Teacher characteristics do not predict self-efficacy for learner-centered instruction in science and mathematics: A null finding from Zambian teacher education

Source abstract

Learner-centered instruction is widely promoted in science and mathematics education. Yet, the extent to which teacher demographic characteristics shape self-efficacy for such practices in developing-country contexts, particularly within university-based teacher education programs in Zambia, remains underexplored. This study examined the self-efficacy for learner-centered instruction among 257 pre-service and in-service science and mathematics teachers enrolled in university-based teacher education programs in Zambia. The study employed a cross-sectional survey design to investigate whether demographic characteristics collectively predict self-efficacy for learner-centered instructional practices. Descriptive findings indicated consistently high self-efficacy across demographic groups. However, multiple linear regression analysis revealed that age, gender, teaching experience, qualification, category of teacher, and subject specialization did not significantly predict teachers’ self-efficacy beliefs. The regression model was not significant, F (7,249) = 1.33, p<i> </i>= 0.237, and explained only 3.6% of the variance in self-efficacy (R² = 0.036). Post-hoc power analysis indicated insufficient power to detect small effects (achieved power ≈ 0.35). These findings tentatively suggest that professional learning experiences may warrant greater attention than demographic characteristics in future better-powered longitudinal research<i>. </i>Implications for the design of STEM teacher education curricula are discussed.