Effects of sample size and nonnormality on the estimation of mediated effects in latent variable models

Research output: Contribution to journalArticle

104 Citations (Scopus)

Abstract

A Monte Carlo approach was used to examine bias in the estimation of indirect effects and their associated standard errors. In the simulation design, (a) sample size, (b) the level of nonnormality characterizing the data, (c) the population values of the model parameters, and (d) the type of estimator were systematically varied. Estimates of model parameters were generally unaffected by either nonnormality or small sample size. Under severely nonnormal conditions, normal theory maximum likelihood estimates of the standard error of the mediated effect exhibited less bias (approximately 10% to 20% too small) compared to the standard errors of the structural regression coefficients (20% to 45% too small). Asymptotically distribution free standard errors of both the mediated effect and the structural parameters were substantially affected by sample size, but not nonnormality. Robust standard errors consistently yielded the most accurate estimates of sampling variability.

Original languageEnglish (US)
Pages (from-to)87-107
Number of pages21
JournalStructural Equation Modeling
Volume4
Issue number2
DOIs
StatePublished - 1997

Fingerprint

Latent Variable Models
Non-normality
Standard error
Sample Size
Distribution-free
Structural Parameters
Small Sample Size
trend
Regression Coefficient
Maximum Likelihood Estimate
Estimate
Maximum likelihood
Sample size
Latent variable models
Sampling
Estimator
regression
simulation
Model
Values

ASJC Scopus subject areas

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

Cite this

@article{28838ffd3fab4a01afbf9c859efd1168,
title = "Effects of sample size and nonnormality on the estimation of mediated effects in latent variable models",
abstract = "A Monte Carlo approach was used to examine bias in the estimation of indirect effects and their associated standard errors. In the simulation design, (a) sample size, (b) the level of nonnormality characterizing the data, (c) the population values of the model parameters, and (d) the type of estimator were systematically varied. Estimates of model parameters were generally unaffected by either nonnormality or small sample size. Under severely nonnormal conditions, normal theory maximum likelihood estimates of the standard error of the mediated effect exhibited less bias (approximately 10{\%} to 20{\%} too small) compared to the standard errors of the structural regression coefficients (20{\%} to 45{\%} too small). Asymptotically distribution free standard errors of both the mediated effect and the structural parameters were substantially affected by sample size, but not nonnormality. Robust standard errors consistently yielded the most accurate estimates of sampling variability.",
author = "Finch, {John F.} and Stephen West and David Mackinnon",
year = "1997",
doi = "10.1080/10705519709540063",
language = "English (US)",
volume = "4",
pages = "87--107",
journal = "Structural Equation Modeling",
issn = "1070-5511",
publisher = "Psychology Press Ltd",
number = "2",

}

TY - JOUR

T1 - Effects of sample size and nonnormality on the estimation of mediated effects in latent variable models

AU - Finch, John F.

AU - West, Stephen

AU - Mackinnon, David

PY - 1997

Y1 - 1997

N2 - A Monte Carlo approach was used to examine bias in the estimation of indirect effects and their associated standard errors. In the simulation design, (a) sample size, (b) the level of nonnormality characterizing the data, (c) the population values of the model parameters, and (d) the type of estimator were systematically varied. Estimates of model parameters were generally unaffected by either nonnormality or small sample size. Under severely nonnormal conditions, normal theory maximum likelihood estimates of the standard error of the mediated effect exhibited less bias (approximately 10% to 20% too small) compared to the standard errors of the structural regression coefficients (20% to 45% too small). Asymptotically distribution free standard errors of both the mediated effect and the structural parameters were substantially affected by sample size, but not nonnormality. Robust standard errors consistently yielded the most accurate estimates of sampling variability.

AB - A Monte Carlo approach was used to examine bias in the estimation of indirect effects and their associated standard errors. In the simulation design, (a) sample size, (b) the level of nonnormality characterizing the data, (c) the population values of the model parameters, and (d) the type of estimator were systematically varied. Estimates of model parameters were generally unaffected by either nonnormality or small sample size. Under severely nonnormal conditions, normal theory maximum likelihood estimates of the standard error of the mediated effect exhibited less bias (approximately 10% to 20% too small) compared to the standard errors of the structural regression coefficients (20% to 45% too small). Asymptotically distribution free standard errors of both the mediated effect and the structural parameters were substantially affected by sample size, but not nonnormality. Robust standard errors consistently yielded the most accurate estimates of sampling variability.

UR - http://www.scopus.com/inward/record.url?scp=0001885793&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0001885793&partnerID=8YFLogxK

U2 - 10.1080/10705519709540063

DO - 10.1080/10705519709540063

M3 - Article

AN - SCOPUS:0001885793

VL - 4

SP - 87

EP - 107

JO - Structural Equation Modeling

JF - Structural Equation Modeling

SN - 1070-5511

IS - 2

ER -