Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

When constraints interact: a caution about reference variables, identification constraints, and scale dependencies in

James H Steiger1

  • 1Department of Psychology, University of British Columbia, Vancouver, Canada. steiger@unixg.ubc.ca

Psychological Methods
|July 2, 2002
PubMed
Summary

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

A Tribute to the Mind, Methodology and Mentoring of Wayne Velicer.

Multivariate behavioral research·2020
Same author

Creativity and technical innovation: spatial ability's unique role.

Psychological science·2013
Same author

A 12-Point Circumplex Structure of Core Affect.

Emotion (Washington, D.C.)·2011
Same author

The insidious effects of failing to include design-driven correlated residuals in latent-variable covariance structure analysis.

Psychological methods·2008
Same author

Empirical and conceptual problems with longitudinal trait-state models: introducing a trait-state-occasion model.

Psychological methods·2005
Same author

Beyond the F test: Effect size confidence intervals and tests of close fit in the analysis of variance and contrast analysis.

Psychological methods·2004

Structural equation modeling requires careful handling of latent variable variances. This study reveals that common constraints for identifying these variances can complicate hypothesis testing and lead to inaccurate results in popular software.

Area of Science:

  • Statistics
  • Psychometrics
  • Quantitative Psychology

Background:

  • Traditional structural equation modeling (SEM) methods face challenges in directly specifying or constraining variances of latent endogenous variables.
  • Latent variable variances are typically unidentified without specific identification strategies.

Purpose of the Study:

  • To critically examine the underlying principles and practical implications of constraints used for latent variable variance identification in SEM.
  • To highlight potential issues and complexities associated with these constraints that may be overlooked in standard SEM training materials.

Main Methods:

  • Review of established identification strategies in structural equation modeling.
  • Analysis of the impact of variance constraints on model identification and hypothesis testing.

Related Experiment Videos

  • Examination of the interaction between variance constraints and other model constraints.
  • Evaluation of the consequences of these constraints on standardized solutions using common SEM software.

    • Main Results:

    • Standard methods for achieving identification, such as fixing factor loadings, are commonly employed.
    • Alternative methods involve fixing variances using constrained estimation algorithms.
    • The application and interpretation of these variance constraints are more complex than typically presented.
    • Constraints on latent variable variances can interfere with hypothesis testing and produce erroneous standardized solutions in some software packages.

    Conclusions:

    • The use of constraints for latent variable variances in SEM is not always straightforward or uncontroversial.
    • Researchers must be aware of the potential for these constraints to impact hypothesis testing and standardized solutions.
    • Careful consideration and validation are necessary when implementing variance constraints in structural equation models.