AI evidence extraction

Interpreting Interaction Patterns and Cognitive Strategies in LLM-Supported Exploratory Learning: A Mixed-Methods Analysis Using the DOK Framework

Yiming Luo, T. Liu, Patrick Cheong‐Iao Pang, Dana McKay, Shanton Chang, George Buchanan · Information · 2026

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

5/10

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/10.3390/info17030288

Methodology & findings

Study design

Observational comparative study systematically investigating exploratory learning strategies of 46 students across two regions in Asia, classifying 25 distinct strategies using the Depth of Knowledge (DOK) model and comparing strategy usage between high and low-performing student subgroups..

Sample

N = 46, 4 groups

Primary method

Mixed-methods analysis incorporating quantitative comparison between student subgroups and qualitative classification using the Depth of Knowledge (DOK) framework; specific statistical tests not detailed in abstract

Main result

The study found that "a declining trend in the utilization of EL strategies across ascending cognitive stages" and "high AWP students employed EL strategies more frequently than their peers, with ten EL strategies exhibiting significant between-group differences." Additionally, among students with different AI experience levels, only a few EL strategies usage and cognitive stages showed significant differences.

Reports effect sizes.

Research paradigm

mixed-methods (quantitative and qualitative)

Author conclusions

The authors conclude that "these insights can help educators and LLM interface designers develop targeted exploratory learning assistance for different types of students and help them build high-level metacognitive processes for effective human–computer interaction."

Risk of bias

Selection bias: participants from two different regions of Asia may not be representative of broader student populations; Confounding variables: AI experience differences across regions not fully controlled; Observer bias: classification of strategies by researchers may introduce subjectivity; Self-selection bias: students who volunteer for LLM-supported learning studies may differ systematically from general population; Selection bias: Limited to 46 students in two Asian regions, potentially not representative of broader populations; Confounding variables: Different AI experience levels among students not fully controlled; Observer bias: Observational study design may introduce bias in strategy classification and interpretation; Selection bias: convenience sampling across two Asian regions may not be representative; Confounding variables: student prior AI experience not fully controlled; Observational design: no random assignment to conditions; Sample size: N=46 limits generalizability

Limitations

The paper states that the study was conducted on "46 students in two different regions of Asia" which represents a geographically limited sample
While specific limitations are not explicitly detailed in the abstract provided, the authors acknowledge that "the underlying mechanisms driving these disparities, particularly in how students interact with LLMs, remain underexplored."

Open questions raised

The study addresses underexplored mechanisms driving disparities in student engagement and effectiveness with LLM-supported learning environments, particularly regarding how students interact with LLMs during exploratory learning.
The paper identifies that "the underlying mechanisms driving these disparities, particularly in how students interact with LLMs, remain underexplored" and suggests future research should focus on understanding how to develop targeted exploratory learning assistance and metacognitive processes for different student types in LLM-supported environments.
The underlying mechanisms driving disparities in student engagement and effectiveness when interacting with LLM-supported learning environments, particularly in how students interact with LLMs, remain underexplored.

Data: not_statedCode: not_statedExtracted from: pdfAgreement 61%

Explore related topics

Human-AI Collaboration Mixed Methods