13

I told these people there was not much to report about their data because they had a very small number of independent samples with a lot of repeated measures. Their measures were all categorical too. So now they chose to just ignore the fact that the study contains repeated measures and treated every sample as independent.

What can I do?

6
  • 4
    I sympathise, but this seems like more of an academic ethics question than a statistics question, and so is probably off-topic. You could consider contacting the journal or your institute ombudsman (if one exists) to raise these issues with them.
    – mkt
    Commented Oct 10, 2023 at 11:04
  • 2
    I have flagged this question and asked the moderators to migrate it to Academia.SE. That said, reputable journals require all authors to sign consent forms before they appear on a paper. Did you sign such a form? Commented Oct 10, 2023 at 11:16
  • 2
    This question seems good for Academia.SE unless the question was supposed to be a discussion/question about the statistics? This, and actually the entire question, is not clear to me. What is going on from the statistical point of view: Their treated dependent samples as independent (overestimating significance) or they didn't use a repeated measures design (underestimating significance)? What's the problem with categorical? And from the academic point of view: What has been your role in the research? How is your name connected to the publication? As an author, reference, acknowledgement? Commented Oct 10, 2023 at 11:42
  • 6
    How is your name "included"? As an author? What?
    – Buffy
    Commented Oct 10, 2023 at 12:18
  • 4
    @RichardErickson I've experienced all of those in just one field! :)
    – Bryan Krause
    Commented Oct 10, 2023 at 15:22

2 Answers 2

24

If you don't want your name associated with a paper, no one can force you to do so. Authorship for publications is by consent of each author. Even mentioning someone as a contributor to a paper requires consent, though it is often implied.

You can inform the others to remove all mention of yourself, possibly leaving them with plagiarism issues in some cases if they still want to publish. If they submit a paper without your permission, you can contact the publisher who will respond if the publication is legitimate.

Of course, if the others are powerful and you are not this leaves you with the possibility of blowback from them. In particular, fights between advisors and students are seldom fruitful for the student.

1
  • 6
    I once used something like "I don't feel my involvement with this paper meets the requirements for authorship" in a similar situation to OP's.
    – G_B
    Commented Oct 11, 2023 at 11:44
11

An interesting part about this question is the history of your other questions which I believe are related

These questions all revolve around a small data set.

The value of small data sets is not so big because it has a weak statistical strength and is for that reason not worthwhile to publish.

However, reasons to publish the data could still exist, and this is when the report is not so much about the significance of the data, but also about the methods or about rough estimates. For example when we have a billion dollar exploration mission to send five probes to Mars to collect samples in 5 places, would we discard it and not publish it because it is not significant?

Also, is a small sample less significant than a large sample when the p-values are the same? Is the 'lady tasting tea' experiment with zero errors among eight tea cups (p-value 1/70) worse than when it would have been with 448 errors out of 966 tea cups (p-value of 1/69.86)? Another aspect about small samples is that you do not always need to estimate the error based on the observed sample variance (and require a large sample to reduce the error or increase the significance). In specific cases you can make good guesses about a distribution value without a large sample.

The exact value of the data and whether there's indeed not much to report is not so clear from your description of the data.

You must log in to answer this question.