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/2142/82140
Description
Title
Regression Models for Paired Comparisons
Author(s)
Verkuilen, John V.
Issue Date
2007
Doctoral Committee Chair(s)
Budescu, David V.
Department of Study
Psychology
Discipline
Psychology
Degree Granting Institution
University of Illinois at Urbana-Champaign
Degree Name
Ph.D.
Degree Level
Dissertation
Keyword(s)
Psychology, Psychometrics
Language
eng
Abstract
"Paired comparisons are among the most widely used experimental methods in psychology and related fields. First employed in the middle of the 19th Century, many models have been developed over the years to analyze data from such experiments, for instance the Thurstone and Bradley-Terry-Luce models. This dissertation proposes an integrative framework in which the vast majority of extant scaling procedures for paired comparisons can be cast. The framework is based on the Generalized Linear Mixed Model (GLMM) and it characterizes paired comparison models according to four characteristics, taken in order: (1) the response format, (2) the link function, (3) the error distribution, and (4) the mixing distribution. It expresses the scaling problem as a regression fitting in the GLMM, and shows how particular aspects of paired comparisons restrict the choices that can be made among these four model components. All known response formats---binary, discrete ordinal, continuous constant sum and continuous constant product---fit as examples in this framework. Given this, it is easy to consider other models based on the four characteristics. In addition to the framework, models designed to analyze data for graded response formats where subjects are heterogeneous in their response styles are proposed and examined. Finally a simulation study illustrates the robustness of the classic ""Case V"" specification."
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.
