The differential predictive validity of a test of academic literacy for students from different English language school backgrounds

Abstract

In the past decade, testing for academic readiness has become common practice across the South African higher education landscape. The aim of this assessment is to identify students who might not be ready for academic education and to put support programmes in place to help bridge this gap. Given the implied contribution that these programmes are expected to make towards higher graduation rates, it is important that the tests used for identifying those who need the support are valid for all students regardless of background. The aim of this study was to investigate the differential predictive validity of a test of academic literacy for two groups of students; those who took English at Home Language level and those who took it as First Additional Language at school. A regression analysis of a total of 564 scores obtained by the two groups on the Test of Academic Literacy Levels (TALL) as a predictor of their end of first-year average scores was carried out to determine this. The results show that although the Home Language group performed better than their counterpart on the test on average, the test predicted the outcome variable marginally better for the First Additional Language group than it did for their counterpart. This difference was too negligible, however, for one to conclude that the test was predictively biased against any of the two groups.