Date of Award


Document Type


Degree Name

Doctor of Philosophy (PhD)

Legacy Department

Industrial and Organizational Psychology


Britt, Thomas W

Committee Member

Moore , D. DeWayne

Committee Member

Sinclair , Robert R

Committee Member

Tafero , Tracey


The current study explores the role of selection in predicting workplace safety using an applied sample of applicants and incumbents in a grocery store chain located in the Southeastern United States. Namely, both personality-based and overt selection assessments, a distinction drawn from the integrity testing literature, were used to predict on-the-job safety performance and safety outcomes. Both types of assessments were hypothesized to predict two forms of safety performance (compliance and participation), which, in turn, were expected to predict both objective (i.e., work days missed, restricted work days, and micro-accidents) and subjective (i.e., near-miss, minor injuries, and musculoskeletal pain) safety outcomes.
The mediated relationships (personality-based and overt variables predicting safety outcomes through safety performance) could not be tested with objective safety outcomes as the dependent variables due to low sample size. When the subjective outcomes were the dependent variables, this hypothesis was only supported when MSK pain was the outcome; indirect relationships with minor injuries or near-misses as the dependent variables were not significant. None of the direct relationships between the personality variables and the safety outcomes were significant. The second set of hypotheses proposed the same mediated relationships with the two overt safety variables as the predictors. Neither direct nor indirect hypothesized relationships reached statistical significance.
The hypothesized relationships between the selection assessments and safety performance were also theorized to be moderated by safety climate strength, which is the degree to which employees view the company and its practices and policies similarly (Siehl & Martin, 1990). A strong climate was expected to weaken the predictor-mediator relationship because strong situations, which provide many cues about how to behave, decrease individual discretion and foster behavioral homogeneity (Hattrup & Jackson, 1996; Meyer, Dalal & Bonaccio, 2009; Mischel, 1977).
The results of these hypothesis tests indicated that the interaction between any of the personality variables with safety climate strength did not uniquely predict safety performance. Likewise, the interaction between both overt safety variables and safety climate strength did not significantly predict safety performance.
Exploratory analyses suggested that average safety climate was a strong predictor of safety performance, accounting for over 39% of the variance in this outcome after controlling for demographics and group membership. Further, safety climate strength was also significantly related to safety performance above and beyond the effects of safety climate average. Safety performance and average safety climate were predictive of all subjective safety outcomes (near-misses, minor injuries, and MSK pain). Limitations and practical implications of the current study are discussed.

Included in

Psychology Commons