|
Activity Number:
|
312
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, August 5, 2008 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Social Statistics Section
|
| Abstract - #300617 |
|
Title:
|
Hierarchical Nonlinear Profile Analysis of Dichotomous Item Responses
|
|
Author(s):
|
Steven A. Culpepper*+
|
|
Companies:
|
Metropolitan State College of Denver
|
|
Address:
|
Department of Human Services, Denver, CO, 80217-3362,
|
|
Keywords:
|
Hierarchical Nonlinear Models ; Profile Analysis ; Item Response Theory ; Experiment ; Map Comprehension
|
|
Abstract:
|
This study linked hierarchical nonlinear profile analysis (NPA) of dichotomous responses with a larger family of item response models. The NPA method offers several benefits over previous profile analysis methods: 1) NPA relies upon maximum likelihood rather than different dissimilarity measures that produce different results; 2) item and person parameters are computed during the same estimation step with an appropriate distribution for dichotomous variables; 3) the estimation procedure produces standard errors of profile coordinates; and 4) subject-level variables can be included to predictor profile parameters. An application examined experimental differences among 288 subjects' topographic map comprehension. The model produced a measure of overall test performance, in addition to correspondence weights with an item-difficulty profile and item-discrimination profile.
|
