Statistical models for school effectiveness rankings are assessed. An analysis of linear models with fixed regression parameters and models with separate regression parameters for each school is presented. In the first case, the regression lines are parallel and the ranking of the schools is the same for all student backgrounds. In the second case, schools may be ranked quite differently for boys and girls, for blue-collar and white-collar children, and for high-IQ and low-IQ students. In the example presented--a 1959 data set of 1,290 children in 37 schools in Groningen, the Netherlands--it appears that slopes do indeed differ; hence, rankings of schools differ across different backgrounds. Analysis of the significance of this effect was attempted. More recently, random coefficient models have been proposed for school effect analysis. These models are considered to be a random sample from a population of schools, and statements are made about this population and not necessarily about the individual schools in the study, interpreted as subpopulations. In random coefficient models, the residuals in the regression equations are not independent for students in the same school. This means that predictions will generally be more conservative than in fixed coefficient models. It is argued that random coefficient models are more appropriate for school effects analysis. Seven data tables are included. (TJH)