This item is only available for download by members of the University of Illinois community. Students, faculty, and staff at the U of I may log in with your NetID and password to view the item. If you are trying to access an Illinois-restricted dissertation or thesis, you can request a copy through your library's Inter-Library Loan office or purchase a copy directly from ProQuest.
Permalink
https://hdl.handle.net/2142/69173
Description
Title
A Model That Combines IRT and Latent Class Models
Author(s)
Yamamoto, Kentaro
Issue Date
1987
Doctoral Committee Chair(s)
Tatsuoka, Kikumi K.
Department of Study
Education
Discipline
Education
Degree Granting Institution
University of Illinois at Urbana-Champaign
Degree Name
Ph.D.
Degree Level
Dissertation
Keyword(s)
Education, Tests and Measurements
Abstract
A measurement model called Hybrid model is developed to overcome the weaknesses of the widely used Item Response Theory (IRT) and Latent Class models by combining them both. The weakness of IRT is that it is unable to express cognitive structure whether that is rule of operations or particular misunderstanding, or to give descriptive information of knowledge possessed by examinees. The Latent Class models require severe constraints, namely possession of complete knowledge of latent classes, and has the indeterminacy of number of ordered classes. The Hybrid model is most applicable in the so-called achievement testing, diagnostic testing, and mastery testing environment, and it is applicable even in some "ability tests" as long as a unique understanding or misunderstanding of the examinee can be characterized by a particular response pattern.
Within the field of educational testing and evaluation, item response models can be classified into two types, one is "continuous" because of the propensity to answer in a certain way on an item changes accordingly as continuous ability value changes, and the other is "discrete" because a set of conditional probabilities are assigned to a limited number of identifiable ability states which may be ordered or not ordered. The hybrid model combines the two continuous and discrete measurement models, namely the two-parameter Logistic model of IRT and the Latent Class models respectively. The marginal maximum likelihood estimation method was used to develop the algorithm for a computer program which estimates parameters of IRT, mixture proportions and vectors of conditional probabilities of correct responses for the latent classes. Several types of latent class models are considered, including both more traditional models as well as the model that uses person characteristic curves. To evaluate the accuracy of the algorithm and also the computer program itself, several simulation data sets were generated. The comparison of estimated parameter values and the original values shows that the model parameters are accurately estimated, in particular, IRT item parameters estimation was found to be robust to the wrong specification of the latent class type.
Use this login method if you
don't
have an
@illinois.edu
email address.
(Oops, I do have one)
IDEALS migrated to a new platform on June 23, 2022. If you created
your account prior to this date, you will have to reset your password
using the forgot-password link below.
