|
Activity Number:
|
243
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Tuesday, August 13, 2002 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Nonparametric Statistics*
|
| Abstract - #301572 |
|
Title:
|
A Semiparametric Approach to Analyzing Repeated Measures from a Finite Mixture Model
|
|
Author(s):
|
Ryan Elmore*+ and Tom Hettmansperger and Hoben Thomas
|
|
Affiliation(s):
|
Pennsylvania State University and Pennsylvania State University and Pennsylvania State University
|
|
Address:
|
326 Thomas Building, University Park, Pennsylvania, 16802, USA
|
|
Keywords:
|
semiparametric ; mixture model ; beta-binomial ; correlation ; asymptotics
|
|
Abstract:
|
This research is motivated by a psychology experiment in which children are asked to complete a series of tests. Theory suggests that each child will use one of several strategies to complete the task; however, it is not known which strategy they will employ. Finite mixture models are an attractive approach to modeling populations which are composed of distinct subpopulations, or components. In this paper, we will present a semiparametric approach to the analysis of repeated measures from a finite mixture model. A model-free approach to inference for finite mixture models was initially developed in Hettmansperger and Thomas (2000). We will extend their model in order to account for the within-subject correlation that is inherently present in the repeated measures context. Several theoretical results are presented in addition to an example from a psychological experiment.
|
