br /></div>
<div class="MsoNormal" style="line-height: 150%; margin-left: 36.0pt; text-indent: -36.0pt;">
<span style="mso-no-proof: yes;">Etindele Sosso, F. A., & Rauoafi, S. (2016).
Brain disorders: Correlation between cognitive impairment and complex
combination. <i style="mso-bidi-font-style: normal;">Mental Health in Family
Medicine</i>, <i style="mso-bidi-font-style: normal;">12</i>, 215–222. <a href="https://doi.org/10.25149/1756-8358.1202010">https://doi.org/10.25149/1756-8358.1202010</a></span></div>
<div class="MsoNormal" style="line-height: 150%; margin-left: 36.0pt; text-indent: -36.0pt;">
<span style="mso-no-proof: yes;">Etindele Sosso, F. A. (2017a).
Neurocognitive</span> game between risk factors, sleep and suicidal behaviour. <i style="mso-bidi-font-style: normal;">Sleep Science</i>, <i style="mso-bidi-font-style: normal;">10</i>(1), 41–46. <a href="https://doi.org/10.5935/1984-0063.20170007">https://doi.org/10.5935/1984-0063.20170007</a></div>
<div class="MsoNormal" style="line-height: 150%;">
<br /></div>
<div class="MsoNormal" style="line-height: 150%;">
In the 2016 article, 1,344
respondents were assessed for cognitive deficiencies; 71.7% of the participants
were aged 18–24, 76.2% were women, and 62% were undergraduates. (These figures
all match those that were reported in the lead author’s Master’s thesis, so we might
tentatively assume that this study used the same sample.) In the 2017 article,
1,545 respondents were asked about suicidal tendencies, with 78% being aged
18–24, 64.3% women, and 71% undergraduates. Although these are clearly entirely
different samples in every respect, the tables of results of the two studies
are remarkably similar. Every variable label is identical across all three
tables, which might not be problematic in itself if similar predictors were
used for all of the different outcome variables. More concerning, however, is
the fact that of the 120 cells in Tables 1 and 2 that contain statistics (mean/SD
combinations, <i style="mso-bidi-font-style: normal;">p</i> values other than
.000, regression coefficients, standard errors, and confidence intervals), 58—that
is, almost half—are identical in every digit. Furthermore, the <i style="mso-bidi-font-style: normal;">entirety</i> of Table 3—which shows the
results of the logistic regressions, ostensibly predicting completely different
outcomes in completely different samples—is identical across the two articles
(52 out of 52 numbers). One of the odds ratios in Table 3 has the value 1133096220.169
(again, in both articles). There does not appear to be an obvious explanation
for how this duplication could have arisen as the result of a natural process.</div>
<div class="MsoNormal" style="line-height: 150%;">
<br /></div>
<div class="separator" style="clear: both; text-align: center;">
<a href="http://4.bp.blogspot.com/-fEMjIZXr05w/Xp7WAgulYDI/AAAAAAAAFFg/YigOo07JdP4uvYPnYggNdsp4nxvT72XTwCK4BGAYYCw/s1600/Etindele-2016-vs-2017.png" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" height="640" src="https://4.bp.blogspot.com/-fEMjIZXr05w/Xp7WAgulYDI/AAAAAAAAFFg/YigOo07JdP4uvYPnYggNdsp4nxvT72XTwCK4BGAYYCw/s640/Etindele-2016-vs-2017.png" width="492" /></a></div>
<div class="MsoNormal" style="line-height: 150%;">
<span style="font-size: 10pt;">Left: The tables of results from Etindele Sosso and Raouafi
(2016). Right: The tables of results from Etindele Sosso (2017a). Cells highlighted
in yellow are identical (same variable name, identical numbers) in both articles.</span></div>
<div class="MsoNormal" style="line-height: 150%;">
<br /></div>
<div class="MsoNormal" style="line-height: 150%;">
<u>The mouse studies<o:p></o:p></u></div>
<div class="MsoNormal" style="line-height: 150%;">
Further evidence that this
laboratory may have, at the very least, a suboptimal approach to quality
control when it comes to the preparation of manuscripts comes from the
following pair of articles, in which the lead author of Etindele Sosso et al.
