Georgia Institute of Technology
Vertical Equating with Longitudinal Data: Unidimensional and Multidimensional Item Response Theory Models
Abstract
dc:description.abstractVertical equating is one context in which item response theory (IRT) model fit is important to evaluate and the measurement of change in students’ performance between test occasions is a central topic in educational research and assessment. The aim of the work presented here is to examine the performances of several multidimensional (MRMLC and SLTM) and unidimensional (Rasch and 2PL) IRT models on linking item parameters in a longitudinal design, in which there are no common items across the occasions. Both an empirical study with real data and a simulation study were conducted. It was found that the multidimensional IRT models for learning and change yielded more plausible results for equating and they have certain advantages over the standard unidimensional IRT models.
Degree
thesis:*- Level thesis:degree_level
- Masters
- Department dc:contributor.department
- Psychology
- Grantor dc:publisher
- Georgia Institute of Technology
- Year dc:date.issued
- 2020
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Yan, Yan
- Advisor dc:contributor.advisor
-
- Embretson, Susan E.
- Committee members dc:contributor.committeemember
-
- Roberts, James S.
- Hunter, Michael D.
Subjects
dc:subject × 4Rights
- Language dc:language.iso
- en_US
Identifiers
dc:identifier.*- Handle dc:identifier.uri
- http://hdl.handle.net/1853/66403