A Monte Carlo evaluation of three methods to detect local dependence in binary data latent class models

D.L. Oberski, G.H. van Kollenburg, J.K. Vermunt

Research output: Contribution to journalArticleScientificpeer-review

Abstract

Binary data latent class analysis is a form of model-based clustering applied in a wide range of fields. A central assumption of this model is that of conditional independence of responses given latent class membership, often referred to as the “local independence” assumption. The results of latent class analysis may be severely biased when this crucial assumption is violated; investigating the degree to which bivariate relationships between observed variables fit this hypothesis therefore provides vital information. This article evaluates three methods of doing so. The first is the commonly applied method of referring the so-called “bivariate residuals” to a Chi-square distribution. We also introduce two alternative methods that are novel to the investigation of local dependence in latent class analysis: bootstrapping the bivariate residuals, and the asymptotic score test or “modification index”. Our Monte Carlo simulation indicates that the latter two methods perform adequately, while the first method does not perform as intended.
Keywords: Conditional dependence, Latent variable models, Score test, Lagrange multiplier test, Modification index, Bivariate residuals
Original languageEnglish
Pages (from-to)267-279
JournalAdvances in Data Analysis and Classification
Volume7
Issue number3
DOIs
Publication statusPublished - 2013

Fingerprint

Latent Class Model
Binary Data
Data Model
Latent Class Analysis
Evaluation
Score Test
Lagrange multipliers
Local Independence
Lagrange multiplier Test
Asymptotic Test
Model-based Clustering
Latent Class
Chi-square Distribution
Latent Variable Models
Conditional Independence
Bootstrapping
Biased
Monte Carlo Simulation
Evaluate
Alternatives

Cite this

@article{1361734ebc52434a90c75758eec6f64c,
title = "A Monte Carlo evaluation of three methods to detect local dependence in binary data latent class models",
abstract = "Binary data latent class analysis is a form of model-based clustering applied in a wide range of fields. A central assumption of this model is that of conditional independence of responses given latent class membership, often referred to as the “local independence” assumption. The results of latent class analysis may be severely biased when this crucial assumption is violated; investigating the degree to which bivariate relationships between observed variables fit this hypothesis therefore provides vital information. This article evaluates three methods of doing so. The first is the commonly applied method of referring the so-called “bivariate residuals” to a Chi-square distribution. We also introduce two alternative methods that are novel to the investigation of local dependence in latent class analysis: bootstrapping the bivariate residuals, and the asymptotic score test or “modification index”. Our Monte Carlo simulation indicates that the latter two methods perform adequately, while the first method does not perform as intended.Keywords: Conditional dependence, Latent variable models, Score test, Lagrange multiplier test, Modification index, Bivariate residuals",
author = "D.L. Oberski and {van Kollenburg}, G.H. and J.K. Vermunt",
year = "2013",
doi = "10.1007/s11634-013-0146-2",
language = "English",
volume = "7",
pages = "267--279",
journal = "Advances in Data Analysis and Classification",
issn = "1862-5347",
publisher = "Springer Verlag",
number = "3",

}

A Monte Carlo evaluation of three methods to detect local dependence in binary data latent class models. / Oberski, D.L.; van Kollenburg, G.H.; Vermunt, J.K.

In: Advances in Data Analysis and Classification, Vol. 7, No. 3, 2013, p. 267-279.

Research output: Contribution to journalArticleScientificpeer-review

TY - JOUR

T1 - A Monte Carlo evaluation of three methods to detect local dependence in binary data latent class models

AU - Oberski, D.L.

AU - van Kollenburg, G.H.

AU - Vermunt, J.K.

PY - 2013

Y1 - 2013

N2 - Binary data latent class analysis is a form of model-based clustering applied in a wide range of fields. A central assumption of this model is that of conditional independence of responses given latent class membership, often referred to as the “local independence” assumption. The results of latent class analysis may be severely biased when this crucial assumption is violated; investigating the degree to which bivariate relationships between observed variables fit this hypothesis therefore provides vital information. This article evaluates three methods of doing so. The first is the commonly applied method of referring the so-called “bivariate residuals” to a Chi-square distribution. We also introduce two alternative methods that are novel to the investigation of local dependence in latent class analysis: bootstrapping the bivariate residuals, and the asymptotic score test or “modification index”. Our Monte Carlo simulation indicates that the latter two methods perform adequately, while the first method does not perform as intended.Keywords: Conditional dependence, Latent variable models, Score test, Lagrange multiplier test, Modification index, Bivariate residuals

AB - Binary data latent class analysis is a form of model-based clustering applied in a wide range of fields. A central assumption of this model is that of conditional independence of responses given latent class membership, often referred to as the “local independence” assumption. The results of latent class analysis may be severely biased when this crucial assumption is violated; investigating the degree to which bivariate relationships between observed variables fit this hypothesis therefore provides vital information. This article evaluates three methods of doing so. The first is the commonly applied method of referring the so-called “bivariate residuals” to a Chi-square distribution. We also introduce two alternative methods that are novel to the investigation of local dependence in latent class analysis: bootstrapping the bivariate residuals, and the asymptotic score test or “modification index”. Our Monte Carlo simulation indicates that the latter two methods perform adequately, while the first method does not perform as intended.Keywords: Conditional dependence, Latent variable models, Score test, Lagrange multiplier test, Modification index, Bivariate residuals

U2 - 10.1007/s11634-013-0146-2

DO - 10.1007/s11634-013-0146-2

M3 - Article

VL - 7

SP - 267

EP - 279

JO - Advances in Data Analysis and Classification

JF - Advances in Data Analysis and Classification

SN - 1862-5347

IS - 3

ER -