Meta-Analysis Data on AI-Assisted Programming and Computational Thinking

19 empirical studies and 19 independent effect sizes were coded for a meta-analysis on AI-assisted programming learning. The dataset includes coded variables like educational level, teaching strategy, and effect-size data, with high inter-coder reliability (Cohen’s kappa = 0.902). It was created by YHP and last updated on 2026-04-27.

Use Cases

• Verify the reported meta-analytic results based on the provided effect-size data and coding.
• Reproduce subgroup analyses based on variables like CT measurement tool and instructional function of AI.
• Inspect coding decisions and the study selection process using the included supporting documentation.
• Conduct sensitivity or additional moderator analyses based on the study-level characteristics.

Strengths

• High inter-coder reliability for study coding (Cohen’s kappa = 0.902) and quality assessment (0.925).
• All included studies were rated as high methodological quality, with scores ranging from 0.818 to 0.917.
• Includes supporting documentation like a coding sheet, PRISMA checklist, and flow diagram for transparency.

Limitations

• Row count is unknown, which may limit suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.

Provenance

Source

Harvard Dataverse, author YHP.

Collection Method

Data extracted and coded from 19 empirical studies by two independent researchers for a meta-analysis.

Time Range

Publication years of the included studies are coded, but the specific range is not stated.

Freshness

Last updated 2026-04-27 16:56:30; freshness should be verified.

Geography

Geographic coverage of the included studies is not specified.