Objective: The aim of this study was to evaluate the reproducibility of a self-administered questionnaire about present and past physical activities at work and during leisure time.
Methods and design: The questionnaire, covering the period 1970-1993, comprised 12 questions on physical activities at work, and 12, with similar wording, for such activities in leisure time. There were also four questions on physical training. Two-week reproducibility (test-retest reliability) concerning the period 1970-1993 was analysed in a group of 44 subjects, and 1-year reproducibility, concerning current activities in 1993, was analysed in a second group of 123 subjects in relation to gender, age and low-back health.
Results: Test-retest reliability calculated as intraclass correlation coefficients (ri) for physical activities at work (ri 0.41-0.98) exceeded that for leisure time and physical training activities (ri 0.33-0.68). Calculated correlations did not differ markedly between past and present activities. No distinct influence of gender or low back health on 1-year reproducibility was found, in contrast to a slight tendency towards higher reproducibility among subjects of 50 years and older compared with younger subjects.
Conclusions: Reproducibility of this questionnaire about physical activities at work showed no clear tendency to deteriorate regarding activities during the immediately preceding two decades. The questions about physical activities during leisure time have to be revised. Reliable retrospective information about physical activities in leisure time could perhaps not be collected by self-administered questionnaires and other methods, e.g. interview-based questionnaires, may be more suitable.