The Relationship between Adolescents' Self-Regulated Learning and Academic Achievement: An Interrelated Mediation Model of Academic Emotions: Evidence from a Nationwide Sample in China

RUI SUN^1,2Sheng Zhang^2*

• ¹Beijing Normal University, Beijing, China
• ²Beijing Normal University Collaborative Innovation Center of Assessment for Basic Education Quality, Beijing, China

ABSTRACT: In the 21st century, self-regulated learning (SRL) plays a vital role in the cultivation of high-quality talent. Grounded in self-regulated learning theory and attachment theory, this study draws on nationwide survey data from China (N = 88,149 students) and employs structural equation modeling (SEM) to systematically examine the relationship between adolescents’ SRL and academic achievement. Academic emotions and teacher–student relationships (TSR) are incorporated as a mediating and a moderating variable, respectively. Building on prior research, the study further investigates whether a reverse pathway exists from academic emotions to SRL, thereby providing an initial empirical test of the recursive mechanism between the two constructs. The results indicate that: (1) adolescents’ levels of SRL are significantly and positively associated with their academic achievement; (2) three types of academic emotions (e.g., positive high-arousal emotions) significantly mediate the relationship between SRL and academic achievement; and (3) TSR moderate both the first stage of the mediation pathway (“SRL → academic emotions → academic achievement”) and the direct effect of SRL on academic achievement. Overall, the findings support the cognition–emotion cyclical model, suggesting that students can enhance SRL through positive academic emotions, while appropriate SRL strategies can in turn foster positive academic emotions, forming a virtuous cycle that strengthens students’ autonomy in the learning process. This study extends Pintrich’s SRL model by empirically validating a recursive loop between SRL and academic emotions. Furthermore, it highlights a "dual empowerment" path of emotional regulation and metacognitive monitoring through optimized TSR, offering vital policy insights for fostering high-order autonomous learning in the AI era.