Im struggling with the power in general..I need assistance with how to calculate sample size with confidence intervals as to be able to conclude that the sample of the present study is within XX % confidence derived from the population. (or something to that effect).
Also, since I have three main hypotheses, how do I go about sample size? do i use the most complicated one involving most predictors or?
So in terms of power, I performed a sensitivity power analysis in G*Power for multiple regression, increase in R2, results indicated that with my sample size of N=160 pairs of participants, I would have sufficient power to detect a MDE of f2 = .06. I thereafter ran an a priori power analysis which suggested that my sample size would be sufficient to detect a large and moderate effect size but not a small (f2 = .02). When I tried calculating sample size as per recommendation by Kelley & Maxwell (2003) I got confused since they used the F-value. So, do I use the F-value of a model reported in previous research (even if that model is not "identical" to mine? or is there a way to calculate individual predictors perhaps? Is it possible to pool F-values (as in e.g metanalysis pooling effect sizes? (which was my initial thought) but I dont know how to do that..