PURPOSE: Exploratory factor analysis is used to identify latent factors for public health interventions. The most popular factor retention criterion, the eigenvalue greater than one (EVG1) rule, leads to the retention of more factors than warranted. The use of parallel analysis (PA) as a factor retention criterion is recommended. DESIGN: Environmental factors that are likely to affect the propensity of individuals to walk are identified. Results from PA are compared with those obtained by using the EVG1 rule. SETTING: Telephone survey data were collected from three communities (spanning 12 counties) in the Appalachian region. PARTICIPANTS: The sample was composed of adults between the ages of 40 and 65 years. There were 1019 completed surveys. MEASURES: The data on 14 variables, including measures of neighborhood safety, beauty, and access to walking facilities were collected. ANALYSIS: A principal components analysis was performed. The factors retained after using the EVG1 rule were compared with the factors retained after using PA. Varimax rotation was used to aid factor interpretation. RESULTS: The EVG1 rule led to the retention of nearly twice the number of latent factors as did the PA criterion. CONCLUSION: The use of the EVG1 rule in research may mislead policy makers to focus on trivial interventions. They are urged to use PA to obtain more parsimonious and externally valid interventions.
PURPOSE: Exploratory factor analysis is used to identify latent factors for public health interventions. The most popular factor retention criterion, the eigenvalue greater than one (EVG1) rule, leads to the retention of more factors than warranted. The use of parallel analysis (PA) as a factor retention criterion is recommended. DESIGN: Environmental factors that are likely to affect the propensity of individuals to walk are identified. Results from PA are compared with those obtained by using the EVG1 rule. SETTING: Telephone survey data were collected from three communities (spanning 12 counties) in the Appalachian region. PARTICIPANTS: The sample was composed of adults between the ages of 40 and 65 years. There were 1019 completed surveys. MEASURES: The data on 14 variables, including measures of neighborhood safety, beauty, and access to walking facilities were collected. ANALYSIS: A principal components analysis was performed. The factors retained after using the EVG1 rule were compared with the factors retained after using PA. Varimax rotation was used to aid factor interpretation. RESULTS: The EVG1 rule led to the retention of nearly twice the number of latent factors as did the PA criterion. CONCLUSION: The use of the EVG1 rule in research may mislead policy makers to focus on trivial interventions. They are urged to use PA to obtain more parsimonious and externally valid interventions.