EDUCATION SUPPORT PARTNERS' INVOLVEMENT AND PUPILS' ACADEMIC PERFORMANCE IN PRIMARY SCHOOLS IN BUKANGA COUNTY, ISINGIRO DISTRICT, UGANDA: EXAMINING THE RELATIONSHIP

Education Support Partners (ESPs) are widely assumed to improve pupils' academic performance, yet this relationship is rarely tested empirically at the level of discrete intervention dimensions within a single rural Ugandan district. This article, drawn from a wider study, examined the relationship between ESP involvement and pupils' academic performance in primary schools in Bukanga County, Isingiro District, Uganda. A cross-sectional correlational design nested within a mixed-methods approach was used. Correlation and regression analyses were conducted using teacher-level data (n = 20), the only respondent category providing paired ratings of both ESP involvement and academic performance, while qualitative data were obtained from headteachers and ESP representatives across four primary schools. Data were analysed using Pearson correlation, multiple linear regression, and thematic analysis. Findings revealed a strong, positive, and statistically significant relationship between overall ESP involvement and pupils' academic performance (r = 0.631, p = 0.003). Among individual dimensions, teacher support and training showed the strongest correlation (r = 0.601, p = 0.005) and was the only dimension to remain an independent, significant predictor in the regression model (β = 0.415, p = 0.020), which jointly explained 52.3% of the variance in academic performance (R² = 0.523, F(5, 14) = 3.071, p = 0.042). ESP involvement is significantly and positively associated with pupils' academic performance, with teacher-focused support exerting the most robust independent influence, although the small correlational sample and cross-sectional design warrant caution in interpreting causality. ESPs and school administrators should prioritise sustained, well-coordinated teacher support and training alongside other service dimensions, while future research should employ larger samples and longitudinal designs to strengthen causal inference.