Exercising at work and self‐reported work performance

The purpose of this paper is to address the interplay of workplace exercising on self‐reported workplace performance.

A mixed methods design combined a randomised cross‐over trial with concurrent focus groups. Three workplaces (two private companies, one public service organisation) were purposefully selected for their provision of on‐site exercise facilities, size (>250 employees) and large proportion of sedentary occupations. Two mood diary questionnaires were distributed to employees exercising on‐site only. Order of questionnaire completion was randomised: self‐selected exercise‐day (ExD) or no‐exercise day (NExD) first. Exercise specifics (duration, intensity, mode) and ExD mood (pre‐/post‐exercise) were recorded. On NExD, mood was measured early and late in the working day. A 15‐item work performance grid was completed at day‐ends. Three on‐site focus groups were held concurrently to explore performance‐related topics.

Among 201 volunteer respondents (67 per cent female, mean age 38.2 years), mood improved on ExD, pre‐to‐post exercise (all p<0.01). Performance indicators were higher on ExD, versus NExD (all p<0.01), independent of exercise specifics and workload. Positive changes in performance outcomes were almost exclusively linked to changes in mood. Inductive analysis of focus groups revealed 13 (of 17) themes exhibiting positive outcomes. Employee tolerance and resilience were central to the subjective findings.

The naturalistic, dual‐paradigm study demonstrated that workday exercise can improve white‐collar workers' mood and self‐reported performance on days when they exercise at work over days when they do not. There are clear implications not only for employee wellbeing, but also for competitive advantage and motivation by increasing opportunities for exercising at work.

This is one of the few studies that addresses the acute effects of exercise in the workplace in the same people. Self‐rated productivity effects attributable to exercising during the working day were strongly mediated by changes in mood. Statistical power is amplified within the cross‐over design.