A Study of school effectiveness in Primary Schools using Hierarchical Linear Models

Abstract

INTRODUCTION. A census study based on a secondary analysis of a previous diagnostic evaluation (DE) is presented on school effectiveness in Primary Education Schools in the Autonomous Community of La Rioja. The impact of contextual variables on the performance of the subjects evaluated was assessed, not using the gross mean scores obtained in the DE, but the adjusted mean scores once the effect of contextual variables were controlled. METHOD. The control was carried out by multilevel statistical modelling with Hierarchical Linear Models at two levels: students and schools. The residues of the models were used as criteria for school effectiveness. RESULTS. A descriptive analysis by levels of the contextual variables and of the performances
is shown, as well as a detailed analysis of the hierarchical models by subjects. From the residues of the models, an arrangement of the schools according to their effectiveness is achieved. DISCUSSION.
The Hierarchical Linear Models proved to be an adequate tool to know the impact of the contextual variables in school learning, beyond the performance of the schools previously provided in DE. The models made it possible to know the schools with the best and worst results of controlled efficacy and the effect of the context. The hierarchical strategy detected that in the relation of schools there are schools from both networks and from very diverse socioeconomic levels, as well as that schools that in one subject obtain a very high (or very low) residue, do not have the same in others, information that is impossible to obtain with more
conventional analysis tools.