1
$\begingroup$

I plan to do a meta-analysis including the effect size of each cohort as the objectives. I'd like to ask if I could apply multiple test corrections (i.e. Bonferroni) on the pooled effect p-value.

I couldn't find a similar paper or other reference. I'd be very grateful if you could give me any suggestions.

Improve this question
$\endgroup$
1
  • $\begingroup$ There are a number of methods for combining p-values which can be useful if you do not have effect sizes but if you do then why not meta-analyses the effect sizes? $\endgroup$
    – mdewey
    Commented Jul 4, 2023 at 11:23

1 Answer 1

Reset to default
0
$\begingroup$

P-values for summary effect sizes are not useful. Typically, we consider the summary effect of a meta-analysis "exploratory", and do not perform adjustments for multiple tests. If you have been concerned with type-I error rates ("false positives"), you can:

  • select the most homogeneous set of studies possible,
  • use 99% confidence intervals.
  • employ a random-effects model
  • describe the variability via 95% predictive intervals.
Improve this answer
$\endgroup$
1
  • $\begingroup$ Thank you for your comments. In this meta, I regard each cohort as each 'study' and I wanna know if the pooled effect is significant across the cohorts. $\endgroup$
    – Joe
    Commented Jul 4, 2023 at 6:21

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.