Self-directed learning self-efficacy in mathematics: The role of anxiety, support, and contextual factors

Abstract

This study examines individual and environmental factors predicting self-directed learning (SDL) self-efficacy among 15-year-old students and investigates the interaction between math anxiety and teacher support. Using nationally representative data from 5,515 students in the PISA 2022 Türkiye sample, weighted logistic regression analysis was employed to estimate the probability of students demonstrating above-average SDL self-efficacy. Missing data were addressed through multiple imputation, and sampling design effects were taken into account. The results indicate that twenty-first-century mathematics self-efficacy, family support for learning, and perseverance in mathematics substantially increase the likelihood of high SDL self-efficacy. In contrast, math anxiety is negatively associated with SDL self-efficacy and shows a significant interaction with teacher support. Specifically, the negative association between math anxiety and SDL self-efficacy was negligible when perceived teacher support was low; however, this negative association became substantially stronger at higher levels of teacher support. Additionally, mathematical motivation and perceiving mathematics as easy are unexpectedly associated with lower SDL self-efficacy. Overall, the findings highlight the importance of contextual and affective factors, alongside individual beliefs, in shaping students’ SDL capacities. This unexpected pattern suggests that the predominantly instructional teacher support measured in PISA may not buffer the adverse correlates of math anxiety but may instead amplify them, possibly because directive and controlling support styles increase performance pressure for anxious students. These results have important implications for educational practice and teacher preparation aimed at fostering students’ autonomy and self-efficacy in learning.