Understanding school effects in South Africa using multilevel analysis: findings from TIMSS 2011

Abstract

Studies of school effect usually evolve around two major issues; understanding of the factors within schools that explain variation in students achievement levels, and how differences among schools account for the variation in their student' achievement levels. Given the limited studies of school effect research in South Africa, this paper explores how different characteristics of students and school are related to performance. Using the 2011 South African TIMSS (Trends in International Mathematics and Science Study) data, the study employs multilevel analysis to identify the most important indicators of successful schools in South Africa. The initial analysis entails the partitioning of the overall variation in students' achievement level in mathematics into within and between schools that allows us to estimate school effects (the extent to which schools make a difference in students' achievement levels). A follow-up analysis involves the addition of student and school characteristics in the multilevel models to profile students and schools that are most successful. The multilevel analysis indicated that differences in schools account for about 62% of the variation in students' achievement levels in mathematics. Students are more successful when they attend schools that are adequately resourced, have teachers that are satisfied with their working conditions and that are specialised in the field of mathematics. The results also show that the attitude of the students toward mathematics has an impact on their performance. The findings demonstrate the need for the South African educational policy makers to continue improving the learning environment of less resourced schools and also encourage teachers to instil a sense value and self-confidence in learning mathematics and to improve teacher qualification.