The Effect of Small Sample Size on Two-Level Model Estimates: A Review and Illustration

Multilevel models are an increasingly popular method to analyze data that originate from a clustered or hierarchical structure. To effectively utilize multilevel models, one must have an adequately large number of clusters; otherwise, some model parameters will be estimated with bias. The goals for...

Full description

Saved in:
Bibliographic Details
Published in:Educational psychology review 2016-06, Vol.28 (2), p.295-314
Main Authors: McNeish, Daniel M., Stapleton, Laura M.
Format: Article
Language:English
Subjects:
Child and School Psychology
Cluster Grouping
Data analysis
Education
Educational aspects
Educational Psychology
Effect Size
Estimation bias
Estimation methods
Hierarchical Linear Modeling
Learning and Instruction
Least Squares Statistics
Literature Reviews
Maximum likelihood estimation
Maximum Likelihood Statistics
Multilevel models
Multivariate Analysis
Population estimates
Research methods
Research Problems
Review Article
REVIEW ARTICLES
Sample Size
Sampling distributions
Simulation
Simulations
Standard error
Statistical Analysis
Statistical variance
Studies
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Multilevel models are an increasingly popular method to analyze data that originate from a clustered or hierarchical structure. To effectively utilize multilevel models, one must have an adequately large number of clusters; otherwise, some model parameters will be estimated with bias. The goals for this paper are to (1) raise awareness of the problems associated with a small number of clusters, (2) review previous studies on multilevel models with a small number of clusters, (3) to provide an illustrative simulation to demonstrate how a simple model becomes adversely affected by small numbers of clusters, (4) to provide researchers with remedies if they encounter clustered data with a small number of clusters, and (5) to outline methodological topics that have yet to be addressed in the literature.
ISSN:1040-726X
1573-336X
DOI:10.1007/s10648-014-9287-x