Specification searches in multilevel structural equation modeling: A monte carlo investigation

James L. Peugh, Craig K. Enders

Research output: Contribution to journalArticle

4 Citations (Scopus)

Abstract

Cluster sampling results in response variable variation both among respondents (i.e., within cluster or Level 1) and among clusters (i.e., between-cluster or Level 2). Properly modeling within- and between-cluster variation could be of substantive interest in numerous settings, but applied researchers typically test only within-cluster (i.e., individual difference) theories. Specifying a between-cluster model in the absence of theory requires a specification search in multilevel structural equation modeling. This study examined a variety of within-cluster and between-cluster sample sizes, intraclass correlation coefficients, start models, parameter addition and deletion methods, and Type I error control techniques to identify which combination of start model, parameter addition or deletion method, and Type I error control technique best recovered the population of the between-cluster model. Results indicated that a "saturated" start model, univariate parameter deletion technique, and no Type I error control performed best, but recovered the population between-cluster model in less than 1 in 5 attempts at the largest sample sizes. The accuracy of specification search methods, suggestions for applied researchers, and future research directions are discussed.

Original languageEnglish (US)
Pages (from-to)42-65
Number of pages24
JournalStructural Equation Modeling
Volume17
Issue number1
DOIs
StatePublished - Jan 2010

Fingerprint

Structural Equation Modeling
Specification
Specifications
Type I error
Error Control
Deletion
Sample Size
Structural equation modeling
Model
Intraclass Correlation Coefficient
Cluster Sampling
Individual Differences
Sampling
Search Methods
Univariate

ASJC Scopus subject areas

  • Modeling and Simulation
  • Decision Sciences(all)
  • Economics, Econometrics and Finance(all)
  • Sociology and Political Science

Cite this

Specification searches in multilevel structural equation modeling : A monte carlo investigation. / Peugh, James L.; Enders, Craig K.

In: Structural Equation Modeling, Vol. 17, No. 1, 01.2010, p. 42-65.

Research output: Contribution to journalArticle

@article{acfaa01bc5db43388172f525a021c7cf,
title = "Specification searches in multilevel structural equation modeling: A monte carlo investigation",
abstract = "Cluster sampling results in response variable variation both among respondents (i.e., within cluster or Level 1) and among clusters (i.e., between-cluster or Level 2). Properly modeling within- and between-cluster variation could be of substantive interest in numerous settings, but applied researchers typically test only within-cluster (i.e., individual difference) theories. Specifying a between-cluster model in the absence of theory requires a specification search in multilevel structural equation modeling. This study examined a variety of within-cluster and between-cluster sample sizes, intraclass correlation coefficients, start models, parameter addition and deletion methods, and Type I error control techniques to identify which combination of start model, parameter addition or deletion method, and Type I error control technique best recovered the population of the between-cluster model. Results indicated that a {"}saturated{"} start model, univariate parameter deletion technique, and no Type I error control performed best, but recovered the population between-cluster model in less than 1 in 5 attempts at the largest sample sizes. The accuracy of specification search methods, suggestions for applied researchers, and future research directions are discussed.",
author = "Peugh, {James L.} and Enders, {Craig K.}",
year = "2010",
month = "1",
doi = "10.1080/10705510903438948",
language = "English (US)",
volume = "17",
pages = "42--65",
journal = "Structural Equation Modeling",
issn = "1070-5511",
publisher = "Psychology Press Ltd",
number = "1",

}

TY - JOUR

T1 - Specification searches in multilevel structural equation modeling

T2 - A monte carlo investigation

AU - Peugh, James L.

AU - Enders, Craig K.

PY - 2010/1

Y1 - 2010/1

N2 - Cluster sampling results in response variable variation both among respondents (i.e., within cluster or Level 1) and among clusters (i.e., between-cluster or Level 2). Properly modeling within- and between-cluster variation could be of substantive interest in numerous settings, but applied researchers typically test only within-cluster (i.e., individual difference) theories. Specifying a between-cluster model in the absence of theory requires a specification search in multilevel structural equation modeling. This study examined a variety of within-cluster and between-cluster sample sizes, intraclass correlation coefficients, start models, parameter addition and deletion methods, and Type I error control techniques to identify which combination of start model, parameter addition or deletion method, and Type I error control technique best recovered the population of the between-cluster model. Results indicated that a "saturated" start model, univariate parameter deletion technique, and no Type I error control performed best, but recovered the population between-cluster model in less than 1 in 5 attempts at the largest sample sizes. The accuracy of specification search methods, suggestions for applied researchers, and future research directions are discussed.

AB - Cluster sampling results in response variable variation both among respondents (i.e., within cluster or Level 1) and among clusters (i.e., between-cluster or Level 2). Properly modeling within- and between-cluster variation could be of substantive interest in numerous settings, but applied researchers typically test only within-cluster (i.e., individual difference) theories. Specifying a between-cluster model in the absence of theory requires a specification search in multilevel structural equation modeling. This study examined a variety of within-cluster and between-cluster sample sizes, intraclass correlation coefficients, start models, parameter addition and deletion methods, and Type I error control techniques to identify which combination of start model, parameter addition or deletion method, and Type I error control technique best recovered the population of the between-cluster model. Results indicated that a "saturated" start model, univariate parameter deletion technique, and no Type I error control performed best, but recovered the population between-cluster model in less than 1 in 5 attempts at the largest sample sizes. The accuracy of specification search methods, suggestions for applied researchers, and future research directions are discussed.

UR - http://www.scopus.com/inward/record.url?scp=75149138257&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=75149138257&partnerID=8YFLogxK

U2 - 10.1080/10705510903438948

DO - 10.1080/10705510903438948

M3 - Article

AN - SCOPUS:75149138257

VL - 17

SP - 42

EP - 65

JO - Structural Equation Modeling

JF - Structural Equation Modeling

SN - 1070-5511

IS - 1

ER -