Examining a four-process GenAI cycle reveals two distinct metacognitive regulation styles: Exploratory-Simplification and Systematic-Methodical, which show that students use combinations of strategies across the AI-SRL cycle.
Abstract
Generative AI (GenAI) is now common in university project work, yet previous studies often examine students’ trust, creativity, overload, or engagement separately. This leaves a key gap: how students regulate GenAI across a full project workflow. This exploratory study addresses that gap by examining a four-process “
AI-mediated Self-Regulated Learning
” (
AI-SRL
) cycle: (1) evaluating and selecting GenAI suggestions, (2) experiencing shifts in creative agency, (3) managing overload through filtering and summarizing, and (4) monitoring time and energy to stop or continue working with GenAI. We conducted a two-course basic qualitative design study with 97 undergraduate and graduate students. Data came from an open-ended questionnaire aligned to the four processes. We used inductive content analysis with a shared codebook, reliability checks, and cross-level comparisons. Findings show that students use combinations of strategies across the
AI-SRL
cycle. They exercise agency through goal alignment, revision, and verification, with graduates reporting stronger cross-checking and source-based justification. Creativity was described as a conditional outcome: it increased when GenAI widened ideas but declined when it replaced personal exploration. Overload was managed through targeted prompts and structured outputs, again more common among graduates. Most students did not lose track of time; they used clear stopping cues such as fatigue, repetition, or satisfaction. Together, results reveal two distinct metacognitive regulation styles: “
Exploratory-Simplification
” and “
Systematic-Methodical
”.
Self-regulated learning (SRL) is an instrumental skill for success in learning computer science (CS) and software engineering. SRL is an active process where students engage in cycles of planning, strategy-use, monitoring and control, and adaptation to accomplish goals. There have been calls to investigate whether SRL theory and measurement need to be made more specific to the CS education context and to integrate SRL theory when interpreting observations of student behavior in computing education research. In this work we examine the self-regulatory behavior of CS students in a 200-level CS course with an emphasis on software engineering. We performed a think-aloud study with thirty-three students working on their programming projects. Researchers then coded the think aloud data using a theory-driven codebook aligning with Winne and Hadwin's COPES model of SRL. Using the coded data we performed a thematic analysis of two codes to better understand how students monitor their understanding of the task and set goals. Our results revealed a more complex picture than prior work of how students monitor their understanding of the task and set subgoals. We discuss the implications of our results for CS instruction and intervention designs to promote more effective SRL and learning.
J. Bacher, Christina L. Hollander, Michael Berro et al.· Annual Conference on Innovat...· 0 citations
This study explores how undergraduate EFL students use chat-based AI tools in essay writing through the lens of Self-Regulated Learning (SRL), focusing on the forethought, performance, and self-reflection phases. A qualitative research design was used, and data were collected through semi-structured interviews with 12 English Education students selected through purposive sampling based on their experience using chat-based AI tools. The data were analyzed using thematic analysis with both deductive and inductive coding. The findings show that in the forethought phase, students mainly set simple goals such as completing essays quickly, getting good grades, and meeting academic requirements. They also used AI to help generate ideas and organize essay structure when they had difficulties. In the performance phase, students used chat-based AI for brainstorming ideas, improving grammar and vocabulary, and developing their writing. However, they did not fully depend on AI output and often rewrote and checked the content to maintain originality and clarity. In the self-reflection phase, students viewed AI as helpful for improving writing speed, ideas, and language quality, but they also expressed concerns about overuse and dependence. Despite this, most students reported improvements in their writing skills and planned to continue using AI carefully. The study concludes that chat-based AI supports students in essay writing, but its effectiveness depends on how students regulate its use across SRL phases. The findings show that students are starting to apply self-regulated learning strategies, although their writing is still strongly influenced by task completion goals and academic demands.
Rifqah· Pubmedia Jurnal Pendidikan B...· 0 citations
LLM-based programming help tools integrated into learning management systems offer new possibilities for supporting students in large programming courses. Yet existing research rarely accounts for the self-regulatory differences that shape who chooses to use these tools or for the ways that technological scaffolds interact with learners’ motivation and help-seeking behaviors. This study addresses that gap through a quasi-experimental design that intentionally delayed the introduction of an LMS-integrated LLM tool, CodeHelp, until after early-semester assessments and substantial measures of student effort had already been collected. Using data from 589 students in a second-level programming course, we find that students who chose to use the tool were already more engaged before it became available: they attended more classes, spent more time in labs, and earned higher scores on the first exam. When these differences were accounted for using logistic regression and propensity score matching, the apparent performance benefit associated with CodeHelp disappeared. These findings suggest that engagement with LLM-based tools reflects underlying self-regulatory behaviors and that the tool functions as a form of technological scaffolding primarily activated by already-engaged learners. Methodologically, the study demonstrates how delaying tool introduction and applying causal inference methods can produce more credible estimates of impact in real classrooms. Pedagogically, it reveals that LLM-based programming support may amplify existing disparities in self-regulation rather than compensate for disengagement. As enthusiasm for generative AI in computing education grows, this study highlights the need for transparent, theory-informed evaluation practices that distinguish genuine learning gains from pre-existing differences in student behavior.
