Assessment of non-verbal intelligence in South African schools : do language and gender bias performance on the Raven's Standard Progressive Matrices?

Abstract

Test bias within the field of psychometrics is an issue of concern in the South African context. Bias refers to whether a test measures what it claims to measure across different groups. In South Africa, psychological testing has been associated with the oppressive and discriminatory practices of the Apartheid era, leading to many tests being banned and considered invalid and unfair (Foxcroft & Roodt, 2004). Research is required in South Africa to investigate the effects and functions of psychometric tests in the context of the country’s history and diverse population groups. This research investigated whether the Raven’s Standard Progressive Matrices (SPM) is a good measure of non-verbal intelligence for high school children in the South African context, or whether it is biased by language and gender among Grade 8 and Grade 9 learners. The subjects were between 13 and 15 years old and attended an English medium government high school in Johannesburg. The archival data was collected over a two year period and the SPM was administered to some subjects in both Grade 8 and 9. A 2-way ANOVA tested for an overall difference on the SPM scores between learners who spoke English as a first language (EFL) and English as an additional language (EAL), as well as the performance of Male and Female learners. There was a significant difference between the language groups in Grade 8 but not Grade 9. There were no significant gender differences. An item analysis tested for a language or gender difference on the item level. Certain test items showed a gender difference while others a language difference, but no significant trend was found across the test. A matched paired t-test for the group that received the test in both Grade 8 and 9 determined that there was a significant increase in scores with greatest gains for the Male and Female EAL learners.