Construct validity evidence for single-response items to estimate physical activity levels in large sample studies.

Valid measurement of physical activity is important for studying the risks for morbidity and mortality. The purpose of this study was to examine evidence of construct validity of two similar single-response items assessing physical activity via self-report. Both items are based on the stages of change model. The sample was 687 participants (men = 504, women = 183) who completed an 8-response (PA8) or 5-response (PA5) single-response item about current level of physical activity. Responses were categorized as meeting or not meeting guidelines for sufficient physical activity to achieve a health benefit. Maximal cardiorespiratory fitness (CRF) and health markers were obtained during a clinical examination. Partial correlation, multivariate analysis of covariance, and logistic regression were used to identify the relations between self-reported physical activity, CRF, and health markers when controlling for gender and age. Single-response items were compared to a detailed measure of physical activity. Single-response items correlated significantly with CRF determined with a maximal exercise test on a treadmill (PA8 = .53; PA5 = .57). Differences in percentage of body fat and cholesterol were in the desired direction, with those self-reporting sufficient physical activity for a health benefit having the lower values. The single-response items demonstrated evidence of construct validity and may provide feasible, cost-effective, and efficient methods to assess physical activity in large-scale studies.