PS296 Chapter Notes - Chapter 16: Squared Deviations From The Mean
Document Summary
Most of our anova computations deal with the sum of squared deviations about the mean because unlike mean squares, they can be added and subtracted. We take the total sum of squares sstotal and partition/decompose it into the part that is due to variation between groups (ssgroup) and that part that is due to variation within groups (sserror) Our charts will have individual observations (xjj), individual group means (xbarj) and the grand mean (xbargm) The ss total (total sum of squares) represents the sum of squared deviations of all the observations from the grand mean, regardless of which treatment produced them. Ssgroup is a measure of differences due to groups (differences between group means) and is directly related to the variance of the group means. We square and then sum the deviations of the group means from the grand mean and multiple it by the sample size to produce our second estimate of population variance if.
