How does mathematics self-concept shape approaches to learning? Evidence from prospective mathematics teachers

We examined how mathematics self-concept influences approaches to learning mathematics among prospective mathematics teachers (PMTs) in Nigeria. Using a cross-sectional survey design within the context of students’ approaches to learning theory, we generated datasets from 200 final-year PMTs across government-owned higher institutions in Nigeria. This theory provides a framework for explaining differences in approaches to learning and the factors that affect them. We measured the constructs using validated instruments and analysed the dataset through structural equation modelling. The measurement models demonstrated appropriate construct validity with omega reliability coefficients ranging from .788 to .879. Contrary to theoretical expectations, mathematics self-concept showed a significant negative effect on deep approaches to learning (β = −.685, p < .001) and a significant positive effect on surface approaches to learning (β = .307, p < .001), with the structural model exhibiting an appropriate fit (comparative fit index = .955, Tucker–Lewis index = .941, standardised root mean square residual = .055). These suggest that within the PMTs context, strong mathematics self-concept may be grounded in examination-oriented success, thereby reinforcing surface learning strategies rather than conceptual engagement. Conversely, lower self-concept appears to prompt compensatory deep learning efforts. These findings are theoretically provocative as they challenge the directions of effects between research constructs.