I am using G*Power 3 (Erdfelder, Faul, & Buchner, 1996) software to calculate the sample size. My study is to find whether there is a relationship between traits A and B by using correlation. I read a paper suggest selecting f test in the software, and another paper suggests selecting correlation of bivarate normal model. However, under similar parameters, the first setting results about 100+ samples, but in the latter setting, it results in only 20 samples. I am not sure which figure I should take and which setting is the correct setting.
Thanks
I'm not sure what you want to do exactly but a good rule is if you have two estimates for n choose the larger as a safety factor. Best wishes David Booth
David Eugene Booth , thanks for your reply. The problem is which setting should I use in the test family, select the "F tests> Linear multiple regression: Fixed model, R2 deviation from zero" or select "Exact> Correlection Bivariate normal model". I have put the screen captures of the UI for both options for easy understanding.@
Reference:
Faul, F., Erdfelder, E., Buchner, A., & Lang, A.-G. (2009). Statistical power analyses using G*Power 3.1: Tests for correlation and regression analyses. Behavior Research Methods, 41, 1149-1160.
I noticed that in the linear multiple regression option, you entered "2" under "number of predictors". However, for a bivariate relationship (A correlated with B), you would only have one predictor variable. This may account for some of the difference in the sample size estimates.
- Badges
- Science topic
- Similar topics
- Statistics
- Survey