Formulating a Dimtest-Based Effect Size Measure and a New DIMTEST Statistic and Their Applications
Seo, Min Hee
This item is only available for download by members of the University of Illinois community. Students, faculty, and staff at the U of I may log in with your NetID and password to view the item. If you are trying to access an Illinois-restricted dissertation or thesis, you can request a copy through your library's Inter-Library Loan office or purchase a copy directly from ProQuest.
Permalink
https://hdl.handle.net/2142/80077
Description
Title
Formulating a Dimtest-Based Effect Size Measure and a New DIMTEST Statistic and Their Applications
Author(s)
Seo, Min Hee
Issue Date
2008
Doctoral Committee Chair(s)
Chang, Hua-Hua
Department of Study
Educational Psychology
Discipline
Educational Psychology
Degree Granting Institution
University of Illinois at Urbana-Champaign
Degree Name
Ph.D.
Degree Level
Dissertation
Keyword(s)
Education, Tests and Measurements
Language
eng
Abstract
This dissertation proposes an interpretable effect size measure for use with DIMTEST, a widely used method for testing the hypothesis of test unidimensionality as represented by local item independence. There are six chapters in this dissertation. Chapter 1 addresses the importance of assessing unidimensionality and provides a review of the literature and methods. Chapter 2 reviews basic concepts of DIMTEST and the DIMTEST procedures for a statistical test of the null hypothesis. In Chapter 3 a DIMTET-based Effect Size Measure (DESM) is proposed. This chapter also evaluates the efficacy of DESM and the amount of bias in the DESM estimator by comparing it with the DESM population parameter. The results indicate that the DESM estimates converge to DESM parameters as test length and sample size increase. Even though DESM exhibits statistical bias with shorter tests, the amount of the bias decreases as test length and PT size increase. Chapter 4 of the dissertation develops a new DIMTEST statistic (NEWDIM) that yields improved Type I error rates relative to the original DIMTEST T statistic. Furthermore, this chapter evaluates the performances of DESM and NEWDIM in terms of Type I error and power. DESM used in tandem with DIMTEST results in better Type I error rates compared to using DIMTEST alone while having approximately equal power for cases of moderate multidimensionality. To better understand the effectiveness of DESM, NEWDIM, and DIMTEST, Chapter 5 focuses on short tests where DIMTEST has shown unexpectedly high Type I error rates. This section simulates short tests with total items less than 11 and compares the performance of DESM, NEWDIM, and DIMTEST under both unidimensional and multidimensional conditions. As expected, DIMTEST has Type I error inflation; however, using either DIMTEST or NEWDIM in tandem with DESM reduces greatly Type I error rates. Finally, the last chapter of the dissertation applies DESM and NEWDIM to real data in order to gain insight about their performances in practice.
Use this login method if you
don't
have an
@illinois.edu
email address.
(Oops, I do have one)
IDEALS migrated to a new platform on June 23, 2022. If you created
your account prior to this date, you will have to reset your password
using the forgot-password link below.
