Context matters: Construct framing in measures of physical activity engagement among African American women

Authors

Stephanie M. McClure, University of AlabamaFollow
Travis Loux, Saint Louis UniversityFollow
Enbal Shacham, Saint Louis UniversityFollow
Eileen Gillespie, Saint Louis UniversityFollow
Denise Hooks-Anderson, Saint Louis UniversityFollow

Keywords

African American women; physical activity; theory of planned behavior; gender norms; role conflict

Disciplines

Community Health and Preventive Medicine | Other Anthropology | Public Health Education and Promotion

Abstract

Assessment of psychosocial factors influencing health behavior typically privileges conceptual consistency (framing constructs similarly across contexts) over conceptual specificity (context-specific framing). Modest statistical relationships between these factors and health behaviors, and persistent racial disparities in health outcomes raise questions about whether conceptually consistent framing fully captures relevant predictors. Ethnographic studies suggest not - that perceptions influencing health behaviors are multifaceted and contextual. To test this, we added items querying contextualized predictors of intention to engage in leisure-time physical activity (LTPA) to a Theory of Planned Behavior (TPB)-based survey and examined the psychometrics of the adapted subscales. We measured internal consistency (Cronbach’s alpha) and construct validity (exploratory factor analysis using polychoric correlations for ordinal data). Participants were a convenience sample of 200 African American women in a Midwestern, suburban University-affiliated family medicine practice.

Reliability of the adapted subscales was notably lower than the original subscales. A two-factor model fit best for the attitudes subscale, but explained slightly less than 50% of the variance. The new items loaded strongly on one factor. A three-factor model best fit the norms subscale and accounted for around 57% of the variance. Two of the three new items loaded strongly on one factor. Factor analysis for the perceived control subscale was not possible due to low number of items; however, two of the three new items were highly correlated (.73). Including context-specific factors may improve assessment of intention to engage in LTPA. Further study of this question with a larger, representative sample is warranted.

Recommended Citation

McClure, Stephanie M.; Loux, Travis; Shacham, Enbal; Gillespie, Eileen; and Hooks-Anderson, Denise (2018) "Context matters: Construct framing in measures of physical activity engagement among African American women," Journal of Health Disparities Research and Practice: Vol. 12: Iss. 3, Article 2.
Available at: https://digitalscholarship.unlv.edu/jhdrp/vol12/iss3/2