Mean Squares in ANOVA

Mean squares are sums of squares divided by their corresponding degrees of freedom:

$$ MSTreat = \frac{SSTreat}{k-1}, \quad MSE = \frac{SSE}{N-k}. $$

The treatment mean square estimates variability between treatment means, while the error mean square estimates the common within-treatment variance.