PSYCH 100A Lecture 14: Psych 100A W7L2 2/23/17
Document Summary
T-tests are only useful for two groups but that is generally not the case. We would have to do multiple t tests if there are more than 2 groups. Each test has a 5% chance of a type i error. If we do multiple t-tests, we increase the probability of a type i error not accurate if there are many groups. The overall rate of error is not 5% x the number of groups, generally less. This is because there is some sort of relationship between the groups. Used with two or more nominal independent variables and an interval dependent variable. Null hypothesis compares all of the different means together. There is no difference anywhere in the data. Omnibus null hypothesis: includes all pairwise and complex hypotheses. H0: 1 = 2 = = n. Complex hypotheses: comparing two/more groups to another two/more groups at a time (not one group to another group)
