18.12 Assumptions

The assumption for F tests in a two-factor ANOVA are similar to those for a one-factor ANOVA. All underlying populations (for each treatment combination or cell) are assumed to be normally distributed , with equal variances. A s with the one-factor ANOVA, you need not too concerned about violations of these assumptions, particularly if all cell sizes are equal and each cell is fairly large (greater than about 10). Otherwise, in the unlikely event that you encounter consspicuous departures from normality of equality of variances, consult a more advanced statistics book.