PSYC 3000 Lecture 9: 3000 oct 8
Document Summary
Must say what test was done, if it is significant and what direction the statistic went. (higher or lower) If you are using a test its because your sample is smaller than 30. This means you need to prove your population is normally distributed in order to run these tests. This means you need to do a test of the skew or test of kurtosis. These tests are done before t test and z test. (observed skew expected skew)/standard error of the skew. Formula of standard error of the skew is squared 6 over sample size. This then gives you a z value which you need to turn into p value. (observed kurt expected kurt)/standard error of the kurt. Formula of standard error of kurt is squared 24 over sample size.