|
Activity Number:
|
541
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Thursday, August 2, 2007 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Section on Bayesian Statistical Science
|
| Abstract - #309511 |
|
Title:
|
Comparing Fits in Nonparametric Bayes Item Response Models
|
|
Author(s):
|
Kristin Duncan*+
|
|
Companies:
|
San Diego State University
|
|
Address:
|
519 W Laurel St Unit 4, San Diego, CA, 92101,
|
|
Keywords:
|
Nonparametric Bayes ; Item Response Theory
|
|
Abstract:
|
Item response models are used to study the relationship between a latent trait or ability level and individuals' responses to items on a test or questionnaire designed to measure that trait. The item characteristic curve (ICC) gives the probability of a correct response as a function of ability. Dirichlet process priors provide a flexible means of modeling the ICC. Here we compare fits from models using two- and three-parameter logistic curves as the prior means for the Dirichlet process and also study the impact of the mass parameter of the Dirichlet process on the fitted curves. Comparisons are made with simulated data and with a data set from an introductory statistics final exam.
|