(2020) reported the results of some psychophysiological experiments conducted on
mice:</div>
<div class="MsoNormal" style="line-height: 150%;">
<br /></div>
<div class="MsoNormal" style="line-height: 150%; margin-left: 36.0pt; text-indent: -36.0pt;">
Etindele Sosso, F. A. (2017b). Visual dot interaction with short-term
memory. <i style="mso-bidi-font-style: normal;">Neurodegenerative Disease Management</i>,
<i style="mso-bidi-font-style: normal;">7</i>(3), 182–190. <a href="https://doi.org/10.2217/nmt-2017-0012">https://doi.org/10.2217/nmt-2017-0012</a></div>
<div class="MsoNormal" style="line-height: 150%; margin-left: 36.0pt; text-indent: -36.0pt;">
Etindele Sosso, F. A., Hito, M. G., & Bern, S. S. (2017). Basic
activity of neurons in the dark during somnolence induced by anesthesia. <i style="mso-bidi-font-style: normal;">Journal of Neurology and Neuroscience</i>, <i style="mso-bidi-font-style: normal;">8</i>(4), 203–207. <a href="https://doi.org/10.21767/2171-6625.1000203">https://doi.org/10.21767/2171-6625.1000203</a> <a href="file:///C:/Users/Nick/Documents/Academic/Reanalysis/Etindele%20Sosso-Gaming/20200421%20Blog%20post%20-%20Some%20issues%20in%20a%20recent%20gaming%20research%20article.doc#_ftn1" name="_ftnref1" style="mso-footnote-id: ftn1;" title=""><span class="MsoFootnoteReference"><span style="mso-special-character: footnote;"><!--[if !supportFootnotes]--><span class="MsoFootnoteReference"><span style="font-family: "times new roman"; font-size: 12.0pt;">[1]</span></span><!--[endif]--></span></span></a></div>
<div class="MsoNormal" style="line-height: 150%;">
<br /></div>
<div class="MsoNormal" style="line-height: 150%;">
In each of these two articles
(which have 28 and 24 Google Scholar citations, respectively), the neuronal
activity of mice when exposed to visual stimuli under various conditions was
examined. Figure 5 of the first article shows the difference between the firing
rates of the neurons of a sample of an unknown number of mice (which could be as
low as 1; I was unable to determine the sample size with any great level of certainty
by reading the text) in response to visual stimuli that were shown in different
orientations. In contrast, Figure 3 of the second article represents the firing
rates of two different types of brain cell (interneurons and pyramidal cells)
before and after a stimulus was applied. That is, these two figures represent
completely different variables in completely different experimental conditions.
And yet, give or take the use of dots of different shapes and colours, they
appear to be exactly identical. Again, it is not clear how this could have
happened by chance.</div>
<div class="MsoNormal" style="line-height: 150%;">
<br /></div>
<div class="MsoNormal" style="line-height: 150%;">
<!--[if gte vml 1]><v:shape id="_x0000_i1030"
type="#_x0000_t75" style='width:451.5pt;height:404.25pt'>
<v:imagedata src="file:///C:\Users\Nick\AppData\Local\Temp\msohtml1\01\clip_image011.png"
o:title=""/>
</v:shape><![endif]--><!--[if !vml]--><!--[endif]--></div>
<div class="separator" style="clear: both; text-align: center;">
<a href="http://4.bp.blogspot.com/-lDvjiLGXY3A/Xp7U3EcrVdI/AAAAAAAAFEk/VoHtFq_pukcUmGiFCf3xEs669fQQxq--gCK4BGAYYCw/s1600/Etindele-mouse-neurons.png" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" height="573" src="https://4.bp.blogspot.com/-lDvjiLGXY3A/Xp7U3EcrVdI/AAAAAAAAFEk/VoHtFq_pukcUmGiFCf3xEs669fQQxq--gCK4BGAYYCw/s640/Etindele-mouse-neurons.png" width="640" /></a></div>
<div class="MsoNormal" style="line-height: 150%;">
<span style="font-size: 10pt;">Top: Figure 5 from Etindele Sosso (2017b). Bottom: Figure 3
from Etindele Sosso et al. (2017). The dot positions and axis labels appear to
be identical. Thanks are due to Elisabeth Bik for providing a second pair of
eyes.</span></div>
<div class="MsoNormal" style="line-height: 150%;">
<br /></div>
<div class="MsoNormal" style="line-height: 150%;">
<u>Conclusion<o:p></o:p></u></div>
