Effectiveness criteria in school effectiveness studies: Further research on the choice for a multivariate model

In order to ascertain which school characteristics can explain the differences in effectiveness between schools, important methodological choices have to be made in school effectiveness research. One of these choices relates to the criterion or criteria the researcher wishes to use to compare schools. Should, for example, a school be deemed effective if it achieves high performances in mathematics, or in mother tongue or in both? In order to safeguard the construct validity of the concept 'school effectiveness' this type of research should include multiple effectiveness criteria and thus reduce the risk of 'mono-operation bias'. In that case two analysis strategies can be used: a series of univariate analyses (one for each output measure); or a multivariate model in which different dependent variables are modelled at the same time. In this methodological review we will examine what the implications of these different analysis methods are for research results. The dataset is derived from a large-scale effectiveness study conducted into technical and vocational secondary education in Flanders. Based on the results of this study we will make methodological suggestions for researchers in the field of educational effectiveness and school leadership. (C) 2009 Elsevier Ltd. All rights reserved.