ABSTRACT
In the realm of statistical inference used to sort the signal from the noise, there is a hierarchy of methods. We have studied these methods from the bottom up. We started with simple t tests and then moved up to the more general ANOVA. Now, we just have one last method of analysis to discover. We have seen that t tests are restricted to investigations of dichotomies, and ANOVA can go beyond the simple comparison of two populations at a time and look at n populations with a single risk. The general linear hypothesis based on least squares is at the top of the hierarchy and can do anything that ANOVA or t tests can plus a lot more.
