Exploring the mediating role of willingness to communicate between self-regulation, professor-student rapport, and learner engagement in GenAI-supported learning: A mixed-methods study

The present study investigates the relationships among self-regulation, professor-student rapport, willingness to communicate, and learner engagement in GenAI-supported higher education from the perspective of Social Cognitive Theory. A mixed-methods design was employed. Quantitative data were collected from 519 undergraduate students through questionnaires, complemented by qualitative interview data from 20 participants. Structural equation modeling using SPSS and AMOS was conducted to test the hypothesized relationships, while qualitative data were analyzed using NVivo to explore underlying mechanisms. The quantitative results revealed that self-regulation significantly predicted willingness to communicate (β = 0.375) and learner engagement (β = 0.263), while professor-student rapport significantly predicted willingness to communicate (β = 0.225) and learner engagement (β = 0.255). Willingness to communicate also positively predicted learner engagement (β = 0.372), confirming its partial mediating role. Qualitative findings further showed that willingness to communicate functions as a process-oriented mechanism by triggering learning action, buffering negative emotions, translating relational support into participation, and amplifying engagement in AI-supported contexts. These findings advance Social Cognitive Theory by elucidating how learner engagement emerges through communicative agency and offer practical implications for designing GenAI-supported instruction in higher education.