USING ITEM RESPONSE THEORY TO DEVELOP A RAVEN’S MATRICES SHORT FORM FOR PAKISTANI ADOLESCENTS AND YOUNG ADULTS
Open Access
- Author:
- Zhong, Hongxuan
- Graduate Program:
- School Psychology
- Degree:
- Doctor of Philosophy
- Document Type:
- Dissertation
- Date of Defense:
- February 22, 2019
- Committee Members:
- James C. DiPerna, Dissertation Advisor/Co-Advisor
James C. DiPerna, Committee Chair/Co-Chair
Barbara A. Schaefer, Committee Member
Pui-Wa Lei, Committee Member
Pamela M. Cole, Outside Member
Barbara A. Schaefer, Committee Chair/Co-Chair - Keywords:
- Item response theory
Measurement efficiency
Short form development
Raven's Matrices
Adolescents and young adults - Abstract:
- The Raven's Matrices are a group of non-verbal tests designed to measure eductive ability (Raven, Raven, & Court, 1998a). The length of the 72-item Raven’s Combined Matrices (RCM) can limit its application in large-scale research studies, as well as potentially cause physical fatigue and/or emotional distress for test takers. Although several research teams have created short forms for different versions of the Raven’s Matrices tests (Arthur & Day, 1994; Bilker et al., 2012; Bors & Stokes, 1998; Sefcek, Miller, & Figueredo, 2016; Wytek, Opgenoorth, & Presslich, 1984), few have used modern test theories to do so. In addition, no short forms have been developed for use in Pakistan to date. As such, the purpose of this study was to use Item Response Theory to develop a RCM short form for potential use with adolescents and young adults in Pakistan. Data were drawn from a longitudinal surveillance follow-up study conducted in Pakistan (N = 1,405). A split-sample approach was used for parameter estimation and validation, along with cross-validation to verify results. Typically developing adolescents and young adults were of interest in this study. As such, items that provided most information around the middle range of the ability continuum were selected to construct the short form. The resulting 10-item RCM short provides similar levels of test information to the RCM long form, maintains the maximum amount of information in the middle of the ability range, demonstrates acceptable reliability for research purposes, and is strongly correlated with the RCM long form. Results, however, also indicated instability in parameter estimation to a certain degree. As such, replication and additional psychometric studies are essential prior to any use of the RCM short form in research or practice.