Across all investigated hypotheses: What % of them is actually true?
What is your Type I error (α; typically 5%)?
Do you want to specify power directly or indirectly by specifying sample size per group and effect size? (Assuming a two-group t-test)
% of studies that report a significant result, although it's not