<div class="MsoNormal" style="line-height: 150%;">
I find it slightly surprising that
16 authors—all of whom, we must assume because of their formal statements to
this effect in the “Author contributions” section, made substantial
contributions to the Etindele et al. (2020) article in order to comply with the
demanding authorship guidelines of <i style="mso-bidi-font-style: normal;">Nature
Research</i> journals (specified <a href="https://www.nature.com/nature-research/editorial-policies/authorship">here</a>)—apparently
failed to notice that this work contained quite so many inconsistencies. It
would also be interesting to know what the reviewers and action editor had to
say about the manuscript prior to its publication. The time between submission
and acceptance was 85 days (including the end of year holiday period), which does not suggest that a particularly extensive revision process took
place. In any case, it seems that some sort of corrective action may be
required for this article, in view of the importance of the subject matter for
public policy.</div>
<div class="MsoNormal" style="line-height: 150%;">
<br /></div>
<div class="MsoNormal" style="line-height: 150%;">
<u>Supporting files<o:p></o:p></u></div>
<div class="MsoNormal" style="line-height: 150%;">
I have made the following
supporting files available <a href="http://nickbrown.fr/blog/etindele-et-al">here</a>: </div>
<div class="MsoNormal" style="line-height: 150%; margin-left: 36.0pt; mso-list: l0 level1 lfo1; tab-stops: list 36.0pt; text-indent: -18.0pt;">
<!--[if !supportLists]--><span style="mso-list: Ignore;">-<span style="font: 7.0pt "Times New Roman";">
</span></span><!--[endif]--><span style="font-family: "courier new"; font-size: 11.0pt; line-height: 150%;">Etindele-et-al-Table3-numbers.xls</span>: An Excel
file containing the numbers from Table 3 of Etindele et al.’s (2020) article,
with some calculations that illustrate the deficiencies in the relations
between the statistics that I mentioned earlier. The basic numbers were
extracted by performing a copy/paste from the article’s PDF file and using text
editor macro commands to clean up the structure.</div>
<div class="MsoNormal" style="line-height: 150%; margin-left: 36.0pt; mso-list: l0 level1 lfo1; tab-stops: list 36.0pt; text-indent: -18.0pt;">
<!--[if !supportLists]--><span style="mso-list: Ignore;">-<span style="font: 7.0pt "Times New Roman";">
</span></span><!--[endif]-->“<span style="font-family: "courier new"; font-size: 11.0pt; line-height: 150%;">(Annotated) Etindele Sosso, Raouafi - 2016 - Brain
Disorders - Correlation between Cognitive Impairment and Complex
Combination.pdf</span>” and “<span style="font-family: "courier new"; font-size: 11.0pt; line-height: 150%;">(Annotated) Etindele Sosso - 2017 - Neurocognitive
Game between Risk Factors, Sleep and Suicidal Behaviour.pdf</span>”: Annotated
versions of the 2016 and 2017 articles mentioned earlier, with identical
results in the tables highlighted.</div>
<div class="MsoNormal" style="line-height: 150%; margin-left: 36.0pt; mso-list: l0 level1 lfo1; tab-stops: list 36.0pt; text-indent: -18.0pt;">
<!--[if !supportLists]--><span style="mso-list: Ignore;">-<span style="font: 7.0pt "Times New Roman";">
</span></span><!--[endif]-->“<span style="font-family: "courier new"; font-size: 11.0pt; line-height: 150%;">(Annotated) Etindele Sosso, Kuss - 2018 (preprint) -
Insomnia and problematic gaming - A study in 9 low- and middle-income
countries.pdf</span>” and <span style="font-family: "courier new"; font-size: 11.0pt; line-height: 150%;">“(Annotated) Etindele Sosso et al. - 2020 -
Insomnia, sleepiness, anxiety and depression among different types of gamers in
African countries.pdf</span>” Annotated versions of the 2018 preprint and the
published Etindele et al. (2020) article, with overlapping text highlighted.</div>
<div class="MsoNormal" style="line-height: 150%; margin-left: 36.0pt; mso-list: l0 level1 lfo1; tab-stops: list 36.0pt; text-indent: -18.0pt;">
