The Effects of Sample Size on the Estimation of Regression Mixture Models

Thomas Jaki, Minjung Kim, Andrea Lamont, Melissa George, Chi Chang, Daniel J Feaster, M. Lee Van Horn

Research output: Contribution to journalArticle

2 Citations (Scopus)

Abstract

Regression mixture models are a statistical approach used for estimating heterogeneity in effects. This study investigates the impact of sample size on regression mixture’s ability to produce “stable” results. Monte Carlo simulations and analysis of resamples from an application data set were used to illustrate the types of problems that may occur with small samples in real data sets. The results suggest that (a) when class separation is low, very large sample sizes may be needed to obtain stable results; (b) it may often be necessary to consider a preponderance of evidence in latent class enumeration; (c) regression mixtures with ordinal outcomes result in even more instability; and (d) with small samples, it is possible to obtain spurious results without any clear indication of there being a problem.

Original languageEnglish (US)
JournalEducational and Psychological Measurement
DOIs
StateAccepted/In press - Jan 1 2018

Fingerprint

Mixture Model
Sample Size
Regression Model
regression
Statistical Models
Small Sample
Regression
Latent Class
indication
Enumeration
Monte Carlo Simulation
simulation
Datasets
ability
Necessary
evidence

Keywords

  • heterogeneous effects
  • regression mixture models
  • sample size

ASJC Scopus subject areas

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

Cite this

The Effects of Sample Size on the Estimation of Regression Mixture Models. / Jaki, Thomas; Kim, Minjung; Lamont, Andrea; George, Melissa; Chang, Chi; Feaster, Daniel J; Van Horn, M. Lee.

In: Educational and Psychological Measurement, 01.01.2018.

Research output: Contribution to journalArticle

Jaki, Thomas ; Kim, Minjung ; Lamont, Andrea ; George, Melissa ; Chang, Chi ; Feaster, Daniel J ; Van Horn, M. Lee. / The Effects of Sample Size on the Estimation of Regression Mixture Models. In: Educational and Psychological Measurement. 2018.
@article{7b17d80fb494419e9fdba1cbdb25ad55,
title = "The Effects of Sample Size on the Estimation of Regression Mixture Models",
abstract = "Regression mixture models are a statistical approach used for estimating heterogeneity in effects. This study investigates the impact of sample size on regression mixture’s ability to produce “stable” results. Monte Carlo simulations and analysis of resamples from an application data set were used to illustrate the types of problems that may occur with small samples in real data sets. The results suggest that (a) when class separation is low, very large sample sizes may be needed to obtain stable results; (b) it may often be necessary to consider a preponderance of evidence in latent class enumeration; (c) regression mixtures with ordinal outcomes result in even more instability; and (d) with small samples, it is possible to obtain spurious results without any clear indication of there being a problem.",
keywords = "heterogeneous effects, regression mixture models, sample size",
author = "Thomas Jaki and Minjung Kim and Andrea Lamont and Melissa George and Chi Chang and Feaster, {Daniel J} and {Van Horn}, {M. Lee}",
year = "2018",
month = "1",
day = "1",
doi = "10.1177/0013164418791673",
language = "English (US)",
journal = "Educational and Psychological Measurement",
issn = "0013-1644",
publisher = "SAGE Publications Inc.",

}

TY - JOUR

T1 - The Effects of Sample Size on the Estimation of Regression Mixture Models

AU - Jaki, Thomas

AU - Kim, Minjung

AU - Lamont, Andrea

AU - George, Melissa

AU - Chang, Chi

AU - Feaster, Daniel J

AU - Van Horn, M. Lee

PY - 2018/1/1

Y1 - 2018/1/1

N2 - Regression mixture models are a statistical approach used for estimating heterogeneity in effects. This study investigates the impact of sample size on regression mixture’s ability to produce “stable” results. Monte Carlo simulations and analysis of resamples from an application data set were used to illustrate the types of problems that may occur with small samples in real data sets. The results suggest that (a) when class separation is low, very large sample sizes may be needed to obtain stable results; (b) it may often be necessary to consider a preponderance of evidence in latent class enumeration; (c) regression mixtures with ordinal outcomes result in even more instability; and (d) with small samples, it is possible to obtain spurious results without any clear indication of there being a problem.

AB - Regression mixture models are a statistical approach used for estimating heterogeneity in effects. This study investigates the impact of sample size on regression mixture’s ability to produce “stable” results. Monte Carlo simulations and analysis of resamples from an application data set were used to illustrate the types of problems that may occur with small samples in real data sets. The results suggest that (a) when class separation is low, very large sample sizes may be needed to obtain stable results; (b) it may often be necessary to consider a preponderance of evidence in latent class enumeration; (c) regression mixtures with ordinal outcomes result in even more instability; and (d) with small samples, it is possible to obtain spurious results without any clear indication of there being a problem.

KW - heterogeneous effects

KW - regression mixture models

KW - sample size

UR - http://www.scopus.com/inward/record.url?scp=85052581087&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85052581087&partnerID=8YFLogxK

U2 - 10.1177/0013164418791673

DO - 10.1177/0013164418791673

M3 - Article

JO - Educational and Psychological Measurement

JF - Educational and Psychological Measurement

SN - 0013-1644

ER -