Using Simulation To Study School Effectiveness

School-effectiveness research is constrained by ambiguous factors of effectiveness and a lack of theory. This paper presents findings of a study that used simulation to improve school-effectiveness theory. Simulation is also used to explore the direct effects of schools on individual learning. After introducing simulation models, the paper describes a multilevel simulation model to simulate learning in a classroom environment over several years. Three experiments were conducted to validate data structures, compare differences between schools and classes, and generate hypothetical effects of policy changes. The first experiment analyzed longitudinal data from about 4,100 students and found a correlation among student-background characteristics (socioeconomic status, gender, IQ, and achievement). The simulated data were structured similarly to the actual-education data. In the second experiment, the model was able to generate differences between schools and classes. Finally, the third experiment generated hypothetical effects for two major policy innovations. Five figures are included. The appendices contain information on secondary education in The Netherlands and a mathematical formulation of the model