<!--[if !supportLists]--><span style="mso-list: Ignore;">-<span style="font: 7.0pt "Times New Roman";">
</span></span><!--[endif]--><span style="font-family: "courier new"; font-size: 11.0pt; line-height: 150%;">Etindele-2016-vs-2017.png</span>, Etindele-et-al-Table3-duplicates.png,
<span style="font-family: "courier new"; font-size: 11.0pt; line-height: 150%;">Etindele-mouse-neurons.png</span>,
<span style="font-family: "courier new"; font-size: 11.0pt; line-height: 150%;">Etindele
Sosso-Kuss-Preprint-Figure1.png</span>, <span style="font-family: "courier new"; font-size: 11.0pt; line-height: 150%;">Preprint-article-discussion-side-by-side.png</span>,
<span style="font-family: "courier new"; font-size: 11.0pt; line-height: 150%;">Preprint-article-intro-side-by-side.png</span>:
Full-sized versions of the images from this blog post.</div>
<div class="MsoNormal" style="line-height: 150%;">
<br /></div>
<div class="MsoNormal" style="line-height: 150%;">
<u>Reference<o:p></o:p></u></div>
<div class="MsoNormal" style="line-height: 150%; margin-bottom: 6.0pt; margin-left: 36.0pt; margin-right: 0cm; margin-top: 0cm; mso-pagination: widow-orphan lines-together; text-indent: -36.0pt;">
Brown,
N. J. L., & Heathers, J. A. J. (2019). Rounded Input Variables, Exact Test Statistics
(RIVETS): A technique for detecting hand-calculated results in published
research.<span style="color: black; mso-bidi-font-style: italic;"> <i>PsyArXiv
Preprints</i></span><span style="color: black;">. </span><a href="https://doi.org/10.31234/osf.io/ctu9z">https://doi.org/10.31234/osf.io/ctu9z</a><span lang="EN-US" style="mso-ansi-language: EN-US;"><o:p></o:p></span></div>
<div class="MsoNormal" style="line-height: 150%;">
<br />
<div class="MsoNormal" style="line-height: 24px;">
[[ Update 2020-04-21 13:14 UTC: Via Twitter, I have learned that I am not the first person to have publicly questioned the Etindele et al. (2020) article. See Platinum Paragon's blog post from 2020-04-17 <a href="http://platinumparagon.info/gaming-addiction-in-africa/">here</a>. ]]</div>
<div class="MsoNormal" style="line-height: 24px;">
<br />
[[ Update 2020-04-22 13:43 UTC: Elisabeth Bik has identified two more articles by the same lead author that share an image (same chart, different meaning). See <a href="https://twitter.com/MicrobiomDigest/status/1252656666698366976?s=20" target="_blank">this Twitter thread</a>. ]]<br />
<br />
<div class="MsoNormal" style="line-height: 24px;">
[[ Update 2020-04-23 22:48 UTC: See my related blog post <a href="https://steamtraen.blogspot.com/2020/04/the-mystery-of-missing-authors.html">here</a>, including discussion of a partial data set that appears to correspond to the Etindele et al. (2020) article. ]]<br />
<br />
<div class="MsoNormal" style="line-height: 24px;">
[[ Update 2020-06-04 11:50 UTC: I blogged about the reaction (or otherwise) of university research integrity departments to my complaint about the authors of the Etindele Sosso et al. article <a href="https://steamtraen.blogspot.com/2020/05/the-silence-of-rios.html">here</a>. ]]<br />
<br />
[[ Update 2020-06-04 11:55 UTC: The Etindele Sosso et al. article has been retracted. The retraction notice can be found <a href="https://www.nature.com/articles/s41598-020-66798-w">here</a>. ]]</div>
</div>
<div class="MsoNormal" style="line-height: 24px;">
<br /></div>
</div>
</div>
<div style="mso-element: footnote-list;">
<hr align="left" size="1" width="33%" />
<!--[endif]-->
<br />
<div id="ftn1" style="mso-element: footnote;">
<div class="MsoFootnoteText" style="line-height: 150%;">
<a href="file:///C:/Users/Nick/Documents/Academic/Reanalysis/Etindele%20Sosso-Gaming/20200421%20Blog%20post%20-%20Some%20issues%20in%20a%20recent%20gaming%20research%20article.doc#_ftnref1" name="_ftn1" style="mso-footnote-id: ftn1;" title=""><span class="MsoFootnoteReference"><span style="font-size: 12.0pt; line-height: 150%;"><span style="mso-special-character: footnote;"><!--[if !supportFootnotes]--><span class="MsoFootnoteReference"><span style="font-family: "times new roman"; font-size: 12.0pt;">[1]</span></span><!--[endif]--></span></span></span></a><span style="font-size: 12.0pt; line-height: 150%;"> This
