Criterion-related validity of the cultural web when assessing safety culture. (January 2019)

• Record Type: Journal Article
• Authors: Cooper, M.D.; Collins, M.; Bernard, R.; Schwann, S.; Knox, R.J.
• Is Part Of: Safety science. Volume 111(2019)
• Page Start: 49
• Page End: 66
• Publication Date: 2019-01
• Subjects: Safety culture assessment -- Cultural web -- Organisational safety -- Criterion-related validity -- Safety culture models; Industrial accidents -- Periodicals; Accident Prevention -- Periodicals; Safety -- Periodicals; Travail -- Accidents -- Périodiques; 363.11
• DOI: 10.1016/j.ssci.2018.09.013
• Languages: English
• ISSNs: 0925-7535

Abstract

Highlights: Researchers have struggled to develop an externally validated method of measuring safety culture. The cultural web proved to be a useful method for identifying, describing & improving safety culture. This study provided qualitative data and quantitative data focused on effectiveness. The cultural web exhibited high reliability and criterion-related validity with injury records. The study data was used to test three different theoretical safety culture models. Abstract: The purpose of the safety culture construct is to reduce organisational and occupational accidents. However, researchers have struggled to develop validated 'measures' of safety culture, that unequivocally link cultural traits with actual safety performance. Johnson's (1992) [Johnson, G., (1992). Managing strategic change—strategy, culture and action. Long Range Planning, 25 (1), pp.2] qualitative cultural web tool was adapted to simultaneously produce quantitative effectiveness ratings of an organisation's current safety arrangements for impacting personnel's safety-related behaviour. Data was collected at 15 safety culture workshops across North America over three-weeks. The population sample comprised 700 personnel, divided into 110 respondent groups. Data were examined from two perspectives: Within the cultural web topics (Routines, Stories, Symbols, Influences, Values, Structures & Measures); and specific safety culture topics (Profit before safety, Culture of Fear, Safety Leadership, Compliance, Competency, Communication, Lessons Learned) derived from thematic content analysis across the cultural web topics. The overall safety culture was shared and stable. Cronbach's Alpha (0.845) indicated reliability. Criterion-related validity between the organisation's Total Recordable Incident Rates (TRIR) for the cultural web topics (r = 0.488, p < 0.01) and specific safety culture topics (r = 0.417, p < 0.01) was found. Multiple regressions against specific incident records returned adjusted R 2 criterion-related validity coefficients between 0.06 and 0.45. Both perspectives confirmed the criterion-related validity of the cultural web tool, albeit stronger relationships tended to be obtained from the safety culture topics. The study results reinforce the conclusion that the tool is a reliable and valid method that can help companies reduce organisational and occupational incidents and improve their safety culture.