Power Analysis

Home > Statistics Every Writer Should Know > The Stats Board > Discusssion

Power Analysis
Message posted by Holly on October 17, 1999 at 12:00 AM (ET)

Is anyone able to describe the procedure for selecting the number of subjects needed to obtain decent power in a correlational design? I have three groups and am trying to correlate 5 variables. Thanks,
Holly

READERS RESPOND:
(In chronological order. Most recent at the bottom.)

Re: Power Analysis
Message posted by JG on October 17, 1999 at 12:00 AM (ET)

Power is type II error and depends on you alternative hypothesis. Usually you set up a test to control alpha or type I error and then you check for type II error. Even if you can control n , you can only control power for one alternative hypothesis.

Your $5 contribution helps cover part the $500 annual cost of keeping this site online.