|
Activity Number:
|
237
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Tuesday, August 5, 2003 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Social Statistics Section
|
| Abstract - #301208 |
|
Title:
|
Bayesian Estimation in Polytomous Logistic Regression Models
|
|
Author(s):
|
Hariharan Swaminathan*+ and Jane Rogers
|
|
Companies:
|
Educational Testing Service and Educational Testing Service
|
|
Address:
|
Rosedale Rd., MS 03T, Princeton, NJ, 08541-0001,
|
|
Keywords:
|
|
|
Abstract:
|
The simultaneous determination of regression functions in several groups has important applications in the areas of prediction and validation. Procedures for estimating and comparing regression functions when the response variable is continuous have been developed and widely used in validation studies. In the multi-group case, Bayes and empirical Bayes approaches have been demonstrated to improve estimation and inference over the classical maximum likelihood approach. The situation is different when the response variable is dichotomous or polytomous. Empirical Bayes approaches, when the data structure is hierarchical in nature, are currently available. However, Bayes approaches for dichotomous and polytomous response variables do not seem to be available currently. Bayes and empirical Bayes are developed in this paper for comparing regression functions across groups when the response variable is polytomous. These procedures are extended to the analysis of hierarchical data. It is shown that Bayes procedures offer considerable advantage over empirical Bayes procedures. The procedures developed are applied to the analysis of large-scale assessment data.
|
- The address information is for the authors that have a + after their name.
- Authors who are presenting talks have a * after their name.
Back to the full JSM 2003 program |
