Using Data-Dependent Priors to Mitigate Small Sample Bias in Latent Growth Models: A Discussion and Illustration Using Mplus

Research output: Contribution to journalArticle

23 Citations (Scopus)

Abstract

Mixed-effects models (MEMs) and latent growth models (LGMs) are often considered interchangeable save the discipline-specific nomenclature. Software implementations of these models, however, are not interchangeable, particularly with small sample sizes. Restricted maximum likelihood estimation that mitigates small sample bias in MEMs has not been widely developed for LGMs, and fully Bayesian methods, while not dependent on asymptotics, can encounter issues because the choice for the factor covariance matrix prior distribution has substantial influence with small samples. This tutorial discusses differences between LGMs and MEMs and demonstrates how data-dependent priors, an established class of methods that blend frequentist and Bayesian paradigms, can be implemented within Mplus 7.1 to abate the small sample bias that is prevalent with LGM software while keeping additional programming to the bare minimum.

Original languageEnglish (US)
Pages (from-to)27-56
Number of pages30
JournalJournal of Educational and Behavioral Statistics
Volume41
Issue number1
DOIs
StatePublished - Feb 1 2016
Externally publishedYes

Fingerprint

trend
programming
paradigm
software

Keywords

  • data-dependent prior
  • latent basis model
  • latent growth models
  • Mplus
  • second-order growth model
  • small samples

ASJC Scopus subject areas

  • Education
  • Social Sciences (miscellaneous)

Cite this

@article{5a6af9a7ec214166be3d36e54fdf911f,
title = "Using Data-Dependent Priors to Mitigate Small Sample Bias in Latent Growth Models: A Discussion and Illustration Using Mplus",
abstract = "Mixed-effects models (MEMs) and latent growth models (LGMs) are often considered interchangeable save the discipline-specific nomenclature. Software implementations of these models, however, are not interchangeable, particularly with small sample sizes. Restricted maximum likelihood estimation that mitigates small sample bias in MEMs has not been widely developed for LGMs, and fully Bayesian methods, while not dependent on asymptotics, can encounter issues because the choice for the factor covariance matrix prior distribution has substantial influence with small samples. This tutorial discusses differences between LGMs and MEMs and demonstrates how data-dependent priors, an established class of methods that blend frequentist and Bayesian paradigms, can be implemented within Mplus 7.1 to abate the small sample bias that is prevalent with LGM software while keeping additional programming to the bare minimum.",
keywords = "data-dependent prior, latent basis model, latent growth models, Mplus, second-order growth model, small samples",
author = "Daniel McNeish",
year = "2016",
month = "2",
day = "1",
doi = "10.3102/1076998615621299",
language = "English (US)",
volume = "41",
pages = "27--56",
journal = "Journal of Educational and Behavioral Statistics",
issn = "1076-9986",
publisher = "SAGE Publications Inc.",
number = "1",

}

TY - JOUR

T1 - Using Data-Dependent Priors to Mitigate Small Sample Bias in Latent Growth Models

T2 - A Discussion and Illustration Using Mplus

AU - McNeish, Daniel

PY - 2016/2/1

Y1 - 2016/2/1

N2 - Mixed-effects models (MEMs) and latent growth models (LGMs) are often considered interchangeable save the discipline-specific nomenclature. Software implementations of these models, however, are not interchangeable, particularly with small sample sizes. Restricted maximum likelihood estimation that mitigates small sample bias in MEMs has not been widely developed for LGMs, and fully Bayesian methods, while not dependent on asymptotics, can encounter issues because the choice for the factor covariance matrix prior distribution has substantial influence with small samples. This tutorial discusses differences between LGMs and MEMs and demonstrates how data-dependent priors, an established class of methods that blend frequentist and Bayesian paradigms, can be implemented within Mplus 7.1 to abate the small sample bias that is prevalent with LGM software while keeping additional programming to the bare minimum.

AB - Mixed-effects models (MEMs) and latent growth models (LGMs) are often considered interchangeable save the discipline-specific nomenclature. Software implementations of these models, however, are not interchangeable, particularly with small sample sizes. Restricted maximum likelihood estimation that mitigates small sample bias in MEMs has not been widely developed for LGMs, and fully Bayesian methods, while not dependent on asymptotics, can encounter issues because the choice for the factor covariance matrix prior distribution has substantial influence with small samples. This tutorial discusses differences between LGMs and MEMs and demonstrates how data-dependent priors, an established class of methods that blend frequentist and Bayesian paradigms, can be implemented within Mplus 7.1 to abate the small sample bias that is prevalent with LGM software while keeping additional programming to the bare minimum.

KW - data-dependent prior

KW - latent basis model

KW - latent growth models

KW - Mplus

KW - second-order growth model

KW - small samples

UR - http://www.scopus.com/inward/record.url?scp=84955483682&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84955483682&partnerID=8YFLogxK

U2 - 10.3102/1076998615621299

DO - 10.3102/1076998615621299

M3 - Article

AN - SCOPUS:84955483682

VL - 41

SP - 27

EP - 56

JO - Journal of Educational and Behavioral Statistics

JF - Journal of Educational and Behavioral Statistics

SN - 1076-9986

IS - 1

ER -