Fitting Residual Error Structures for Growth Models in SAS PROC MCMC

Research output: Contribution to journalArticle

Abstract

In behavioral sciences broadly, estimating growth models with Bayesian methods is becoming increasingly common, especially to combat small samples common with longitudinal data. Although Mplus is becoming an increasingly common program for applied research employing Bayesian methods, the limited selection of prior distributions for the elements of covariance structures makes more general software more advantages under certain conditions. However, as a disadvantage of general software’s software flexibility, few preprogrammed commands exist for specifying covariance structures. For instance, PROC MIXED has a few dozen such preprogrammed options, but when researchers divert to a Bayesian framework, software offer no such guidance and requires researchers to manually program these different structures, which is no small task. As such the literature has noted that empirical papers tend to simplify their covariance matrices to circumvent this difficulty, which is not desirable because such a simplification will likely lead to biased estimates of variance components and standard errors. To facilitate wider implementation of Bayesian growth models that properly model covariance structures, this article overviews how to generally program a growth model in SAS PROC MCMC and then demonstrates how to program common residual error structures. Full annotated SAS code and an applied example are provided.

Original languageEnglish (US)
Pages (from-to)587-612
Number of pages26
JournalEducational and Psychological Measurement
Volume77
Issue number4
DOIs
StatePublished - Aug 1 2017
Externally publishedYes

Fingerprint

Growth Model
Markov Chain Monte Carlo
Covariance Structure
Software
Bayes Theorem
Bayesian Methods
Growth
Research Personnel
Behavioral Sciences
Components of Variance
Research Methods
Bayesian Model
Longitudinal Data
Standard error
Covariance matrix
Prior distribution
Small Sample
behavioral science
Simplification
Biased

Keywords

  • Bayes
  • latent growth model
  • MCMC
  • SAS

ASJC Scopus subject areas

  • Education
  • Developmental and Educational Psychology
  • Applied Psychology
  • Applied Mathematics

Cite this

Fitting Residual Error Structures for Growth Models in SAS PROC MCMC. / McNeish, Daniel.

In: Educational and Psychological Measurement, Vol. 77, No. 4, 01.08.2017, p. 587-612.

Research output: Contribution to journalArticle

@article{8b83216588bd4f6c8b9f2478129939ad,
title = "Fitting Residual Error Structures for Growth Models in SAS PROC MCMC",
abstract = "In behavioral sciences broadly, estimating growth models with Bayesian methods is becoming increasingly common, especially to combat small samples common with longitudinal data. Although Mplus is becoming an increasingly common program for applied research employing Bayesian methods, the limited selection of prior distributions for the elements of covariance structures makes more general software more advantages under certain conditions. However, as a disadvantage of general software’s software flexibility, few preprogrammed commands exist for specifying covariance structures. For instance, PROC MIXED has a few dozen such preprogrammed options, but when researchers divert to a Bayesian framework, software offer no such guidance and requires researchers to manually program these different structures, which is no small task. As such the literature has noted that empirical papers tend to simplify their covariance matrices to circumvent this difficulty, which is not desirable because such a simplification will likely lead to biased estimates of variance components and standard errors. To facilitate wider implementation of Bayesian growth models that properly model covariance structures, this article overviews how to generally program a growth model in SAS PROC MCMC and then demonstrates how to program common residual error structures. Full annotated SAS code and an applied example are provided.",
keywords = "Bayes, latent growth model, MCMC, SAS",
author = "Daniel McNeish",
year = "2017",
month = "8",
day = "1",
doi = "10.1177/0013164416652441",
language = "English (US)",
volume = "77",
pages = "587--612",
journal = "Educational and Psychological Measurement",
issn = "0013-1644",
publisher = "SAGE Publications Inc.",
number = "4",

}

TY - JOUR

T1 - Fitting Residual Error Structures for Growth Models in SAS PROC MCMC

AU - McNeish, Daniel

PY - 2017/8/1

Y1 - 2017/8/1

N2 - In behavioral sciences broadly, estimating growth models with Bayesian methods is becoming increasingly common, especially to combat small samples common with longitudinal data. Although Mplus is becoming an increasingly common program for applied research employing Bayesian methods, the limited selection of prior distributions for the elements of covariance structures makes more general software more advantages under certain conditions. However, as a disadvantage of general software’s software flexibility, few preprogrammed commands exist for specifying covariance structures. For instance, PROC MIXED has a few dozen such preprogrammed options, but when researchers divert to a Bayesian framework, software offer no such guidance and requires researchers to manually program these different structures, which is no small task. As such the literature has noted that empirical papers tend to simplify their covariance matrices to circumvent this difficulty, which is not desirable because such a simplification will likely lead to biased estimates of variance components and standard errors. To facilitate wider implementation of Bayesian growth models that properly model covariance structures, this article overviews how to generally program a growth model in SAS PROC MCMC and then demonstrates how to program common residual error structures. Full annotated SAS code and an applied example are provided.

AB - In behavioral sciences broadly, estimating growth models with Bayesian methods is becoming increasingly common, especially to combat small samples common with longitudinal data. Although Mplus is becoming an increasingly common program for applied research employing Bayesian methods, the limited selection of prior distributions for the elements of covariance structures makes more general software more advantages under certain conditions. However, as a disadvantage of general software’s software flexibility, few preprogrammed commands exist for specifying covariance structures. For instance, PROC MIXED has a few dozen such preprogrammed options, but when researchers divert to a Bayesian framework, software offer no such guidance and requires researchers to manually program these different structures, which is no small task. As such the literature has noted that empirical papers tend to simplify their covariance matrices to circumvent this difficulty, which is not desirable because such a simplification will likely lead to biased estimates of variance components and standard errors. To facilitate wider implementation of Bayesian growth models that properly model covariance structures, this article overviews how to generally program a growth model in SAS PROC MCMC and then demonstrates how to program common residual error structures. Full annotated SAS code and an applied example are provided.

KW - Bayes

KW - latent growth model

KW - MCMC

KW - SAS

UR - http://www.scopus.com/inward/record.url?scp=85025087243&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85025087243&partnerID=8YFLogxK

U2 - 10.1177/0013164416652441

DO - 10.1177/0013164416652441

M3 - Article

AN - SCOPUS:85025087243

VL - 77

SP - 587

EP - 612

JO - Educational and Psychological Measurement

JF - Educational and Psychological Measurement

SN - 0013-1644

IS - 4

ER -