PSYC 2022 Lecture 5: Notes stats
22 Feb 2017
School
Department
Course
Professor
Document Summary
Both f-ratio & tstat compare actual mean diff bw sample means (num) w/ the diff that would be expected if no treat effect (h0 is true). If num is sig bigger than the den, conclude there is a sig diff bw treat. 1. : (cid:2869)=(cid:2870)=(cid:2871) (no diff bw pop (treat) or diff bw m are caused by random factors (sampling error)) (cid:1006) (cid:1007) (cid:894)all (cid:1007) (cid:373)ea(cid:374)s are diff(cid:895) H1:at least 1 treat m is diff treat cond are not all the same diff are caused by systematic diff bw sample m h1: 1 = 2 but 3 is diff or h(cid:1005): (cid:1005) . State alpha repeated/related notes differences bw the den of f in independent measures vs repeated is that for an imd, the variability within treatments is the appropriate error term. For repeated measures, however, you must subtract out variability due to id from the variability within treatments to obtain a measure of error.