article was accepted 12 days after submission, which is presumably entirely
unrelated to the fact that the lead author is listed <a href="https://www.jneuro.com/editors.php">here</a> as the journal’s
specialist editor for Neuropsychology and Cognition.</span></div>
<div class="separator" style="clear: both; text-align: center;">
<a href="http://1.bp.blogspot.com/-P3dxxv2uksg/Xp7UtFV4UxI/AAAAAAAAFEc/1c1SFCVHZPMWdRSUiPoe1YNNgxGXwubywCK4BGAYYCw/s1600/JNeuro-editors.png" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" height="85" src="https://1.bp.blogspot.com/-P3dxxv2uksg/Xp7UtFV4UxI/AAAAAAAAFEc/1c1SFCVHZPMWdRSUiPoe1YNNgxGXwubywCK4BGAYYCw/s640/JNeuro-editors.png" width="640" /></a></div>
<div class="MsoFootnoteText" style="line-height: 150%;">
<br /></div>
</div>
</div>
<br /></div>
Nick Brownhttp://www.blogger.com/profile/07481236547943428014noreply@blogger.com15tag:blogger.com,1999:blog-7890764972166411105.post-6331305593648952582020-04-19T20:44:00.010+02:002023-05-23T01:54:26.619+02:00In psychology everything mediates everything<br />
<div class="MsoNormal" style="line-height: 150%;">
In the past couple of years I have
reviewed half a dozen manuscripts with abstracts that go something like this:</div>
<div class="MsoNormal" style="line-height: 150%;">
<br /></div>
<div class="MsoNormal" style="line-height: 150%; margin-left: 36pt;">
<Construct
X> is known to be associated with higher levels of well-being and healthy
psychological functioning, as indexed by <Construct Y>. However, to date,
no study has investigated the role of <Construct M> in this association.
The present study bridges this gap by testing a mediation path model in a
sample of undergraduates (<i style="mso-bidi-font-style: normal;">N</i> = 100).
As predicted, M fully mediated the positive association between X and Y.<span style="mso-spacerun: yes;"> </span>These results suggest that X predicts higher levels
of M, which subsequently predicts higher levels of Y. These results provide new
insight that may advance a coherent theoretical framework on the pathways by
which M enhances psychological well-being.</div>
<div class="MsoNormal" style="line-height: 150%;">
<br /></div>
<div class="MsoNormal" style="line-height: 150%;">
There is typically a description of
how the 100 participants completed measures of constructs X, M, and Y, with a
table of correlations that might look like this:</div>
<div class="MsoNormal" style="line-height: 150%;">
<br /></div>
<div class="MsoNormal" style="line-height: 150%;">
<span style="font-family: "courier new";"><span style="mso-spacerun: yes;"> </span>X<span style="mso-spacerun: yes;">
</span>Y<o:p></o:p></span></div>
<div class="MsoNormal" style="line-height: 150%;">
<span style="font-family: "courier new";">Y .24*<o:p></o:p></span></div>
<div class="MsoNormal" style="line-height: 150%;">
<span style="font-family: "courier new";">M .52***<span style="mso-spacerun: yes;"> </span>.32**<o:p></o:p></span></div>
<div class="MsoNormal" style="line-height: 150%;">
<br /></div>
<div class="MsoNormal" style="line-height: 150%;">
* <i style="mso-bidi-font-style: normal;">p</i> < .05; ** <i style="mso-bidi-font-style: normal;">p</i> < .01;
*** <i style="mso-bidi-font-style: normal;">p</i> < .001.</div>
<div class="MsoNormal" style="line-height: 150%;">
<br /></div>
<div class="MsoNormal" style="line-height: 150%;">
Then we get to the mediation
analysis. More often than not this is done using the PROCESS macro in SPSS, but
it can also be done “by hand” using a few ordinary least-squares regressions. Here
are the steps required (cf. Baron & Kenny, 1986):</div>
<div class="MsoNormal" style="line-height: 150%;">
