HSS 2381 Lecture Notes - Lecture 7: Effect Size, Data Matrix, Family-Wise Error Rate
Document Summary
Z and t stats can be negative. Used for 2 or more measurements (2 or more groups, 1 or more factors, 2 or more levels) T-tests can only have 2 groups, with 1 factor that has 2 levels. Most common procedure for interval and ratio data is anova. Human brain cannot interpret results of anova"s with more than 5 factors. Interaction: the effect of one factor on another factor. H0: u1 = u2 = u3 etc. H1: there is at least 1 difference among the population means. Source of variance, sum of squares, degrees of freedom, mean squares, f-ratio, significance. S/a = unsystematic variance (individual differences error) F = variance of group means / avg variance within groups. Multiple factor designs let us simultaneously examine several effects. Anova has experiment-wise error rate (type i error rate is constant for all comparisons of groups)