Files in this item

FilesDescriptionFormat

application/pdf

application/pdfKANG-DISSERTATION-2016.pdf (2MB)
(no description provided)PDF

Description

Title:Likelihood estimation for jointly analyzing item responses and response times
Author(s):Kang, Hyeon-Ah
Director of Research:Chang, Hua-Hua
Doctoral Committee Chair(s):Chang, Hua-Hua
Doctoral Committee Member(s):Anderson, Carolyn J.; Culpepper, Steven A.; Douglas, Jeffrey A.; Koehn, Hans-Friedrich; Zhang, Jinming
Department / Program:Educational Psychology
Discipline:Educational Psychology
Degree Granting Institution:University of Illinois at Urbana-Champaign
Degree:Ph.D.
Genre:Dissertation
Subject(s):response time
item response theory
Abstract:Response time has become increasingly important for analyzing the relationship between the proficiency and speed of an examinee. In this thesis, statistical estimation procedures are presented for jointly modeling responses and response times in educational and psychological testing. The models under consideration include the three-parameter logistic response model, the lognormal response time model, and the proportional hazards latent trait model. The individual models are conjoined within the hierarchical framework so that parameters in the respective models can be characterized under a unified scheme. The thesis presents estimation methods for each of the combinations of the response and response time models by maximizing the likelihood functions. A series of simulation studies verify that the estimation methods perform appropriately, and the parameters are robustly estimated. The likelihood-based approach provides a practical and efficient alternative to Bayesian estimation procedures, which often comes with high computational intensity and dependence on priors.
Issue Date:2016-07-12
Type:Thesis
URI:http://hdl.handle.net/2142/92803
Rights Information:2016 by Hyeon-Ah Kang. All rights reserved.
Date Available in IDEALS:2016-11-10
Date Deposited:2016-08


This item appears in the following Collection(s)

Item Statistics