A Primer on Bootstrap Factor Analysis as Applied to Health Studies Research

Abstract

Objectives: To demonstrate how the bootstrap method could be conducted in exploratory factor analysis (EFA) with a syntax written in SPSS. Methods: The data obtained from the Texas Childhood Obesity Prevention Policy Evaluation project (T-COPPE project) were used for illustration. A 5-step procedure to conduct bootstrap factor analysis (BFA) was introduced: (1) conduct principal component analysis, (2) create the target matrix, (3) resample with replacement, (4) conduct EFA and Procrustes rotation, and (5) calculate BFA results. Results: Results of the illustrated BFA example indicated that the 15 variables from this study were stable across the resamples. Conclusions: The bootstrap method, when applied in EFA for health-related research, is a particularly powerful internal replicability analysis.

Acknowledgments

The data used as the example for this article were derived the Texas Childhood Obesity Prevention Policy Evaluation (T-COPPE), which was funded by the Robert Wood Johnson Foundation. Additional important contributions were also made by the Texas A&M Health University Science Center School of Rural Public Health, The Michael & Susan Dell Center for Healthy Living at The University of Texas School of Public Health, and the Texas Department of State Health Services. We give our thanks to all members of the T-COPPE team, particularly Carolyn Smith and other researchers for their data collection efforts.