PSY 350 Lecture Notes - Lecture 42: Type I And Type Ii Errors, Analysis Of Variance, Complement Factor B
Document Summary
If we conducted three separate t-tests (or anovas) to test those three hypotheses, our type i error rates would add up: By testing them all in one analysis, we keep our type i error rate under control. We set an alpha level for the whole test, all together. We can also better understand our effects in context. Interactions are easier to interpret in a two-way anova. Just as we did in repeated-measures anova, we partition the variance a little further in a two-way anova. But this time, we focus on the between-groups variance. We"re going to calculate three related f statistics, all at the same time: One for the main effect of factor a. One for the main effect of factor b. In repeated-measures anova, we subdivided the within-groups variance. In two-way anova, we subdivide the between-groups variance: This looks just likethe way we calculated fin the one-way anova. We find a ss within each group (condition)
