Abstract
BACKGROUND: In educational practice, the interaction between students' learning behaviors and psychological states is crucial. Learning engagement reflects the degree of active participation in learning activities, academic self-efficacy embodies students' confidence in their own learning abilities, and learning anxiety may affect students' academic performance and mental health. PURPOSE: This study aims to break through traditional linear thinking by using Latent Profile Analysis to identify heterogeneous groups of high school students in terms of learning engagement, reveal differences in their learning anxiety and nonlinear characteristics, and validate the differentiated mediating mechanism of academic self-efficacy between learning engagement and anxiety. It fills the theoretical gap from the perspective of group heterogeneity and provides a basis and support for constructing a learning psychological theory model and precise stratified intervention in educational practice. METHODS: In this study, the method of cluster convenience sampling was adopted. Three ordinary high schools in a city of Hubei Province were selected, and students from six classes in each grade from the first year to the third year of senior high school were sampled as the research subjects. A total of 1,008 original questionnaires were obtained, and 936 valid questionnaires were collected. LPA was employed to identify the potential categories of high school students' learning engagement and the impact of these categories on learning anxiety. Additionally, the mediating role of academic self-efficacy between different types of learning engagement and learning anxiety was examined based on the classification results. Furthermore, a generalized additive mixed model was used to analyze the linear relationships between different types of learning engagement, academic self-efficacy, and learning anxiety. RESULTS: Learning engagement can be categorized into four types, among which students with "High Vigor-High Dedication-High Absorption" type exhibit the highest levels of learning anxiety. Academic self-efficacy plays a mediating role between various types of learning engagement and learning anxiety. Moreover, for students with "Medium Vigor-Low Dedication-High Absorption" type, as well as those with "Low Vigor-Medium Dedication-Medium Absorption" type, there exists a non-linear relationship between learning engagement and both academic self-efficacy and learning anxiety. CONCLUSIONS: This study reveals the heterogeneous characteristics of learning engagement and its differentiated impact on learning anxiety. Furthermore, the study finds that the relationship between learning engagement types and self-efficacy and learning anxiety exhibits a nonlinear characteristic with a turning point, breaking through the traditional assumption of a linear relationship and revealing the possible existence of a 'threshold effect' in changes in learning engagement levels. IMPLICATIONS: This study divides high school students' learning engagement into four potential categories through LPA, breaking away from traditional single-dimensional or dichotomous research perspectives and providing categorical empirical evidence for learning engagement theory, which fills a gap in the sufficient segmentation of groups with moderate engagement levels.