Working to Advance
Health Equity in Africa

Determinants of school dropout in the Asenze cohort study in KwaZulu-Natal, South Africa

Determinants of school dropout in the Asenze cohort study in KwaZulu-Natal, South Africa

South Africa remains characterised by high rates of school dropout, exacerbated by the COVID-19 pandemic. The consequences lead to compromised employment opportunities and other negative social and health outcomes associated with out of school youth (OSY). Persistently high dropout rates have prompted calls for more evidence on their underlying causes, to inform interventions that support school retention in South Africa. Data for this paper comes from the Asenze study, a longitudinal epidemiologic study of health and psychosocial needs, which began among preschool children in 2008 in KwaZulu-Natal, South Africa. This paper draws on data collected in the third (2019–2021) and fourth (2022) waves. Multiple logistic regression models assessed the independent association between three sets of variables; individual, cognitive/academic, and household-level on the likelihood of dropping out of school. Approximately 5% (n = 54) of the sample had dropped out before completing secondary school by the fourth wave of the study. School dropout was associated with individual factors; currently engaging in an age-disparate sexual relationship, having had a child, and scoring poorly on the WHO Disability Assessment Schedule. Among the cognitive measures assessed, only planning ability was significantly associated with school dropout; none of the household variables showed a similar link. Pregnancy and child-rearing responsibilities remain the primary reasons cited for school dropout amongst women, affirming associations found in other studies undertaken in South Africa. The association between school dropout and cognition, one of the key components that underly academic performance, has not previously been explored in this context. This research contributes to a limited evidence base by employing longitudinal data and a multi-dimensional approach that includes individual, cognitive, academic, and household-level factors, with a focus on rural and peri-urban settings.