Ears for Exams: AI-Generated Revision Podcasts and Medical Student Performance in India

Replication data for a mixed-methods study evaluating AI-generated revision podcasts at two Indian medical colleges. The repository contains anonymized longitudinal test scores and qualitative focus group transcripts from the 'Ears for Exams' project. It includes pre-test, immediate post-test, and 30-day delayed post-test data for 60-item MCQ assessments, linked by participant codes.

Use Cases

• Analyze the impact of AI-generated audio on learning retention based on longitudinal test scores.
• Evaluate self-reported usability and adherence to the intervention from post-test survey data.
• Conduct a framework analysis of student perceptions based on cleaned focus group transcripts.
• Correlate demographic factors with learning outcomes using the anonymized baseline data.
• Map assessment items to cognitive domains using the provided Bloom's Taxonomy blueprint.

Strengths

• Data supports paired longitudinal analysis via Self-Generated Identification Codes (SGIC).
• Includes both quantitative scores and qualitative transcripts for a mixed-methods evaluation.
• Assessment instruments are mapped to Bloom's Taxonomy (Recall, Application, Analysis).
• All direct identifiers have been removed to maintain student anonymity.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• The dataset is small in scale at 838.0 KB, indicating a limited sample size.

Provenance

Source

figshare, author Urvish Joshi.

Collection Method

Collected via a mixed-methods explanatory sequential design in two Indian medical colleges.

Freshness

Last updated 2026-05-21 13:15:16; freshness should be verified.

Geography

India (two medical colleges).