science education

ENA traced NLP-coded science ideas across an adaptive AI dialog

Weiying Li, Hsin-Yi Chang, Allison Bradford, Libby Gerard, Marcia C. Linn

Journal of Science Education and Technology

A study of 196 eighth-grade students combined NLP-generated idea codes with ENA to represent how science explanations changed across two adaptive prompts. Higher knowledge-integration scores were associated with more connections among microscopic and mechanistic ideas, but the one-school design does not isolate a causal dialog effect.

Проверенное резюме статьи пока доступно на английском языке.