2

I'm in the field of Psychology. One of my manuscripts is now under review after the 1st round of revision. However, I found one of the variables was calculated in a wrong way, the results are better with the corrected variable because the significance levels were higher (i.e., p values were lower), but it didn't change the conclusions of my study, all the significant results still remain significant and non-significant results are still non-significant. I would like to know what would be the best option to deal with it.

option 1: ignore it.

option 2: correct it when I receive reviewers' feedbacks WITHOUT letting the reviewers and editor know.

option 3: correct it when I receive reviewers' feedbacks AND let the reviewers and editor know.

I'm worried that letting the editor and reviewers know would leave a bad impression on me and the quality of my work.

Thank you in advance!

superarrow
  • 87
  • 7
  • Definitely option 3. And do not fear leaving a bad impression - in my mind, it's only a good sign that you spotted it yourself and acted honest. Plus, I've reviewed very bad papers of very good scientists - what kind of an impression that makes, huh? So, you're safe. –  Aug 09 '18 at 21:36

2 Answers2

4

I work in biomedical field so am not sure how vindictive psychology specialists are. The few I worked with are pretty nice, but n=3.

My opinion is definitely option 3. It's not just the p-value that matters, but also i) the statistics itself (e.g. the regression coefficient) could be different and ii) replicability could be threatened as well. If someone would like to follow your work they may be perplexed by your errors.

As for leaving a bad impression. I'd rather get rejected because I was being diligent than getting an acceptance with a flaw that I once convinced myself was trivial, but occasionally came up in the back of my mind. Just bad flavor.

I'd say no need to drum up what was corrected. Simply point out that in a final code review you spotted an error, and corrections to xxx variable was made. Overall associations/differences and conclusion remain unchanged. Give a short apology. And hope for the best.

Penguin_Knight
  • 20,930
  • 2
  • 49
  • 92
  • You certainly want the correction in the published version. You will also probably have to do some revisions in any case. – Buffy Aug 09 '18 at 19:59
  • Thank you Penguin_Knight for your suggestion. I kind of know it deep in my heart that I should go with option 3. However, because I fear that I may be perceived negatively by my supervisor and reviewers, I tend to persuade myself and justify that it's okay to ignore it or change with our letting others know. But I know I won't have a peace of mind if I do so. – superarrow Aug 09 '18 at 20:16
  • I've heard stories in the past about how researchers, especially some established professors, don't actually care about how data was coded and results were analyzed by research assistants or graduate students; they probably won't bother to check their data as long as the results are consistent with the literature and their beliefs. I guess this is not a good practice... – superarrow Aug 09 '18 at 20:16
1

This is potentially problematic, but probably not for the reasons you expect.

By all means, ensure that you raise this when you make your second submission. Provide a detailed description of the problem and your modifications. Set this apart from the revisions requested in response to reviewer comments, though. In addition, inform the deputy editor in charge of your manuscript that this has happened.

The reason I state that this is problematic is because some journals have been taken to task by late additions to the text that were inserted during the later stages of review when only minor modifications are expected. Let me give you an example of one such incident. A manuscript describes the effectiveness of a new way to mitigate the harmful effects of soil erosion in the coastal areas of Malaysia. It's reported primary and secondary outcomes, while showing a positive effect, was lacking in practical significance. The manuscript's reviews were generally positive. During the third round, the authors inserted (without any prompting from reviewers or editorial staff) a new outcome that was substantially positive. This insertion consisted of a single sentence in a 3,500-word third revision. No one spotted it and it proceeded to publication.

When it was published, the authors issues press releases highlighting the effectiveness of the program but using the newly inserted outcome as the headline. "New technique reduces cuts soil erosion by 90%!" The journal received complaints from the reviewers stating that internal processes were breached, that the outcome did not undergo proper peer review, etc etc. Internal investigations were conducted. External arbiters were engaged. The main author was traumatised and had to go on leave because of the flak she was getting.

This incident was described to us by staff of another journal in my field during our regular get-togethers a few years ago now. It's caused us to be very wary about late insertions.

In the final analysis, you need to protect yourself from these allegations by being forthcoming about the revisions you plan on making.

Good luck.

  • Hi, St. Inkbug, thank you very much for your insightful input. I really appreciate it. I actually never thought it would be lead to such a negative consequence by just changing something mirror in the manuscript. In my case, my paper is not accepted at all, the editor has clarified in his first feedback that he cannot accept this paper as major revision is required; and he also cannot guarantee acceptance even after we revise our manuscripts. – superarrow Aug 10 '18 at 01:14
  • I think whether it would be a good idea to send an email to the editor now, and let him know I've spotted errors and would like to correct them. If reviewers also found the errors in their 2nd round of review and pointed it out, would it make me look untrustworthy to state after reviewers' feedbacks that we indeed made mistakes? – superarrow Aug 10 '18 at 01:14
  • Thanks. I think that a major revision provides you with quite significant cover to make the changes you want to make. Regardless of the reviewers picking up the error or not, your first priority should be to produce a manuscript that is scientifically sound. Else, you'll be wearing this error for a while. –  Aug 10 '18 at 01:17
  • Okay, Thank you again! I'm quite clear now about what to do. To be honest is the best policy :) – superarrow Aug 10 '18 at 01:37