Abstract

AJOODHA, Ritesh. Identifying academically vulnerable learners in first-year science programmes at a South African higher-education institution. SACJ [online]. 2022, vol.34, n.2, pp.120-148. ISSN 2313-7835.  http://dx.doi.org/10.18489/sacj.v34i2.832.

The Admission Point Score (APS) is used by most South African universities to identify a university programme in which a learner is likely to succeed. While the APS appears helpful to gauge the aptitude of a learner and predict their success, the reality is that between 2008 and 2015 almost 50% of learners who made the required APS for a Science programme failed to complete the requirements for that programme. This paper delineates and diagnoses learner vulnerability, using a learner attrition model, for early intervention and as an alternative to using the APS. The analysis shows that various predictive models achieve higher accuracy to predict learner vulnerability, by incorporating factors of the learner attrition model, rather than just using the APS score. This paper argues for a more complex view of predicting learner vulnerability for early interventions by incorporating the learner's background, individual characteristics, and schooling data. It does not agree with the aggregation of National Senior Certificate (NSC) subjects into APS scores since this normalises the complexity of the subtle relations between the schooling system, learner attrition, and pre-schooling pedagogical dynamics. This paper points to a more nuanced view of predicting learner vulnerability. CATEGORIES: • Applied computing ~ Education • Computing methodologies ~ Machine learning

Keywords : Predicting learner vulnerability; Identifying at-risk learners; Early intervention; South African higher-education universities; Model-based machine learning; Education.

        · text in English     · English ( pdf )