<![CDATA[This study examines how Self-Regulated Learning (SRL) shapes academic self-efficacy in Generation Z students in Islamic Religious Education and Character Education (PAI & BP) classes. The design is qualitative-dominant mixed-methods: quantitative data supply a contextual baseline; qualitative data carry the main analytical weight. The research site is SMA Praja Nusantara in West Sumatra, a semi-military boarding school where students have no routine access to electronic devices. This condition makes SRL both harder and, arguably, more consequential. Forty-three students participated in the quantitative component; four were selected for in-depth interviews based on post-test score profiles. Pre–post score comparisons showed statistically significant change in both the SRL group and the comparison group (paired t-test, p < 0.001). That result, however, understates what actually happened: aggregate scores flattened differences that mattered. Thematic analysis drawing on observations, semi-structured interviews, and learning documentation analyzed through Braun and Clarke's framework identified six themes: a shift in the locus of learning control, externalization of cognition through mind mapping, widespread difficulty with self-monitoring, domain-specific self-efficacy gains, SRL as a resilience mechanism, and gradual adaptation. Tentative patterns also pointed to SRL's role in the internalization of religious values, though the evidence for this is limited and warrants further investigation. The SRL–self-efficacy link is neither direct nor consistent across students. It works through psychological mechanisms that vary by individual, domain, and context. These findings extend Rahman et al. by tracing an alternative pathway and by testing Bandura and Zimmerman's reciprocity postulates under real boarding school conditions rather than assumed ones. Practical implications for PAI pedagogy in similar settings are discussed.]]>
