Classification effects, or main effects of classification factors, define a different parameter in the model for each different value or level of the corresponding factor. The levels of a classification factor group the observations and in essence such parameters represent the true expected response for observations in the respective groups, all other factors being the same.