<br /></div>
<ol start="1" style="margin-top: 0cm;" type="1">
<li class="MsoNormal" style="line-height: 150%; mso-list: l0 level1 lfo1; tab-stops: list 36.0pt;">Show that X is a significant predictor of Y. You probably don’t actually need to do the regression for this, as the standardized regression coefficient and its associated <i style="mso-bidi-font-style: normal;">p</i>
value will be identical to the correlation coefficient between X and Y, but sometimes the manuscript
will show the SPSS output to prove that the authors conducted this regression anyway. (In the last manuscript
that I reviewed, the authors performed the single-predictor regression and managed to obtain a standardized regression coefficient
that was <i>different</i> to the zero-order correlation, which did not enhance my
confidence in the rest of their analyses.) Here the <i style="mso-bidi-font-style: normal;">p</i> value will be .016.</li>
<li class="MsoNormal" style="line-height: 150%; mso-list: l0 level1 lfo1; tab-stops: list 36.0pt;">Show that X is a significant predictor of M. Again,
no regression is required for this, as it’s just the correlation coefficient. The <i style="mso-bidi-font-style: normal;">p</i> value in this example is about 3E−8.</li>
<li class="MsoNormal" style="line-height: 150%; mso-list: l0 level1 lfo1; tab-stops: list 36.0pt;">Regress Y on both X and M. If you get a significant
regression coefficient for M then you have at least a “partial” mediation
effect. If, in addition, the regression coefficient for X is
non-significant then you have “full" mediation. Here, this produces the
following standardized coefficients:</li>
<ul style="margin-top: 0cm;" type="disc">
<li class="MsoNormal" style="line-height: 150%; mso-list: l0 level2 lfo1; tab-stops: list 72.0pt;">M: β = 0.268, <i style="mso-bidi-font-style: normal;">p</i> = .018</li>
<li class="MsoNormal" style="line-height: 150%; mso-list: l0 level2 lfo1; tab-stops: list 72.0pt;">X: β = 0.101, <i style="mso-bidi-font-style: normal;">p</i> = .368</li>
</ul>
</ol>
<div class="MsoNormal" style="line-height: 150%;">
Ta-da!
In this example, we have complete mediation: The <i>p</i> value for the mediator, M, is significant and the <i>p</i> value for X isn’t. We conclude that Construct
M fully mediates the relation between Construct X and Construct Y. We write
it up and celebrate our fine contribution to understanding
the mechanisms that lead to well-being. Surely the end of mental distress is only one more grant away.</div>
<div class="MsoNormal" style="line-height: 150%;">
<br /></div>
<div class="MsoNormal" style="line-height: 150%;">
The problem is this: Absolutely any
other variable that you might put in place of M, and which is correlated in the
same way with X and Y, <i style="mso-bidi-font-style: normal;">will also show exactly
the same mediation effect</i>. And there is no shortage of things you can
measure—in psychology, at least—that are correlated at around .5 and .3 with two other variables, themselves intercorrelated at around .2, that
you might have measured. Let’s say that X is some aspect of socioeconomic status and Y is
subjective well-being. You can easily come up with any number of ideas for M:
gratitude, optimism, self-esteem, all of the Big Five personality traits (if
you reverse-score neuroticism as emotional stability), etc., without even
needing to resort to Lykken and <span style="mso-no-proof: yes;">Meehl’s</span>
“crud factor” (“in psychology and sociology everything correlates with everything”; Meehl, 1990, p. 204). Does it make sense for multiple third variables all to apparently fully mediate the relation between a predictor and an outcome variable?</div>
<div class="MsoNormal" style="line-height: 150%;">
<br /></div>
<div class="MsoNormal" style="line-height: 150%;">
I wrote some R code, which you can find <a href="http://nickbrown.fr/blog/mediation/mediation.R">here</a>, to demonstrate
the example that I gave above.
