PSY 350 Lecture Notes - Lecture 37: Type I And Type Ii Errors, John Tukey
Document Summary
We know there"s a difference among these means overall, so the highest and lowest must be different. It tests the equality of several means all at the same time. H0: 1 = 2 = 3 = = k. This keeps our type i error under control. If = . 05, we will only reject h0 5% of the time when the null is true, no matter how many means we are testing. When the overall f statistic is significantly larger than 1. 0, we know that at least two of the means are different. But we don"t know yet whether only the extremes are different, or if there are also differences among the means in the middle. To fully understand our anova results, we need to compare all possible pairs of means against one another. We could do this with a bunch of t-tests but that loses the benefit of an anova. Uses the information from our anova to set a standard: