Statistical Dynamic Analysis of Complex Problem-Solving Items: Inference, Prediction, and Intervention

Yunxiao Chen

Complex problem-solving (CPS) ability has been recognized as a central skill in this information era. For its measurement, computer-based items have been developed that dynamically simulate CPS scenarios that mimic real-life challenges. In such items, students’ problem-solving processes are collected in computer log files, which not only provide the final outcome (success/failure) of task completion, but also the preceding steps and actions that result in the outcome. Albeit the potential value of CPS process data, they are seldom researched, possibly due to the lack of effective quantitative methods. This project has three interrelated research goals. The first is to develop dynamic regression tools for analyzing CPS process data, which have the potential to greatly benefit educational research on non-cognitive abilities in general and CPS ability in specific. The second is to analyze real data from large-scale international assessments and make statistical inferences, whose results can be used to validate and improve the designs of CPS items. Finally, based on the learned dynamic regression model, an automatic hint system will be designed and evaluated via simulation study that provides real-time hints to help students improve their CPS ability. Such a system may significantly contribute to the personalized and technology-enabled learning of CPS.

About Yunxiao Chen

Yunxiao Chen is an assistant professor in the Department of Statistics at London School of Economics and Political Science. He completed his Ph.D. in Statistics at Columbia University in 2016. Before joining LSE, Dr. Chen was an assistant professor in the Department of Psychology and the Institute for Quantitative Theory and Methods at Emory University. Dr. Chen’s primary scholarly focus is on the development of statistical and computational methods for solving problems in educational and psychological sciences, under three interrelated topics including (1) large-scale item response data analysis, (2) measurement and predictive modeling based on students’ dynamic behavioral data and (3) adaptive designs in educational measurement and learning. His articles have been published in a number of leading journals including Psychometrika, Journal of the American Statistical Association, British Journal of Mathematical and Statistical Psychology, and Applied Psychological Measurement.