Files in this item
|(no description provided)|
|Title:||Some theoretical and applied results concerning item response theory model estimation|
|Doctoral Committee Chair(s):||Stout, William F.|
|Department / Program:||Statistics|
|Degree Granting Institution:||University of Illinois at Urbana-Champaign|
|Abstract:||It has long been part of the Item Response Theory (IRT) folklore that under the usual empirical Bayes unidimensional IRT modeling approach, the posterior distribution of examinee ability given test response is approximately normal for a long test. Under very general non-parametric assumptions, we make this claim rigorous for a broad class of latent models.
An asymptotic theory is developed by William Stout for an ordinal-scale-based ability estimation based procedure. Based on this theory a new joint MLE algorithm--Percentile Joint Maximum Likelihood Estimation (PJMLE) is developed for the estimation of examinee ability and item structure. An extensive simulation study has been conducted to examine the statistical properties of PJMLE and to compare the statistical performance of PJMLE with BILOG and LOGIST, the two most widely used procedures for joint estimation of ability and item structure.
|Rights Information:||Copyright 1992 Chang, Hua-Hua|
|Date Available in IDEALS:||2011-05-07|
|Identifier in Online Catalog:||AAI9236416|