AI evidence extraction

Beyond Automation: Cultivating Self-Regulated Learning through AI and Reflective Practice

Technology Knowledge and Learning · 2026

AI-generated evidence extraction, verified across multiple analytical personas. Not a substitute for the peer-reviewed original.

6/10

Relevance

1/4

Quality (LMQS)

Evidence

Citations

0.00

FWCI

This is an AI analysis. Read the peer-reviewed original at the publisher: https://doi.org/10.1007/s10758-026-09970-x

Methodology & findings

Study design

Mixed-methods study with 2 × 2 factorial design intervention.

Sample

N = 288, 9 groups

Primary method

ANCOVA (Analysis of Covariance) for quantitative data analysis. Thematic analysis for qualitative data from reflective journals and interviews.

Main result

The study found that "reflection type significantly influenced identified regulation, with group reflection fostering greater internalization of academic goals" and that "AI utilization style significantly impacted all four regulation types (external, introjected, identified, and intrinsic), indicating that ChatGPT as a writing partner can both heighten performance pressure and foster deeper engagement." Additionally, "Students' engagement evolved from an initial over-reliance on ChatGPT to more strategic use over time, a development linked to improved prompt literacy and the regulatory role of reflection."

Reports effect sizes.

Research paradigm

mixed_methods (quantitative + qualitative)

Author conclusions

The authors conclude that "These findings highlight the need for teacher guidance and metacognitive scaffolds in AI-supported tasks, demonstrating how emerging technologies interact with motivation and self-regulation in complex educational settings." They also note that "While students valued ChatGPT for its speed, they expressed distrust in its accuracy and a strong demand for structured AI literacy education."

Risk of bias

Selection bias: Vietnamese undergraduates only (limited generalizability); Attrition: Final matched analytic sample was n=288 (unclear if this represents attrition from initial enrollment); Self-report bias: SRQ-A is a self-report questionnaire; Confounders: Not explicitly discussed in abstract; Selection bias: Participants were Vietnamese undergraduates; generalizability to other populations may be limited; Attrition: Study reports 'final matched analytic sample: n = 288', suggesting some participant loss; Confounders: No explicit mention of controlling for prior AI experience, language proficiency, or academic ability; Intervention fidelity: No mention of blinding or standardization checks; Self-report bias: Primary outcome measure (SRQ-A) relies on self-reported questionnaire data; Selection bias: Matched analytic sample may differ from initial recruitment; Attrition: Study mentions 'final matched analytic sample' suggesting some participant loss; Confounders: Multiple intervention conditions (AI utilization style × reflection type) with potential interaction effects

Open questions raised

The study identifies the need for teacher guidance and metacognitive scaffolds in AI-supported tasks. Future research directions include exploring the interaction between reflection type and AI utilization style more deeply, and developing structured AI literacy education programs.
The paper identifies needs for: (1) teacher guidance in AI-supported learning, (2) metacognitive scaffolds, (3) structured AI literacy education, and (4) further research on how emerging technologies interact with motivation and self-regulation
The authors identify the need for teacher guidance in AI-supported tasks, structured AI literacy education programs, and further investigation into how reflection and AI use interact to support self-regulation in educational settings.

Data: not_statedCode: not_statedExtracted from: pdfAgreement 68%

Explore related topics

AI Literacy Training Human-AI Collaboration Mixed Methods