You will see that I performed the calculations in two ways. The first was to
generate (with a bit of trial and error) some random data with the correct
correlations. (This produces a bit of rounding error, so the <i style="mso-bidi-font-style: normal;">p</i> value for the beta for M in the
regression is reported at .019, not .018.) The second—my preferred method,
since you can generally use this starting with the table of descriptives that
appears in an article—is to start with the correlations and perform the
regression calculations from there. (A surprising number of people do not seem
to know that you can generally determine the standardized coefficients of multiple
regression models just from the correlation table. The standard errors—and, hence,
the <i style="mso-bidi-font-style: normal;">p</i> values—can then be derived from
the sample size. If you have the standard deviations as well, you can get the <span style="mso-no-proof: yes;">unstandardized</span> coefficients. Add in the means and
you can calculate the intercepts too. Again, this can all be done from the descriptive statistics, which is probably why the complete table of descriptives and correlations used to be standard in every paper. You don't need the raw data for any of this.)</div>
<div class="MsoNormal" style="line-height: 150%;">
<br /></div>
<div class="MsoNormal" style="line-height: 150%;">
If your initial choice for variable
X is more strongly correlated with Y than M is, then you can very often just swap
M and Y around, because there is typically nothing to say that whatever X is
measuring occurs “before” whatever M is measuring, or vice versa—especially if you just hauled a bunch of undergraduates in and gave them measures of their current levels of X, M, and Y to complete. The reason
why you want your mediator, M, to be more strongly correlated than X with the
outcome (Y), is a little-known phenomenon of two-variable regression that I
like to think of as a sort of “Matthew effect”. Feel free to skip the next
paragraph in which I explain this in tedious detail.</div>
<div class="MsoNormal" style="line-height: 150%;">
<br /></div>
<div class="MsoNormal" style="line-height: 150%;">
When the two predictors are
moderately strongly correlated with each other (.52, in our case), then
although their zero-order correlations with the outcome variable might be quite
close together (.32 and .24 here), their standardized regression coefficients will diverge
by quite a bit more than their respective correlation
coefficients. Here, M’s correlation of .32 led to a beta of 0.268, which is a
16% reduction, but X’s correlation of .24 was reduced by 58% to a beta of
0.101. If the correlation between M and X had been a little higher (eg, .60 instead of .52), the beta
for M would actually have been larger (0.275) and the beta for X would have been even smaller (0.075). At
some point along the M–X correlation continuum (around .75), the beta for M
would be exactly equal to the correlation coefficient of M with Y (as if X wasn't in the regression model at all), and the beta
for X would be zero. Continuing even further, we would hit “negative
suppression” territory, with M’s standardized regression coefficient being <i style="mso-bidi-font-style: normal;">greater</i> than the original correlation
coefficient of .32, and X’s standardized regression coefficient being negative. Many people seem to have a rather naïve
view of multiple regression in which the addition of a new predictor results in the betas
for all of the predictors being reduced in some roughly equal proportion, but the
reality is often nothing like that. You can explore what happens with just two
predictors (with more, things get even wilder) <a href="http://shiny.ieis.tue.nl/suppressiongraphics/">here</a> using my Shiny app.</div>
<div class="MsoNormal" style="line-height: 150%;">
<br /></div>
<div class="MsoNormal" style="line-height: 150%;">
So it’s possible to build an almost
infinite number of mediation studies, all of which will appear to tell us
something about the mediation of the relation between two psychological variables by a third, although almost all of them are just
illustrating a known phenomenon of multiple regression. Again, <i>everything is
determined by the three correlations between the variables</i>, plus the sample
size if you care about statistical significance. (Alert readers will have
noticed that whether or not mediation is “full” or “partial” will depend to a
large extent on the sample size; with enough participants even the residual
effect of X on Y will be large enough that its p value doesn’t drop below .05.
