Developing exploratory cognitive diagnosis models for educational research

Principal Investigator: Steven Culpepper | Department of Statistics | University of Illinois at Urbana-Champaign

Principal Investigator(s):

Jeffrey Douglas | Department of Statistics | University of Illinois at Urbana-Champaign

Summary

Cognitive diagnosis models (CDMs) are a popular psychometric framework for designing instructionally relevant assessments. The CDM framework has the potential to promote student learning by providing educators and researchers with detailed diagnostic information about students’ mastery of skills in a given content area. However, the application of CDMs requires robust cognitive theory that clearly specifies the necessary skills for success on educational tasks. For CDMs, cognitive theory is recorded in the Q matrix. The unavailability of cognitive theory to specify Q limits widespread application of CDMs in educational research. Furthermore, relying upon imprecise theory (i.e., an incorrect Q matrix) has been shown to yield inaccurate student skill diagnoses, which could impact the decisions of educators and researchers. A fundamental statistical problem for educational research and practice is the estimation of Q. This project develops statistical methodology to estimate Q for several CDMs. The new statistical procedures are disseminated to applied researchers as an open source R package. The research from this project will improve education by providing researchers with new tools to evaluate cognitive theory. This project serves as a critical next step for realizing the promise of CDMs as a methodology for supporting student learning with instructionally relevant assessments.,

Grant Type:

Small Grant

Grant Amount:

$50,000.00

Year:

2016

Research Area:

Field-Initiated

Topic / Subject:

Methodology, Cognition, Assessment

Methods / Approach:

Psychometric analyses, Other - Quantitative, Latent variable models

Disciplinary Perspective:

Statistics