Building an identifiable latent class model with covariate effects on underlying and measured variables

Guan Hua Huang, Karen J Bandeen Roche

Research output: Contribution to journalArticle

Abstract

In recent years, latent class models have proven useful for analyzing relationships between measured multiple indicators and covariates of interest. Such models summarize shared features of the multiple indicators as an underlying categorical variable, and the indicators' substantive associations with predictors are built directly and indirectly in unique model parameters. In this paper, we provide a detailed study on the theory and application of building models that allow mediated relationships between primary predictors and latent class membership, but that also allow direct effects of secondary covariates on the indicators themselves. Theory for model identification is developed. We detail an Expectation-Maximization algorithm for parameter estimation, standard error calculation, and convergent properties. Comparison of the proposed model with models underlying existing latent class modeling software is provided. A detailed analysis of how visual impairments affect older persons' functioning requiring distance vision is used for illustration.

Original languageEnglish (US)
Pages (from-to)5-32
Number of pages28
JournalPsychometrika
Volume69
Issue number1
StatePublished - Mar 2004

Fingerprint

Latent Class Model
Vision Disorders
Covariates
Software
Latent Class
Predictors
Visual Impairment
Categorical variable
Direct Effect
Model Identification
Expectation-maximization Algorithm
Standard error
Model
class membership
Parameter Estimation
visual impairment
Person
Modeling
Parameter estimation
Identification (control systems)

Keywords

  • EM algorithm
  • Finite mixture model
  • Identifiability
  • Multiple discrete indicators
  • Visual functioning

ASJC Scopus subject areas

  • Mathematics (miscellaneous)
  • Psychology(all)
  • Psychology (miscellaneous)
  • Social Sciences (miscellaneous)

Cite this

Building an identifiable latent class model with covariate effects on underlying and measured variables. / Huang, Guan Hua; Bandeen Roche, Karen J.

In: Psychometrika, Vol. 69, No. 1, 03.2004, p. 5-32.

Research output: Contribution to journalArticle

@article{7d3edef5463c4d228d782c52ca53144b,
title = "Building an identifiable latent class model with covariate effects on underlying and measured variables",
abstract = "In recent years, latent class models have proven useful for analyzing relationships between measured multiple indicators and covariates of interest. Such models summarize shared features of the multiple indicators as an underlying categorical variable, and the indicators' substantive associations with predictors are built directly and indirectly in unique model parameters. In this paper, we provide a detailed study on the theory and application of building models that allow mediated relationships between primary predictors and latent class membership, but that also allow direct effects of secondary covariates on the indicators themselves. Theory for model identification is developed. We detail an Expectation-Maximization algorithm for parameter estimation, standard error calculation, and convergent properties. Comparison of the proposed model with models underlying existing latent class modeling software is provided. A detailed analysis of how visual impairments affect older persons' functioning requiring distance vision is used for illustration.",
keywords = "EM algorithm, Finite mixture model, Identifiability, Multiple discrete indicators, Visual functioning",
author = "Huang, {Guan Hua} and {Bandeen Roche}, {Karen J}",
year = "2004",
month = "3",
language = "English (US)",
volume = "69",
pages = "5--32",
journal = "Psychometrika",
issn = "0033-3123",
publisher = "Springer New York",
number = "1",

}

TY - JOUR

T1 - Building an identifiable latent class model with covariate effects on underlying and measured variables

AU - Huang, Guan Hua

AU - Bandeen Roche, Karen J

PY - 2004/3

Y1 - 2004/3

N2 - In recent years, latent class models have proven useful for analyzing relationships between measured multiple indicators and covariates of interest. Such models summarize shared features of the multiple indicators as an underlying categorical variable, and the indicators' substantive associations with predictors are built directly and indirectly in unique model parameters. In this paper, we provide a detailed study on the theory and application of building models that allow mediated relationships between primary predictors and latent class membership, but that also allow direct effects of secondary covariates on the indicators themselves. Theory for model identification is developed. We detail an Expectation-Maximization algorithm for parameter estimation, standard error calculation, and convergent properties. Comparison of the proposed model with models underlying existing latent class modeling software is provided. A detailed analysis of how visual impairments affect older persons' functioning requiring distance vision is used for illustration.

AB - In recent years, latent class models have proven useful for analyzing relationships between measured multiple indicators and covariates of interest. Such models summarize shared features of the multiple indicators as an underlying categorical variable, and the indicators' substantive associations with predictors are built directly and indirectly in unique model parameters. In this paper, we provide a detailed study on the theory and application of building models that allow mediated relationships between primary predictors and latent class membership, but that also allow direct effects of secondary covariates on the indicators themselves. Theory for model identification is developed. We detail an Expectation-Maximization algorithm for parameter estimation, standard error calculation, and convergent properties. Comparison of the proposed model with models underlying existing latent class modeling software is provided. A detailed analysis of how visual impairments affect older persons' functioning requiring distance vision is used for illustration.

KW - EM algorithm

KW - Finite mixture model

KW - Identifiability

KW - Multiple discrete indicators

KW - Visual functioning

UR - http://www.scopus.com/inward/record.url?scp=7244221851&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=7244221851&partnerID=8YFLogxK

M3 - Article

AN - SCOPUS:7244221851

VL - 69

SP - 5

EP - 32

JO - Psychometrika

JF - Psychometrika

SN - 0033-3123

IS - 1

ER -