Ensayo

Safety Science 41 (2003) 575–590 www.elsevier.com/locate/ssci

Safety diagnosis criteria—development and testing
Ranveig Kviseth Tinmannsvik*, Jan Hovden
Norwegian University of Science and Technology, Department of Industrial Economics and Technology Management, N-7065 Trondheim, Norway

Abstract Criteria for evaluating occupational safety in industrial organisations have been developedand tested. These criteria were designed for use in safety diagnoses, i.e. planned and systematic investigations of the organisation and the administrative procedures to control safety. The safety criteria were derived from previous performed safety management analyses and safety audit tools. A questionnaire survey was carried out in order to validate the criteria for evaluating safety in industrialorganisations. Production personnel, safety professionals and line managers from 14 companies (aluminium production plants and meat production industry) answered the questionnaire. Strong correlation were found between the respondents’ overall evaluation of safety and the evaluations of safety specific factors (management factors mainly to promote safety), while the injury frequency rate wasstrongly correlated to the evaluations of general management factors (management factors to improve the production system and organisation in general). The overall evaluation of safety correlated significantly with the injury frequency rate. This indicates that the proposed Safety Diagnosis Criteria can provide a valid measurement of safety in industrial organisations. # 2002 Elsevier Science Ltd. Allrights reserved.
Keywords: Safety audit; Safety management; Diagnosis criteria; Internal control

1. Introduction The main goal for this empirical study was to develop and validate a set of safety diagnosis criteria for evaluating safety in industrial organisations (Tinmannsvik, 1991). These criteria are to be used in safety assessment and control by the line management of the productionsystems. The safety diagnosis criteria were derived

* Corresponding author. Tel.: +47-73593581; fax: +47-73592896. E-mail address: ranveig.k.tinmannsvik@iot.ntnu.no (R.K. Tinmannsvik). 0925-7535/02/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved. PII: S0925-7535(02)00012-7

576

R.K. Tinmannsvik, J. Hovden / Safety Science 41 (2003) 575–590

from previous performed safetyanalyses, and were developed within the framework of the so called SMORT method (Safety Management and Organisation Review ´ ´ Technique; Kjellen and Tinmannsvik, 1989; Kjellen, 1992, 2000). The empirical study thus provides a validation of the SMORT method. The main question to be studied was; Are the safety diagnosis criteria valid as safety measurements? More specific hypotheses regardingvalidity were: 1. There is a significant correlation between the overall evaluation of safety, and the injury frequency rates. 2. There is a significant correlation between evaluation of safety management factors and the injury frequency rate. 3. There is a significant correlation between evaluation of safety management factors and the overall evaluation of safety. Safety analyses and safety managementhave become more and more important. The MORT concept (Management Oversight and Risk Tree; Johnson, 1980) represented a new approach to the scientific management of safety, and has formed a basis for further development of safety analyses and safety assurance methodology in industry. MORT is a logic tree that provides a disciplined method of analysing an accident, and provides a format for safetyprogram evaluation. The SMORT method is a further development of this systematic way of investigating safety in organisations intended for use in safety audits, as well as in accident investigations. SMORT provides a systematic and step wise means of unfolding relevant causal factors by starting with identification of risk influencing factors at the workplace level and proceeding through the various...