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.