The Multidimensional Generalized Graded Unfolding Model for Assessment of Change across Repeated Measures
The Multidimensional Generalized Graded Unfolding Model for Assessment of Change across Repeated Measures
Loading...
Files
Publication or External Link
Date
2008-05-13
Authors
Cui, Weiwei
Advisor
Roberts, James S
Dayton, Chan M
Dayton, Chan M
Citation
DRUM DOI
Abstract
A multidimensional extension of the generalized graded unfolding model for repeated measures (GGUM-RM) is introduced and applied to analyze attitude change across time using responses collected by a Thurstone or Likert questionnaire. The model conceptualizes the change across time as separate latent variables and provides direct estimates of both individual and group change while accounting for the dependency among latent variables. The parameters and hyperparameters of GGUM-RM are estimated by fully Bayesian estimation method via WinBUGS. The accuracy of the estimation procedure is demonstrated by a simulation study, and the application of the GGUM-RM is illustrated by the analysis of attitude change toward abortion among college students.