Cross-National Cognitive Diagnostic Profiles in PISA 2015 Computer-Based Mathematics Assessment: A Joint Analysis of Attribute Mastery and Response Times
Abstract This study examined cross-national cognitive diagnostic profiles in the PISA 2015 computer-based mathematics assessment by integrating attribute mastery and response-time information. Eleven education systems were purposively selected to reflect variation in geographic regions, educational contexts, and levels of economic development, and all students included in the analysis responded to the same set of nine test items. Specifically, this study compared students’ attribute mastery probabilities across four PISA 2015 mathematics content domains, i.e., Change and Relationships, Space and Shape, Quantity, and Uncertainty and Data, examined knowledge-state distributions and cross-sectional mastery-order patterns, and evaluated response-time-informed item diagnostics, including slipping, guessing, and discrimination parameters. Using a joint cognitive diagnosis and response time modeling framework, we estimate students’ mastery probabilities and knowledge-state distributions, identify cross-sectional mastery-order patterns, and evaluate item-level diagnostic characteristics. The findings indicate relatively high mastery of Change and Relationships in several education systems, substantial cross-national variation in Quantity and Uncertainty and Data, and a common cross-sectional mastery-order pattern in which Change and Relationships tends to appear earlier than Space and Shape, Quantity, and Uncertainty and Data in the distribution of knowledge states. Response-time-informed item diagnostics further reveal differences in slipping, guessing, and discrimination across items and education systems, suggesting that similar response scores may reflect different response processes and item characteristics. These results demonstrate how cognitive diagnosis and response-time modeling can extend the interpretation of PISA mathematics data beyond overall performance comparisons and provide diagnostic information for curriculum refinement, targeted instructional support, and assessment development.
Authors
- Guanghui Cheng (ORCID: https://orcid.org/0000-0003-1227-6306)
- Zhaoyuan Zhang (ORCID: https://orcid.org/0009-0002-4692-4607)
- Jing Lu
- Jiwei Zhang
Institutions
- Northeast Normal University (CN)
- Yili Normal University (CN)
- Guangzhou University (CN)
Publication Details
- Journal
- Research Square
- Published
- 2026-11-04
- DOI
- https://doi.org/10.21203/rs.3.rs-10525755/v1
- Primary Topic
- Psychometric Methodologies and Testing
- Type
- preprint