But of course, alert readers will also know that these days statistical
significance doesn’t mean very much on its own, right?)</div>
<div class="MsoNormal" style="line-height: 150%;">
<br /></div>
<div class="MsoNormal" style="line-height: 150%;">
Now, am I saying that all of the
mediation articles that I get to review are based on an atheoretical “throw
some numbers at the wall and see what sticks” approach, which might be an implication of what I have argued here? Well, no... but I’m also not saying that that never happens. I have heard first-hand from grad students, in several cases, what
happens when they have a bunch of variables and no obvious result: Their
supervisor suggests that they write them up as a mediation analysis.</div>
<div class="MsoNormal" style="line-height: 150%;">
<br /></div>
<div class="MsoNormal" style="line-height: 150%;">
I don’t think that preregistration
will necessarily help all that much here, because it is quite predictable from
previous knowledge that X, M, and Y will have the pattern of correlations
needed to produce an apparent mediation effect. I’m going to suggest that the
only solution is to refrain from doing this kind of mediation analysis altogether
in the absence of (a) much better theoretical justification than we currently see, and (b) some kind of constraint on the temporal order in which changes in X, M, and Y occur.
Without a demonstration that the causal arrows are running from X to M and M to Y (MacKinnon & Pirlott, 2014), and not vice versa, we have no
way of knowing whether we are dealing with mediation or confounding,
especially since in many cases the constructs X, M, and Y may themselves be
caused by multiple other factors, and so ad infinitum (cf. Arah, 2008). In the absence of experimental manipulation, causality is hard to demonstrate, especially in psychology. </div>
<div class="MsoNormal" style="line-height: 150%;">
<br />
<br /></div>
<div class="MsoNormal" style="line-height: 150%;">
<u>References</u></div>
<div class="MsoNormal" style="line-height: 150%; margin-left: 36pt; text-indent: -36pt;">
Arah, O. A. (2008). The role of causal reasoning in understanding Simpson's
paradox, Lord's paradox, and the suppression effect: Covariate selection in the
analysis of observational studies. <i style="mso-bidi-font-style: normal;">Emerging
Themes in Epidemiology</i>, <i style="mso-bidi-font-style: normal;">5</i>, 5.
https://doi.org/10.1186/1742-7622-5-5</div>
<div class="MsoNormal" style="line-height: 150%; margin-left: 36pt; text-indent: -36pt;">
<span lang="DE" style="text-indent: -36pt;">Baron, R. M., & Kenny, D.
A. (1986). </span><span style="text-indent: -36pt;">The moderator–mediator variable distinction in social
psychological research: Conceptual, strategic, and statistical considerations. </span><i style="text-indent: -36pt;">Journal of Personality and Social Psychology</i><span style="text-indent: -36pt;">,
</span><i style="text-indent: -36pt;">51</i><span style="text-indent: -36pt;">(6), 1173–1182. https://doi.org/10.1037/0022-3514.51.6.1173</span></div>
<div class="MsoNormal" style="line-height: 150%; margin-left: 36pt; text-indent: -36pt;">
<span style="text-indent: -36pt;">MacKinnon, D. P., & Pirlott, A. G. (2014). </span><span style="text-indent: -36pt;">Statistical approaches for enhancing c</span><span style="text-indent: -36pt;">ausal interpretation of the M to Y r</span><span style="text-indent: -36pt;">elation in mediation analysis. </span><span style="text-indent: -36pt;"><i>Personality and Social Psychology Review</i>, </span><span style="text-indent: -36pt;"><i>19</i>(1), 30</span><span style="text-indent: -36pt;">–43. </span><span style="text-indent: -36pt;">https://doi.org/</span><span style="text-indent: -36pt;">10.1177/1088868314542878</span></div><div class="MsoNormal" style="line-height: 150%; margin-left: 36pt; text-indent: -36pt;"><span style="text-indent: -36pt;">Meehl, P. E. (1990). Why summaries of research on psychological theories are often uninterpretable. </span><i style="text-indent: -36pt;">Psychological Reports</i><span style="text-indent: -36pt;">, </span><i style="text-indent: -36pt;">66</i><span style="text-indent: -36pt;">(1), 195–244. https://doi.org/10.2466/pr0.1990.66.1.195</span><br /></div>
<br />
<div class="MsoNormal" style="line-height: 24px;">
(Thanks to Julia Rohrer for her helpful comments on an earlier draft of this post. If the whole thing is garbage, it's probably because I didn't incorporate more of her thoughts.)</div>
<div class="MsoNormal" style="line-height: 24px;">
<br />
<br /></div>
Nick Brownhttp://www.blogger.com/profile/07481236547943428014noreply@blogger.com2