Robust Methods for Factorial Experiments with Outliers

Two factorial experiments with possible outliers are reanalyzed by means of robust regression techniques. We show that using M-estimates of regression results in efficient analyses which are easier to implement than the previous methods.