Cross-Cultural Analysis of GenAI in Collaborative Learning, 2005-2025

2,382 English-language publications from 2005 to 2025 were analyzed using LDA topic modeling and semantic network analysis. The dataset, created by Shengwu Pan, compares Chinese and international scholarship on Generative AI in Computer-Supported Collaborative Learning (CSCL). It reveals divergent conceptual paradigms between systemic-collectivized and instrumental-individualized approaches.

Use Cases

• Compare topic evolution over time based on the evolutionary topic tracing analysis.
• Analyze sentiment trends in academic discourse on GenAI-supported CSCL based on the sentiment detection feature.
• Map conceptual divergences between Chinese and international research paradigms based on the semantic network analysis.
• Train models to classify academic articles by cultural paradigm based on the described topic modeling results.

Strengths

• Analysis of 2,382 publications provides a substantial corpus for macro-level research.
• Explicit time range from 2005 to 2025 allows for longitudinal analysis.
• Integrated framework combining LDA, semantic networks, and sentiment detection offers multiple analytical dimensions.
• Clear source breakdown: 1,273 international articles and 1,109 Chinese samples.

Limitations

• Row count and specific column-level documentation are unknown, limiting suitability assessment.
• Data may reflect geographic and source bias inherent to the selected Web of Science and CNKI publications.
• The 3.3 MB file size suggests the dataset is relatively small, potentially containing aggregated results rather than raw text.

Provenance

Source

Publications from Web of Science and CNKI, plus policy documents.

Collection Method

Computational text-mining framework applied to collected publications.

Time Range

2005-2025

Freshness

Last updated 2026-04-12 15:10:48

Geography

Comparative analysis of Chinese and international (primarily English-language) scholarship.