Applying machine learning models in multi-institutional studies can generate bias
Open Access
- 12 September 2024
- conference paper
- Published by American Association of Physics Teachers (AAPT)
- p. 144-149
Abstract
There is increasing interest in deploying machine learning models at scale for multi-institutional studies in physics education research. Here we investigate the efficacy of applying machine learning models to institutions outside of their training set, using natural language processing to code open-ended survey responses. We find that, in general, changing institutional contexts can affect machine learning estimates of code frequencies: either previously documented sources of uncertainty increase in magnitude, new unknown sources of uncertainty emerge, or both. We also find an example where uncertainties do not change between the institution used in the training data and an institution not in the training data. Results suggest that attention to uncertainty is critical, especially when making measurements of student writing across multi-institutional data sets.Keywords
This publication has 13 references indexed in Scilit:
- What topics of peer interactions correlate with student performance in physics courses?European Journal of Physics, 2024
- Evaluating IBM’s Watson natural language processing artificial intelligence as a short-answer categorization tool for physics education researchPhysical Review Physics Education Research, 2024
- Method to assess the trustworthiness of machine coding at scalePhysical Review Physics Education Research, 2024
- Understanding interaction network formation across instructional contexts in remote physics coursesPhysical Review Physics Education Research, 2022
- Machine learning for automated content analysis: characteristics of training data impact reliability2022
- Classification of open-ended responses to a research-based assessment using natural language processingPhysical Review Physics Education Research, 2022
- Automated Analysis of Reflection in Writing: Validating Machine Learning ApproachesInternational Journal of Artificial Intelligence in Education, 2019
- Educational commitment and social networking: The power of informal networksPhysical Review Physics Education Research, 2018
- Automated analysis of short responses in an interactive synthetic tutoring system for introductory physicsPhysical Review Physics Education Research, 2016
- Improving class probability estimates for imbalanced dataKnowledge and Information Systems, 2013