Three tests of dimensionality in structural equation modeling: a Monte Carlo simulation study

Public Deposited
Analytics

Downloadable Content

Download PDF
Request Version for Screen Reader
Last Modified
  • March 21, 2019
Creator
  • Jiang, Niantao
    • Affiliation: College of Arts and Sciences, Department of Sociology
Abstract
  • The issue of dimensionality is essential to social science research but few researchers have empirically tested the dimensionality of theoretical constructs. One main reason is the uncertainty of how best to proceed. With the development of structural equation modeling with latent variables, several tests are available for researchers to choose. In this study, drawing on statistical theory and prior researches, I empirically assess the performance of likelihood ration test, confidence interval test, and vanishing tetrads test using data generated from Monte Carlo simulations. The study results show the likelihood ratio test did reasonably well. It does not show obvious signs of impact of the violation of boundary condition when testing for dimensionality. While overall the confidence interval test method appears to be too conservative, the vanishing tetrads tests for dimensionality works best for models with few indicators, but less well in larger models and smaller sample sizes.
Date of publication
DOI
Resource type
Rights statement
  • In Copyright
Advisor
  • Bollen, Kenneth
Degree granting institution
  • University of North Carolina at Chapel Hill
Language
Access right
  • Open access
Date uploaded
  • October 19, 2010

Relations

Parents:

This work has no parents.

Items

Thumbnail Title Date Uploaded Visibility Actions
file details Three tests of dimensionality in structural equation modeling : a Monte Carlo simulation study 2019-04-11 Public Download