A Latent Profile Analysis of Emotions in AI-Mediated IDLE: Associations with Emotion Regulation Strategies and Perceived AI Affordances

The rapid development and easy accessibility of artificial intelligence (AI) technology have led to a significant rise in informal digital learning of English (IDLE). However, the emotional experiences across different cohorts of learners remain underexplored. Contextualized in AI-mediated IDLE, the present study integrated the control-value theory of achievement emotions and the process model of emotion regulation to investigate the latent profiles of emotions and further examine their relations to emotion regulation strategies (cognitive reappraisal and expressive suppression) and perceived AI affordances. Questionnaires were administered to 613 English as a foreign language undergraduates in China. Latent profile analysis revealed three emotion profiles, including moderate positive and moderate negative emotions group (Profile 1, 43%); high positive and low negative emotions group (Profile 2, 21%); and high positive and high negative emotions group (Profile 3, 36%). The Bolck-Croon-Hagenaars (BCH) analysis indicated that students in Profile 2 scored the highest on perceived AI affordances, followed by those in Profile 3 and Profile 1. Additionally, multinomial logistic regression analysis showed that cognitive reappraisal was a stronger predictor of membership in Profiles 2 and 3 compared with Profile 1, while expressive suppression predicted membership in Profile 3 to the greatest extent, followed by Profiles 1 and 2. Pedagogical implications were provided to cultivate learners' optimal emotional state.