The Emotional Equation: A Mixed Methods Approach into Affective Factors Shaping AI-Assisted ESL Writing Through a Positive Psychology Lens

• In Event: AERA Roundtable Session Sunday 11:45 am - Gold 2
  In Roundtable Session: Educational Psychology and Literacy Development: Examining the role of Affective and Psychological Factors in Language Learning

Abstract

This study explores how learner profiles emerge at the intersection of AI competence and affective experiences—specifically anxiety, enjoyment, and willingness to communicate (WTC)—in AI-assisted ESL writing. Drawing on Positive Psychology, cluster and ANOVA analyses of 291 questionnaires identified three clusters: (1) Emotionally Ambivalent, Moderately Competent; (2) Low-Engaged, Low-Competence; and (3) Positively Engaged, Highly Competent. Follow-up interviews with six participants revealed distinct engagement patterns. Cluster 3 showed reflective, dialogic use driven by high enjoyment and WTC. Cluster 1 reported moderate competence but emotional ambivalence, while Cluster 2 used AI mechanically with minimal affective investment. These findings underscore roles of emotions in shaping human–AI interaction and offer theoretical, methodological and pedagogical implications for affect-informed engagement in AI-assisted L2 writing.

Authors

• Xiangning Li, University of Cambridge

• YUNAN ZHANG, University of Hong Kong

• Belle Li, Purdue University