Adequate information about physical activity habits is essential for surveillance, implementing, and evaluating public health initiatives in this area. Previous studies have shown that question order and differences in wording result in systematic differences in people's responses to questionnaires; however, this has never been shown for physical activity questions. The aim was to study the influence of different formulations and question order on self-report physical activity in a population-based health interview survey. Four samples of each 1000 adults were drawn at random from the National Person Register. A new question about physical activity was included with minor differences in formulations in samples 1–3. Furthermore, the question in sample 2 was included in sample 4 but was placed in the end of the questionnaire. The mean time spent on moderate physical activity varied between the four samples from 57 to 100 min/day. Question order was associated with the reported number of minutes spent on moderate-intensity physical activity and with prevalence of meeting the recommendation, whereas physical inactivity was associated with the differences in formulation of the question. Questionnaire context influences the way people respond to questions about physical activity significantly and should be tested systematically in validation studies of physical activity questionnaires.
Scandinavian Journal of Medicine and Science in Sports, 2013, Vol 23, Issue 5, p. 651-656