Physical Activity Behavior Change of EFNEP Adult Participants: Item Generation, Validity, and Reliability Testing

Abstract

To develop and validate self-report physical activity (PA) items designed for assessing physical activity behavioral change among low-income women who participate in the Expanded Food and Nutrition Education Program (EFNEP). The Community Nutrition Education logic model, adapted with constructs of Theory of Reasoned Action, and Theory of Planned Behavior provided the theoretical framework to develop the PA items. The study design for questionnaire development and testing included 6 phases: curricula content analysis; conceptualization and item generation; expert reviews; revisions; cognitive interviews and rigorous psychometric testing. The study was conducted in 6 South Carolina rural counties and with EFNEP eligible mothers. Seventeen items were developed on PA psychosocial mediating variables including attitudes, subjective norm, perceived behavioral control and intention; and 12 items were developed in 4 PA dimensions: home, yard, walk, and work. Content validity was assessed through curricula content analysis, expert reviews (n=8) and cognitive interviews (n=14). 302 mothers from low-income communities in South Carolina completed the PA questionnaire. Factor validity was assessed through Exploratory Factor Analysis (EFA). Reliability (n=50) was assessed using Cronbach for internal consistency and Spearman correlation for test-retest reliability. Predictive validity was assessed (n-50) using accelerometer data. Factor analysis revealed three interpretable factors for the set of PA psychosocial items, with adequate factor loading, acceptable internal consistency and test-retest reliability. No significant association was found for predictive validity analysis. Although further research is needed, this study provided an important methodological foundation for the evaluation and measurement of PA behavior among low-income EFNEP participants.