A reviewer asks for sample size and power in my retrospective study. Can I ignore it?

I-Chen Tsai, MD, PhDTranslated from InnovaRad's public Q&A (Taiwan)Original published Nov 24, 2022Read the original

The question. I wrote a retrospective observational study with multivariable adjustment. A reviewer asked us to add a sample size calculation and a power calculation. Do retrospective studies really report these? We analyzed many subgroups, and giving sample size and power for each would produce about 14 sets. Can we ignore the request?

You cannot ignore it. If you ignore a reviewer, the reviewer gets angry, and then rejects you.

I agree with your point. In a retrospective study you have as many patients as you have, and nobody calculated a sample size beforehand. Post-hoc power analysis is also controversial. But this is what the reviewer wants, and if we can give it, the reviewer will usually accept the answer.

My advice is to do it. A power calculation is not hard; an online tool such as the ClinCalc post-hoc power calculator will give you the numbers quickly.

You do not need to change the manuscript. List the results in a table and attach it to the reply letter. In the reply, politely note that post-hoc power has limited interpretive value and that the study reports confidence intervals to show the precision of its estimates. Keep the tone polite, and then wait for the acceptance.

這些回答是經驗與建議，不是規定。你的稿件以目標期刊的規定為準，你的研究以所屬機構的倫理審查為準。