Tired, strained, and hurt: The indirect effect of negative affect on the relationship between poor quality sleep and work injuries

ABSTRACT

We conducted 3 studies to investigate how poor quality sleep relates to work injuries. First, using a sample of employed people living in the United Kingdom (N = 4,238; Study 1), we found that poor quality sleep was related to more frequent workplace injuries via negative affect rather than cognitive failures. Second, we again compared parallel pathways using a sample of USA employees (N = 202; Study 2): poor quality sleep was related to more frequent work injuries via work-related negative affect but not work-related cognitive failures. Third, we used a 2-wave sample of employees from the United Kingdom (N = 71; Study 3) finding that poor quality sleep was related to more frequent work injuries 7 weeks later via negative affect. Comparing high arousal and low arousal negative affect as competing pathways showed that there was a significant indirect effect of the former on the poor quality sleep-work injuries relationship but not the latter. Across 3 studies, we implicated the role of self-control failure stemming from poor quality sleep in predicting more frequent work injuries and suggested initiatives targeting high arousal negative affect as a way of reducing work injuries.

KEYWORDS:

• Cognitive failures
• injuries
• negative affect
• safety
• sleep
• strain

Acknowledgement

We thank Toon Taris and our anonymous reviewers for their time in providing constructive feedback. We also thank Sharon Clarke, Mark Griffin, and Tahira Probst for providing feedback on earlier versions of this paper, which were presented at the 6th Biennial Conference of Work, Well-being, and Performance, Sheffield, UK. Amanda Harrison and Peter Totterdell generously provided access to the unpublished data used in Study 3. The first author was supported by a Canadian Institutes of Health Research Fellowship.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 Analysis was conducted with all four GHQ negative affect items as four parallel indirect effects, and the CI for “Constantly under strain” was the only indirect effect that did not include zero. We decided on the two parallel indirect effects model for a more succinct result section.

2 According to modern perspectives on indirect effect analyses, the predictor variable need not be significantly associated with the criterion variable to proceed with testing for indirect effects (Hayes, 2009; Preacher & Hayes, 2008). The total effect encompasses paths of influence not included in the tested model, some of which may be opposite in sign of the indirect effect being tested, thus cancelling out each other in the summation of the total effect.

Additional information

Funding

This work was supported by Canadian Centre for Advanced Leadership in Business; Social Sciences and Humanities Research Council, and the Health and Safety Executive also provided financial support for this study.