Introduction to Analysis of Variance
To develop an understanding of how ANOVA works, lets assume, we really didn’t do our treatment, and so the subjects are really just a random sample drawn from the population and randomly assigned to the various cells.
Chance accounts for all observed differences
In this case the best estimate of the variance (S2) is found by summing the squared deviations of each score from the GRAND MEAN. This is variance as we have always known it. In this context we call it TOTAL VARIANCE.