METHOD

This study was conducted in three phases. Phase I involved the development of an interview protocol to assess the safety practices of participating organizations. Phase II involved identifying construction companies with high and low injury claim rates in Nova Scotia, Canada. Phase III involved conducting interviews with a sample of the companies identified in phase II.

Phase I: Interview Protocol

In order to investigate the organizational practices that are associated with effective safety management in small construction companies it was necessary to develop an instrument to assess these practices. Table 3.1 summarizes the organizational practices that distinguish low accident companies from high accident companies identified in previous research. These eight practices were used as the starting point for developing the interview protocol for the current study. We increased the number of indicators assessed with the interview protocol from eight to 11 by specifying the target group for two of the practices. Specifically, the practice of providing safety training was expanded by separating out training for: (1) managers, (2) supervisors, and (3) workers. Furthermore, the practice of safety performance evaluation was expanded to include: (1) managers’, and (2) supervisors’ safety performance evaluations. Therefore, the interview protocol included the following eleven sections:

• 1. Hazard assessment
• 2. Incident investigation
• 3. Worker involvement in safety
• 4. Manager safety training
• 5. Manager performance evaluation
• 6. Manager safety inspections
• 7. Supervisor safety training
• 8. Supervisor performance evaluation
• 9. Supervisor safety inspections
• 10. Safety communication
• 11. Worker safety training

The comparison of small construction companies on these organizational safety practices requires a simple framework to classify performance on a continuum from poor to good. There are a large number of models that classify company performance on such a continuum, or by level of maturity. One of the earliest maturity models was the Quality Management Maturity Grid developed by Crosby (1979). This typology consisted of five levels: (1) Uncertainty, (2) Awakening, (3) Enlightenment, (4) Wisdom, and (5) Certainty, to describe an organization’s approach to quality management. The maturity model framework became popular after the very successful Capability Maturity Model was developed for software (Paulk et al., 1993). More recently there has been significant interest in safety maturity frameworks (Ashcroft et al., 2005; Fleming, 2000; Lawrie et al., 2006; Parker et al., 2006). Safety maturity frameworks build on the fundamental assumption that safety can range from poor to good. Maturity models also aid in identifying an organization’s areas of particular strength or weakness (National Patient Safety Agency, 2006), and what actions need to be taken to improve (Paulk et al., 1993).

Westrum’s (1984) typology of an organization’s culture identifies three basic styles of organizations: pathological, bureaucratic and generative. Pathological organizations are described as environments where there is a focus on personal needs, power and glory. Bureaucratic organizations are described as environments where there is a fixation with rules, positions and departmental territory. Generative organizations are described as environments where there is a focus on the mission, rather than on persons or positions (Westrum, 2004). Westrum (1996, 2004) proposes that this typology can be used to categorize organizational culture. Specifically, Westrum argues that in pathological cultures, information is only important if it will affect personal interests. In bureaucratic cultures, information is only used to advance the goals of the department, and in generative cultures, emphasis is placed on using information to aid in accomplishing the mission (Westrum, 2004).

Westrum (1996, 2004) also proposes that this typology can be used to characterize organizations on how they respond to failure. For example, in pathological organizations, information is hidden and failures are dealt with by blaming a scapegoat. In bureaucratic organizations, information may be ignored and failures are explained away or resolved, with no deeper inquiry into them, whereas generative organizations actively seek information and inquiries occur after failures are discovered, which serve to attack the underlying conditions, not just the immediate causes of the failures.

Reason (1993) adapted Westrum’s tripartite typology, so that it applied to occupational safety by expanding the typology to five and modifying one element. The two additional levels are reactive and proactive. Reactive organizations state that safety is important to them, but respond only after accidents have occurred. Proactive organizations try to anticipate safety issues before they happen (Reason, 1998). Reason (1993; 1998) also modified Westrum’s model by renaming bureaucratic organizations to cal- culative organizations, as it provided a better description of how safety is managed within these organizations. Therefore, Reason’s organizational typology describes five levels of safety maturity: pathological, reactive, calculative, proactive and generative (see Table 3.2).

Although maturity models describe different approaches to managing safety on a continuum from poor to good, they are not necessarily stage models, which require organizations to progress through each level in sequence. These frameworks facilitate making judgements about the sophistication of an organization’s approach to safety management. At the pathological level organizations are not working to ensure worker safety, the reactive level is typified by a focus on compliance with regulation, the calculative level involves using proven methods to manage common hazards, the proactive level involves anticipating and managing infrequent hazardous situations, and the generative level involves an innovative and comprehensive approach to safety management.

Reason’s (1993, 1998) framework provides a useful structure for the interview protocol, as it allows not only for the evaluation of what safety practices are currently present in an organization, but it also provides a

Table 3.2 Safety culture maturity framework

basis for evaluation of the degree of sophistication of each safety practice. Therefore, the 11 organizational practices were mapped onto the maturity framework outlined by Reason (1998), resulting in five possible levels of sophistication for each practice. An example of one of the eleven sections of the interview protocol is presented in Table 3.3.

Table 3.3 Safety practice example from interview protocol: management’s involvement in safety inspections