Hi All, I came across this problem which is more a statistical question, I am hoping that some one can clarify this for me Is there a rule of thumb to determine how large your sample size should be before you perform multiple regression. I have search the web and I found some online tools where you provide the Alpha level, number of predictors, the effect size and the desired power. I have googled the effect size and I found that the effect size= R2/(1-R2) Where R2 is the coefficient of determination. I do not understand how to provide the effect size before doing the regression and what it means? From my understanding, the power of a test is the probability of rejecting a false null hypothesis. In this case what is the hypothesis we are testing? Would the null be testing All the BETA coefficients are ZERO simultaneously? Is it possible to look at the power to test for each BETA separately? i.e H0: B=0 vr Ha: B not equal to 0 I appreciate your help -- View this message in context: http://www.nabble.com/Power-for-Multiple-regression--tp25787420p25787420.html Sent from the R help mailing list archive at Nabble.com.