Abstract:

The use of an overall measure of school effectiveness even in value addition could prove misleading and be of little use in school improvement because it does not indicate whether schools differentiate among different groups of students. The aim of this study is to investigate whether senior secondary schools in Botswana are effective for students of different prior ability and gender. A value added methodology controlling for prior attainment and gender was employed. MLwiN 2.10 Beta (4), based on multilevel modelling analysed the Botswana General Certificate of Secondary Education (BGCSE) examination results from a sample of 5 662 candidates for the period 2005 - 2007. There was some evidence of differential school effectiveness for students of differing prior attainment levels and gender. Most schools promoted the progress of boys more than that of girls (89% compared to 11% of schools) and mid ability students (67% compared to 11% for high ability and 48% for low ability students). Effective schools also differentiated by gender and prior ability. In addition, some schools which were not effective overall were effective for some groups of students. The implications for school improvement and the evaluation of schools’ performance are discussed.

Keywords:

Differential school effectiveness, value addition, multilevel modelling, school improvement.