Percentage-method improves properties of workers' sitting- and walking-time questionnaire

Abstract

Background: Does asking for the percentage of time spent sitting during work (P-method) instead of asking for the absolute length of time spent sitting (T-method) improve properties of the workers' sitting- and walking-time questionnaire (WSWQ)? The purpose of this study was to investigate whether questioning technique influences testretest reliability and criterion validity of the WSWQ. Methods: Sixty-five Japanese workers completed each version of the WSWQ in random order. Both questionnaires assessed quantities of time spent sitting or walking (including standing) during work time, non-working time on a workday, and anytime on a non-workday. Participants wore the thigh-worn inclinometer (activPAL) as criterion measure. Intraclass correlation coefficients (ICC) and Spearman's ρ were used for the analyses. Results: For all three domains, values of reliability and validity with the P-method tended to be higher than with the T-method: ICC values ranged from 0.48-0.85 for the T-method and from 0.71-0.85 for the P-method; Spearman's ρ values ranged from 0.25-0.58 for the T-method and from 0.42-0.65 for the P-method. The validities with both methods on a workday (0.51-0.58 for the T-method and 0.56-0.65 for the P-method) were higher than validities on a non-workday (0.25-0.45 for the T-method and 0.42-0.60 for the P-method). In post-survey interviews, 48 participants (77%) chose the P-method as their preferred questioning style. Conclusions: The study revealed that the P-method WSWQ had better reliability, validity, and ease of answering than the T-method, suggesting that the P-method can improve properties of the WSWQ and consequently advance the quality of epidemiological surveys in this field.