Annual meeting of the American Educational Research Association (AERA) location:San Diego date:13-17 April 2009
When assessing the quality of an educational system through national assessments, it is not only important to look at the average performance of students. One should also check whether there are no systematic differences between the schools in the number of students reaching the attainment targets. In case such differences exist, it is important to investigate whether these differences covary with certain student, classroom, or school characteristics. For this purpose, background questionnaires are administered to the students, the parents, and the teachers of the schools participating in the Flemish national assessment. In the present paper, the results of multilevel analyses of some recent national assessments will be discussed. More specifically, the national assessments of French and biology in secondary education and of reading and listening in primary education will be discussed. In general, the background characteristics allow accounting for a high percentage of the observed school differences. At the students level, the proportion of explained variance is relatively low. It will be shown how the schools receive feedback about these analyses.