To estimate the mean error in the population we need to divide not by the number of scores contributing to the total, but by the degrees of freedom (df), which is the number of scores used to compute the total adjusted for the fact that we’re trying to estimate the population value
