Hello,
I am confused about the best practice regarding power analysis for my meta-analysis in my specific case.
1. There is no earlier meta-analysis or even systematic review which gives me an a-priori estimation of average study sample size or the number of studies.
2. My meta-analysis is exploratory, aiming to synthesize the available (conflicting) research on the association between A and B.
--> Therefore no a-priori power analysis was performed.
But:
The number of studies IS very small and in the discussion, the possibility of low power (caused by and expected because of small k of studies and large heterogeneity) will most likely be addressed anyway.
Now, I wonder if it'd be bad scientific practice to calculate the power of the meta-analysis based on the final set of studies included in the meta-analysis - which would provide me with a metric for the discussion, but ultimately does not change the point I am probably about to be making.
I understand that retrospective power analysis are frowned upon in primary study design. But as with meta-analysis I have no control about the amount and design of primary studies .... I wondered if the explanatory aspects of power-analysis can be utilized.
Thank you for your answer.