ABSTRACT
This chapter presents a catalogue of different ANOVA designs, listed below. For designs 1–17, all factors other than ‘subject’ factors are assumed to be fixed. If you wish to use other random factors, the expected mean square calculations are different (pp. 48, 164, 168, 170), and so the effects of factors may need to be tested against unusual terms. That may also be the case if the predictors are themselves correlated (pp. 170→,174→). It is possible to work such EMSs out by hand (e.g. Myers & Well, 1995,p. 262) but it's much easier to tell your software package what you want and trust it to sort out the maths.
