Deciding on the Number of Classes in Latent Class Analysis and Growth Mixture Modeling: A Monte Carlo Simulation Study
Определение числа классов в латентном классовом анализе и моделях смеси роста: исследование методом Монте-Карло
2007-10-23
SCID: 54.1/n9dzjzwb
Discuss with AI
Bayesian Information Criterionbootstrap likelihood ratio testfactor mixture modelgrowth mixture modelslatent class analysis
Figures from the paper
Abstract (AI)
Mixture modeling is a widely applied data analysis technique used to identify unobserved heterogeneity in a population. Despite mixture models' usefulness in practice, one unresolved issue in the application of mixture models is that there is not one commonly accepted statistical indicator for deciding on the number of classes in a study population. This article presents the results of a simulation study that examines the performance of likelihood-based tests and the traditionally used Information Criterion (ICs) used for determining the number of classes in mixture modeling. We look at the performance of these tests and indexes for 3 types of mixture models: latent class analysis (LCA), a factor mixture model (FMA), and a growth mixture models (GMM). We evaluate the ability of the tests and indexes to correctly identify the number of classes at three different sample sizes (n = 200, 500, 1,000). Whereas the Bayesian Information Criterion performed the best of the ICs, the bootstrap likelihood ratio test proved to be a very consistent indicator of classes across all of the models considered.
Key Findings
1
Bayesian Information Criterion (BIC) performed best among the information criteria evaluated.
2
Convergence problems occurred for badly misspecified models (e.g., GMM with true k=3: convergence rates for 3-, 4-, 5-class models at n=500 were 100%, 87%, and 68%).
3
The bootstrap likelihood ratio test was a very consistent indicator of the correct number of classes across all models considered.
4
The study compared likelihood-based tests and information criteria (ICs) for LCA, factor mixture models, and growth mixture models across sample sizes n=200, 500, 1000.
5
There is no single commonly accepted statistical indicator for determining number of classes in mixture models.
Research Object
Mixture modeling procedures (latent class analysis, factor mixture models, and growth mixture models) used to determine the number of latent classes in a population
Research Subject
Performance of likelihood-based tests and information criteria (e.g., bootstrap likelihood ratio test, Bayesian Information Criterion) for correctly identifying the number of classes across model types and sample sizes
Publication Details
Publication Date
2007-10-23
Journal
Publisher
ISSN
Cited by
12165
Access Type
Author Information
Download PDF
Subscribe to digest
Cited by8
Latent burnout profiles: A new approach to understanding the burnout experience2016
Longitudinal Leisure Activity Profiles and Their Associations with Recovery Experiences and Job Performance2017
Associations between relational sexual behaviour, pornography use, and pornography acceptance among US college students2014
Preferences for community renewable energy investments in Europe2021
Identifying higher education students’ profiles of academic engagement and burnout and analysing their predictors and outcomes2024
Delinquency in Emerging Adulthood: Insights into Trajectories of Young Adults in a German Sample and Implications for Measuring Continuity of Offending2020
Developmental stability and change in emotion regulation strategies and strategy repertoires across adolescence2026
Co-occurring stigmas and associations with HIV testing, mental health, and substance use: a latent class analysis among youth in Kazakhstan2026