The test-retest reliability of habitual incidental physical activity

Objective: To determine the test-retest reliability of items measuring habitual incidental physical activity, incidental physical activity behaviour and sedentary behaviours. Methods: Eighty-four subjects aged 18-65 years were interviewed from randomly selected households within metropolitan and rural Western Australia. Subjects were administered the interview on two separate occasions 10 days apart. Subjects were asked about habitual incidental physical activity (i.e. bouts performed for 10 minutes or less) performed during the week and on weekends. Results: Intraclass correlations performed for items measuring incidental physical activity were generally found to be low to moderate. The frequency of incidental physical activity was more reliably recalled than the average duration spent on short trips (ICC 0.582 to 0.872 compared with 273 to 0.551). Kappa coefficients for habitual incidental physical activity behaviour items were observed to have moderate-to-high (0.478 to 0.939) agreement between administrations. Sedentary behaviours overall showed high test-retest reliability (ICC 0.646 to 0.925). Conclusion: These findings demonstrate that adults can reliably recall the frequency, but not the duration, of habitual incidental physical activity. Thus, the frequency measured by these items could be used to determine the effectiveness of current Australian physical activity guidelines and possibly identify shifts in incidental physical activity behaviour over time. Implications: Measurement of physical activity in all domains including incidental physical activity is required to determine the effectiveness of current guidelines and implemented health promotion interventions.