Abstract
Detection of students’ ability levels is one of the common aims in educational studies. Cognitive Diagnosis Modeling approach has been used recently for the purpose of ability level detection by defined Q-matrices. This paper aims to use Cognitive Diagnosis Modeling (CDM) in order to investigate the definition of a Q-matrix across the cognitive skills of different years and countries in TIMSS. There is a subjective way in defining Q-matrices; for this purpose, an application of building Q-matrices under specific CDMs, from a set of expert proposed attributes is examined. The proposed attributes are used to build Q-matrices for TIMSS mathematics questions across its cycles, and across different nations.
Highlights
Nowadays International assessments have become more important in recent decades as a consequence of globalization
The Rule Space Model (RSM; Tatsuoka, 1983) applied to Trends in International Mathematics and Science Study (TIMSS) 1999 mathematics items, and 30 attributes are defined to show the gap across gender and ethnicities (Birenbaum, Tatsuoka, & Yamada, 2004; Birenbaum, Nasser, & Tatsuoka, 2007)
The aims of this paper are (1) to fit the DINA and generalized DINA (G-DINA; de la Torre, 2011) models to TIMSS 2011 fourth grade mathematics items using the attributes that are proposed for TIMSS 2007 by Lee, Park, and Taylan (2011), and (2) to compare different Q-matrices and validate the most appropriate Q-matrix for the TIMSS 2011 data based on the G-DINA model fit results
Summary
Nowadays International assessments have become more important in recent decades as a consequence of globalization. In addition to the comparison of achievement scores, students’ mastery level of specific attributes has been evaluated by using the TIMSS (Lee, Park, Taylan, 2011). The Rule Space Model (RSM; Tatsuoka, 1983) applied to TIMSS 1999 mathematics items, and 30 attributes are defined to show the gap across gender and ethnicities (Birenbaum, Tatsuoka, & Yamada, 2004; Birenbaum, Nasser, & Tatsuoka, 2007). CDM is a latent variable model where skills are defined as attributes, and represented by the binary vector α to assess student mastery and nonmastery of the skills (de la Torre, 2011). Q-matrix is a cognitive design matrix that identifies the cognitive specification for each item (de la Torre, 2008)
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
