Hopes and cautions in implementing Bayesian structural equation modeling

Robert C. MacCallum, Michael Edwards, Li Cai

Research output: Contribution to journalComment/debate

23 Citations (Scopus)

Abstract

Muthén and Asparouhov (2012) have proposed and demonstrated an approach to model specification and estimation in structural equation modeling (SEM) using Bayesian methods. Their contribution builds on previous work in this area by (a) focusing on the translation of conventional SEM models into a Bayesian framework wherein parameters fixed at zero in a conventional model can be respecified using small-variance priors and (b) implementing their approach in software that is widely accessible. We recognize potential benefits for applied researchers as discussed by Muthén and Asparouhov, and we also see a tradeoff in that effective use of the proposed approach introduces increased demands in terms of expertise of users to navigate new complexities in model specification, parameter estimation, and evaluation of results. We also raise cautions regarding the issues of model modification and model fit. Although we see significant potential value in the use of Bayesian SEM, we also believe that effective use will require an awareness of these complexities.

Original languageEnglish (US)
Pages (from-to)340-345
Number of pages6
JournalPsychological Methods
Volume17
Issue number3
DOIs
StatePublished - Jan 1 2012
Externally publishedYes

Fingerprint

Bayes Theorem
Software
Research Personnel

Keywords

  • Bayesian statistics
  • Factor analysis
  • Structural equation modeling

ASJC Scopus subject areas

  • Psychology (miscellaneous)

Cite this

Hopes and cautions in implementing Bayesian structural equation modeling. / MacCallum, Robert C.; Edwards, Michael; Cai, Li.

In: Psychological Methods, Vol. 17, No. 3, 01.01.2012, p. 340-345.

Research output: Contribution to journalComment/debate

MacCallum, Robert C. ; Edwards, Michael ; Cai, Li. / Hopes and cautions in implementing Bayesian structural equation modeling. In: Psychological Methods. 2012 ; Vol. 17, No. 3. pp. 340-345.
@article{25ff61495751403d86d038c28a2f0816,
title = "Hopes and cautions in implementing Bayesian structural equation modeling",
abstract = "Muth{\'e}n and Asparouhov (2012) have proposed and demonstrated an approach to model specification and estimation in structural equation modeling (SEM) using Bayesian methods. Their contribution builds on previous work in this area by (a) focusing on the translation of conventional SEM models into a Bayesian framework wherein parameters fixed at zero in a conventional model can be respecified using small-variance priors and (b) implementing their approach in software that is widely accessible. We recognize potential benefits for applied researchers as discussed by Muth{\'e}n and Asparouhov, and we also see a tradeoff in that effective use of the proposed approach introduces increased demands in terms of expertise of users to navigate new complexities in model specification, parameter estimation, and evaluation of results. We also raise cautions regarding the issues of model modification and model fit. Although we see significant potential value in the use of Bayesian SEM, we also believe that effective use will require an awareness of these complexities.",
keywords = "Bayesian statistics, Factor analysis, Structural equation modeling",
author = "MacCallum, {Robert C.} and Michael Edwards and Li Cai",
year = "2012",
month = "1",
day = "1",
doi = "10.1037/a0027131",
language = "English (US)",
volume = "17",
pages = "340--345",
journal = "Psychological Methods",
issn = "1082-989X",
publisher = "American Psychological Association Inc.",
number = "3",

}

TY - JOUR

T1 - Hopes and cautions in implementing Bayesian structural equation modeling

AU - MacCallum, Robert C.

AU - Edwards, Michael

AU - Cai, Li

PY - 2012/1/1

Y1 - 2012/1/1

N2 - Muthén and Asparouhov (2012) have proposed and demonstrated an approach to model specification and estimation in structural equation modeling (SEM) using Bayesian methods. Their contribution builds on previous work in this area by (a) focusing on the translation of conventional SEM models into a Bayesian framework wherein parameters fixed at zero in a conventional model can be respecified using small-variance priors and (b) implementing their approach in software that is widely accessible. We recognize potential benefits for applied researchers as discussed by Muthén and Asparouhov, and we also see a tradeoff in that effective use of the proposed approach introduces increased demands in terms of expertise of users to navigate new complexities in model specification, parameter estimation, and evaluation of results. We also raise cautions regarding the issues of model modification and model fit. Although we see significant potential value in the use of Bayesian SEM, we also believe that effective use will require an awareness of these complexities.

AB - Muthén and Asparouhov (2012) have proposed and demonstrated an approach to model specification and estimation in structural equation modeling (SEM) using Bayesian methods. Their contribution builds on previous work in this area by (a) focusing on the translation of conventional SEM models into a Bayesian framework wherein parameters fixed at zero in a conventional model can be respecified using small-variance priors and (b) implementing their approach in software that is widely accessible. We recognize potential benefits for applied researchers as discussed by Muthén and Asparouhov, and we also see a tradeoff in that effective use of the proposed approach introduces increased demands in terms of expertise of users to navigate new complexities in model specification, parameter estimation, and evaluation of results. We also raise cautions regarding the issues of model modification and model fit. Although we see significant potential value in the use of Bayesian SEM, we also believe that effective use will require an awareness of these complexities.

KW - Bayesian statistics

KW - Factor analysis

KW - Structural equation modeling

UR - http://www.scopus.com/inward/record.url?scp=84874392393&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84874392393&partnerID=8YFLogxK

U2 - 10.1037/a0027131

DO - 10.1037/a0027131

M3 - Comment/debate

C2 - 22962888

AN - SCOPUS:84874392393

VL - 17

SP - 340

EP - 345

JO - Psychological Methods

JF - Psychological Methods

SN - 1082-989X

IS - 3

ER -