Rethinking “Elaborate Equals Effective”: Unpacking the Impact of Feedback Complexity and Text Type on Learner Engagement With GenAI's Feedback

ABSTRACT While GenAI‐assisted corrective feedback (GACF) shows great promise for personalized learning, how to tailor it for maximum educational impact remains largely underexplored. Pivotal to this issue is feedback complexity. Simple feedback generally confines itself to basic checks and results, whereas elaborated feedback offers more comprehensive scaffolding, including task explanations, hints, and error analysis. Despite the intuitive appeal of elaborated feedback, theoretical and empirical evidence is inconclusive regarding its definitive superiority over simple feedback. Indeed, determining optimal feedback complexity requires careful consideration of the educational context, individual learner characteristics, and task specificity. Contextualized in the Machine Translation Post‐Editing (MTPE) education—a novel and under‐researched scenario—this study employed a 2 × 2 factorial within‐subjects design to examine the effects of feedback complexity (simple vs. elaborated) and text type (literary vs. technical) on student engagement (emotional, cognitive, and behavioral) and revision outcomes (successful, unsuccessful, or no revision). Our findings revealed text type as a key predictor. Literary texts elicited significantly higher emotional and cognitive engagement with GACF compared to technical texts. However, successful revisions for literary texts were paradoxically lower, indicating that higher engagement may not directly translate into more effective revisions. Moreover, while elaborated GACF led to greater behavioral engagement for literary texts, simple GACF elicited higher behavioral engagement for technical texts. These results underscore the need for providing adaptable, task‐sensitive GACF to optimize learning outcomes in GenAI‐supported educational environments.