I am doing a meta-analysis and my data is in the format of response ratios (bioconcentration factor = concentration in a plant/ concentration in the soil). The r functions I found for power calculation (e.g. dmetar) always contained either a standardized mean difference (SMD) or an odds ratio (OR), either of which does not seem to be appropriate in my case.

I tried to plug in the difference of my expected effect size (ratio 1.5) - the null hypothesis (ratio 1) = 0.5/SD using the standardized mean difference - would you consider this appropriate?

Do you know any other r-function for power analysis that would cover my data format of response ratios?

  • 1
    $\begingroup$ As you can see by your example, response ratios aren't additive - you should use log response ratios instead. See this question, potentially: stats.stackexchange.com/questions/141443/… $\endgroup$
    – rishi-k
    Jul 14, 2021 at 5:59
  • $\begingroup$ Strictly speaking questions asking for software advice are off-topic here. You might be interested in the mailing list for meta-analysis in R stat.ethz.ch/mailman/listinfo/r-sig-meta-analysis/ You need to register first. $\endgroup$
    – mdewey
    Jul 14, 2021 at 12:35
  • $\begingroup$ Thank you for your replies - and sorry if my question is not appropriate. I am aware that for the meta-analysis I need to use log response ratios, however, my focus is on the power-analysis here, and I am looking for a formula or method to do a power-analysis with response ratios. $\endgroup$
    – Milena
    Aug 6, 2021 at 6:09


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.