SOC202H1 Lecture : SOC202 FEBRUARY 28.docx
Document Summary
T test only two groups, often more then two groups, do different t test between but increases making type 1 error, reject null when we should have retained it. Differences in sample means across categories, interval level data. Asks are differecens between sample large we conclude pop reps by samples are different. Not have data for every indi in the world. Null rep by h not mu 1 = mu 2 = m 3 Null rejected at least two means are different not all are different. So alt hypo, 2 of 3 or more significicalty cary from each other. No slash bc we do not know where difference occurs. If null true sample means be same value. Test not specify which pop mean is different. Reject null dif in pop rep by sample. Anova f stat and f distribution page 368. Normal curve not curve anymore, no two tailed tests only one tail.
