Analyzing Clustered Data with OLS Regression: The Effect of a Hierarchical Data Structure
Year: 2014
Type: article
Abstract: A previous study by Mundfrom and Schultz (2002) has shown that coefficient estimates between Ordinary Least Squares (OLS) regression and Hierarchical Linear Models (HLM) for clustered data are, for all intents and purposes, equivalent. However, when comparing these two methods, the standard error estimates are the true cause for concern. A simulation study is conducted to demonstrate that the standard error estimates between OLS and HLM are quite different for clustered data, and to show that the differential standard errors leads to inflated Type I error rates using OLS under such conditions.
Source: General Linear Model Journal
Author: Daniel McNeish
Language: English
Subfield: Statistics and Probability
Field: Mathematics
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W2183094219