Document Type


Publication details

Sardoh, M 2013, 'Assessing safety culture in a high reliability organisation', DBA thesis, Southern Cross University, Lismore, NSW.

Copyright M Sardoh 2013


In the wake of a number of catastrophic safety related mishaps that have occurred worldwide in recent decades there has been increased interest in building a strong safety culture amongst high reliability organisations (HRO) as a way of immunising themselves against accidents in the workplace. It is thought that a strong safety culture will proactively ensure a strong safety performance in high reliability organisations. Although there is no shortage of measures of safety culture and safety climate they have yet to be causally linked to safety performance in the workplace.

One such measure of workplace safety culture is the survey questionnaire developed by the International Atomic Energy Agency (IAEA). Although the agency’s theoretical model of safety culture has been widely adopted by high reliability organisations, the model has never been subjected to empirical scrutiny.

This study set out to assess the face validity of the five characteristics and 37 attributes which comprise the IAEA model. The main research instrument was the IAEA survey questionnaire modified slightly with the permission of the IAEA. It was used to provide a quantitative safety culture assessment of a high reliability organisation. 252 completed returns were collected from the organisation's employees. Initial analyses (Pearson r) confirmed a significant and positive correlation between each of the five characteristics and an efficient workplace safety culture. While a subsequent regression analysis revealed a statistically significant level of support for the model only two of the characteristics (value and accountability) made a unique and significant contribution to the regression equation. The composite variable (integrated) seemed to be measuring something other than it was meant to. This lack of face validity along with excessive multicollinearity amongst the five independent variables served to weaken the IAEA model. In addition, a deductive thematic analysis of the more than 607 recommendations made by the respondents as to how their workplace safety culture might be improved showed that the five characteristics of the IAEA model could only absorb 560 of those recommendations. The inability of the model and its five characteristics to accommodate the remaining 71 (12 %) recommendations is seen as a serious flaw in the IAEA model.

Subsequently a series of post hoc analyses was undertaken beginning with an exploratory factor analysis in an attempt to capture the psychometric properties of the survey as perceived by the respondents. Analyses confirmed the importance of value and learning to an effective workplace safety culture. Three new components were extracted from the factor analysis including teams, direct supervisor and autonomy. Moreover, the new set of predictor variables for the concept of efficient workplace safety culture were found to be free of multicollinearity. Hence, this newer model is seen to be more robust than the IAEA model.

These new dimensions were subsequently shown to be positively related to employee beliefs, values and attitudes regarding the manner in which safety attributes were managed in their organisation. The theoretical implications of this new model are fully discussed including the conceptual blurring in the literature as to the nature and utility of the twin concepts of safety culture and safety climate. In the absence of general agreement regarding the conceptual definition of climate and culture it remains a challenge for future research as to how to operationalise these two constructs as a prelude to establishing their role in determining effective safety performance in the workplace.