Abstract
<p>Most research on students’ GenAI use relies on frequency. Yet frequency captures exposure, not how students use GenAI. This study examined whether patterns of use are more informative for outcomes than frequency. Exploratory factor analysis of items developed inductively from pilot data identified two dimensions: evaluative integration (EI, use to support understanding), and low-verification uptake (LVU, shortcut use). Survey responses from 157 undergraduates were used to test associations with performance and indirect associations through academic self-efficacy (ASE), with frequency included for comparison. EI was positively associated with performance, whereas LVU was negatively associated with it. EI showed a significant indirect association through ASE and no remaining direct association; LVU showed both a significant indirect association and a remaining direct negative association. Frequency predicted neither performance nor ASE. In a separate model, frequency strengthened the positive association between EI and ASE, but the interaction was nonsignificant when the EI and LVU interactions were estimated jointly. No moderation was found for LVU. How students engage with GenAI may therefore be more informative for academic outcomes than frequency alone. The findings support further development of measures of student-GenAI interaction and testing of pedagogical approaches that promote evaluative integration.</p>