Laura M. Cruz Castro, Maryam K. Multani, Gabriel Castelblanco et al.· IEEE Access· 0 citations
First-year students in STEM programs face significant academic and personal challenges that can undermine retention and success, particularly for those navigating new institutional environments without prior college experience. While self- regulated learning (SRL) theory offers a well-established framework for understanding how students plan and reflect, less attention has been paid to the performance phase, the stage where students must translate plans into action amid real academic and social demands. This qualitative study examines the experiences of 15 first-year life science students across three institution types, a Hispanic-Serving Institution, a predominantly white institution, and a liberal arts college, to investigate what plans students formed at the end of their first semester and what factors facilitated or hindered implementation during their second semester. Using thematic analysis of semi-structured interviews, three major plan themes emerged: help-seeking, internal academic adjustments, and managing social and emotional well-being. Facilitating factors for these plans included small class sizes, anonymized participation tools, approachable instructors, peer and family support, counseling services, and structured planning tools, while hindering factors included fear of judgment, high instructor-student ratios, scheduling conflicts, academic burnout, and unsupportive living environments. The findings reveal that plan implementation depended on the interplay of intersecting psychological, social, and structural factors, which created unique conditions that influenced whether students were able to enact their plans. Importantly, the findings reveal that plan implementation unfolded not as a linear process but through nested micro-cycles of forethought, performance, and reflection within the performance phase, triggered by specific events throughout the semester. These findings have implications for how institutions design learner-centered support for STEM students not only at key transition points, but also throughout the semester, to address the conditions that influence whether students are able to successfully implement, adapt, or abandon their regulatory efforts.
Mehri Azizi, Nicole Chlebek, Bryan M. Dewsbury· Trends in Higher Education· 0 citations
Generative artificial intelligence (GenAI) is increasingly used for feedback in higher education, yet evidence remains limited on how alternative human–AI feedback designs shape learning processes and durable outcomes. This study addresses that gap through a multisite, cluster-randomized, longitudinal field experiment comparing four feedback designs in introductory university science courses: peer feedback only, direct GenAI-supported feedback, reflective GenAI-supported feedback, and a hybrid design combining self-evaluation, peer feedback, and GenAI critique. The analytic sample comprised 1,176 first-year undergraduate students from 48 course sections across four universities and three science domains. Primary and secondary outcomes were argument-quality gain on four shared rubric dimensions—claim quality, evidence relevance and sufficiency, coherence of reasoning, and treatment of limitations or alternative explanations—conceptual learning, and delayed AI-free transfer; feedback uptake and self-regulated learning during revision were modeled as process mediators. Direct GenAI-supported feedback improved immediate argument-quality gain relative to peer feedback, whereas reflective and hybrid designs produced stronger feedback uptake and self-regulated learning. The hybrid condition yielded the highest adjusted mean for immediate argument-quality gain and showed the clearest advantage on conceptual learning; the reflective condition showed a positive but non-significant adjusted contrast on conceptual learning relative to direct GenAI-supported feedback. Both reflective and hybrid conditions outperformed direct GenAI-supported feedback on delayed AI-free transfer. Multilevel mediation analyses indicated that feedback uptake and self-regulated learning partially explained these advantages. By comparing four feedback designs, modeling revision processes, and assessing delayed AI-free transfer in a multisite field experiment, the findings suggest that the educational value of GenAI in higher education may depend less on AI access per se than on whether feedback environments preserve student agency, evaluative judgment, and ownership during revision.
Hüseyin Ateş· International Journal of Edu...· 2 citations
This study examines how digitally mediated learning conditions shape students’ self-perceived employability (SPE) by focusing on two explanatory learning resources: academic self-efficacy (ASE) and online self-regulated learning (OSRL). Drawing on social cognitive theory and social cognitive career theory, this study proposes that digital learning design quality (DLQ), assessment transparency (AT), and perceived challenge-based learning (PCBL) contribute to students’ perceived career readiness by strengthening their academic confidence and self-regulatory ability. A two-stage research design was employed. In the qualitative stage, group discussions with lecturers and interviews with students were conducted to refine the measurement scales and ensure contextual appropriateness. In the quantitative stage, survey data were collected from 335 third- and fourth-year university students in Ho Chi Minh City, Vietnam, who had recently completed at least one online or blended course. The proposed model was tested using partial least squares structural equation modeling (PLS-SEM). The results indicate that DLQ, AT, and PCBL are positively associated with ASE and OSRL. AT is the strongest predictor of both ASE and OSRL, whereas compared with DLQ, PCBL is more strongly directly associated with SPE. In addition, both the ASE and OSRL positively predict the SPE, with the effect of the ASE being greater. The model explains a substantial proportion of the variance in SPE, highlighting the importance of psycho-behavioral learning resources in the translation of digital learning experiences into perceived career readiness. This study contributes to the literature by linking digital learning environment factors with employability perceptions through internal learning resources, while also offering practical implications for improving online and blended course design in higher education.
Thanh Long Vu, Phuc Khanh Nguyen, Doan Anh Tuan Lai et al.· Multidisciplinary Science Jo...· 0 citations