Some predictors of mathematics achievement among black secondary school learners

This study was conducted to identify predictors of mathematics achievement among grade 9 learners of a random sample of five township schools. A series of regression analyses were performed for boys and girls separately to obtain Cohen's (1992) effect size estimate (uniquely explained criterion variance expressed as a proportion of unexplained criterion variance) for various predictor variables, Cognitive predictors were verbal and non-verbal General Scholastic Aptitude Test scores. Non-cognitive variables included the hierarchical levels of self-concept: global (Rosenberg Self-Esteem Scale), and academic and mathematics self-concept (relevant scales of Brookover, Erickson and Joiner). Socio-economic predictors included home-related variables (parental education, parental occupation, family size) and school-related factors (class size, teacher's qualification, teacher's experience). Gender differences favouring boys were found. Non-verbal and verbal scholastic aptitude and teacher's general training correlated significantly with mathematics achievement for boys and girls, with nonverbal scholastic aptitude showing the highest correlation and effect size estimate for girls and teacher's general training occupying this position for boys. Teacher's mathematics training and class size showed correlations in excess of 0.35 for boys but not for girls. The negative corrrelation obtained for teacher's general training suggested that learners whose teachers held a three-year teaching diploma performed better in mathematics than did learners whose teachers held a degree and a teacher's diploma.