Test-retest reliability of the Brunel lifestyle physical activity questionnaire

Objectives: To establish the test–retest reliability of planned physical activity (PPA) and unplanned physical activity (UPA) components of the Brunel Lifestyle Physical Activity Questionnaire (BLPAQ). To provide evidence of the BLPAQ's stability using the proportion of agreement (PoA) method over a 5-week period. Design: Test–retest over a 5-week period using three diverse samples of adults. Methods: The 277 participants were subdivided into three adult samples: gymnasium users (n = 80), undergraduate students (n = 111), and university staff members (n = 86). They were asked to complete the test–retest measure in their places of exercise, study, or work respectively. Results: Correlation coefficients between test–retest administrations were calculated for each participant group and intraclass correlations were calculated for each item. Pearson's product-moment correlations ranged from r = 0.95 to r = 0.96 for the PPA subscale and r = 0.93 to r = 0.98 for the UPA subscale. Intraclass correlations ranged from R = 0.52 to R = 0.99 for PPA and R = 0.87 to R = 0.99 for UPA. Fisher's z tests indicated that the test–retest correlation coefficients for the BLPAQ subscales were, on the whole, significantly stronger than those of older, comparable subscales from lifestyle physical activity questionnaires. The PoA analysis for each item revealed that the test–retest administrations were in high agreement (>95%). Conclusions: Overall, the PPA and UPA factors of the BLPAQ demonstrated high reliability and stability. The present study also illustrates the utility of PoA analysis in establishing the stability of physical activity measures.