Citation

Fullarton C, Stokes M. Accid. Anal. Prev. 2007; 39(1): 28-37.

Affiliation

School of Psychology, Deakin University, 221 Burwood Highway, Burwood, 3125 Victoria, Australia.

Copyright

(Copyright © 2007, Elsevier Publishing)

DOI

10.1016/j.aap.2006.05.015

PMID

16950149

Abstract

Workplace injuries are common and destructive to persons, organisations, and society. Various instruments presently exist that are designed to assess the factors underlying workplace injury. The study reports on the construct and predictive validity of a 46-item instrument, the safety perception survey (SPS), currently used to assess safety climate in industrial organisations throughout Australia. Initially, factor analysis was conducted on the data from a sample of 1238 employees from nine organisations, which indicated a one-factor solution, was the best fit. A structural equation model (SEM) linking injury rates to the safety climate measure for 16 sub-groups of six industrial organisations indicated that the measure contributed just 23% of the variance in injury rates. Interestingly, the results indicated that the number of employees was a better and more significant predictor of injury (R(2)=0.48). It is proposed that the SPS as is would need to be modified significantly from its current form to produce improvements in validity, as in its current form the survey is no more predictive of injury than organisational size. Future research into safety climate measures should incorporate predictive validity analysis on injury rates, as for many organisations; this is a performance outcome measure.

Language: en