PSYC 2002 Lecture Notes - Lecture 12: Effect Size, Statistical Hypothesis Testing, Level Of Measurement
Document Summary
Simple solution to problems with the z test: (cid:1852)=(cid:4666) (cid:4667, (cid:1852)= (cid:1872)= . Shows up in t tests and in variance: using the sample mean to calculate the sample sd, the use of the sample mean restricts the variability of the remaining scores. Ex: the critical value depends on your sample size. If n=2 and m=6, then the sum of the scores in the sample must be 12. If one of the scores=8, then a restriction is placed on the identity of the other score, it must be a 4: degrees of freedom is going to determine what the critical value is going to be. If you have access to the population sd, use the z test. It takes advantage of the best information that you have available: resea(cid:396)(cid:272)he(cid:396)s do(cid:374)"t usually ha(cid:448)e this i(cid:374)fo(cid:396)(cid:373)atio(cid:374). If you do(cid:374)"t ha(cid:448)e the sd, use the t test. If you are thinking of critical values, you look at the alpha you want to use.
