Safety Behaviour in the Construction Industry

While there is a good understanding of the extent and patterns of accidents in the construction industry, there is only limited evidence about the full range of factors contributing to those. From a psychosocial point of view, the available literature on safety on construction sites is relatively scarce when compared with other industrial sectors in our society. Important areas of interest when predicting individual safety behavior and organizational safety performance are those related to compliance with procedures, effective management systems, organizational climates and cultures, and performance in the face of risks.

A review of the relevant literature addressing those areas follows. A selection of the most important research and conclusions that have guided the theoretical frame of this project is offered to outline the reasons behind this research. The main contributions to highlight from the literature review is the choice of compliance rather than safety accidents as our dependent variable, and the selection of those factors that seems to influence safety behavior and compliance on site: individual factors, site factors and also factors outside the site boundaries.

Non-compliance versus accidents as measure of safety performance

All interventions oriented to improve safety in the workplace have accident reduction as the ultimate objective and so the criteria against which to measure the effectiveness of such interventions. According to Duff et al. (1993), accident frequency is "the most objective measure of safety performance." However the scientific literature has highlighted that an accident is, to some degree, a chance event requiring a combination of circumstances. Additionally, a HSE (1988) investigation has shown that accidents, especially less serious ones, are subject to gross underreporting. Managers may also be careful in releasing information that could affect the future stability of their business or be used in any way against them.

Alternatively, better safety performance of a business, or progressive improvement of safety when a successful intervention has been carried out, brings the accident rate to a lower level than before, increasing the difficulties of identifying it accurately. This also decreases the accuracy in measuring the amount of improvement due to the intervention over time (Duff et al., 1993).

The combination of above features makes accident frequency difficult to measure in any research. Whatever the objective characteristics, it can be said that accident frequency cannot be considered to be a robust measure for research purposes.

Duff et al. (1993) designed an alternative measure of safety performance based on compliance with safety procedures in particular situations or by means of individual behavior. The authors validated an observational tool that turned up to objectively and reliably measure safety performance in percentages of compliance. This tool was composed by 41 items representing critical safety standards that offer data appropriate to be used in many different safety management strategies, such as performance auditing, training design, incentive system design, and goal-setting (Robertson et al, 1999).

Safety consequences of noncompliance with procedures

A key defense against hazards being realized in high-risk industries is the proceduralization of tasks - defining key rules and procedures which need to be followed to ensure that the work is done both efficiently and safely. One very promising line of inquiry concerning the behavioral antecedents of accidents concerns the relationship between these procedural instructions governing work and the way in which work is done. Not following procedures has been implicated as a contributory factor in incident and accident occurrence in a wide range of industries. Some of these studies have provided measures of the normative level of procedure violations, which varies from under 20 to over 50 percent of recorded tasks. In the construction industry in the U.K., Duff et al. (1993) found percentages of noncompliance with specific categories in 6 construction sites ranging from 22-38% of noncompliance in housekeeping, from 12-43% of noncompliance in scaffolding, from 20-26% of noncompliance in access-to-heights, and from 21-65% of noncompliance in using PPE. Also from the U.K., Robertson et al. (1999) found that noncompliance was around 19% for housekeeping, 16% for scaffolding, 15% for access-to heights, 21% for P.P.E., and 13% for plant. Lingard et al. (1997) reported from Hong Kong percentages ranging from 30-49% of noncompliance in housekeeping, from 30-66% of noncompliance in bamboo scaffolding, from 50-74% of noncompliance in access-to-heights, and from 49-69% of noncompliance in PPE.

The importance of this work is that it provides a mechanism for relating observed or reported behavior with an identifiable safety standard (the official procedure). This has opened up productive avenues for researching into the situational factors (task, organization, resources, individual), which make such violations more likely. Not all violations are serious from a safety point of view (sometimes there are more efficient ways of working). However, it is possible to identify violations that could result in serious consequences.

Interventions to improve safety

One major goal of safety research is to support interventions that can demonstrably improve safety. However, thorough empirical studies are rare and they normally focus on changing individual behavior.

Duff et al. (1993) and Robertson et al. (1999) carried out a two - phased study on the effectiveness of different intervention strategies. Specifically, they looked at the effects of feedback, goal setting and training, on safety performance. They found that the techniques used produced marked improvements in site safety, participative goal-setting being the more effective of the three. However, a continuous and consistent intervention along the lifecycle of the site was recommended to achieve the maximum benefit. The results also highlighted the vital role of management commitment in the effectiveness of the intervention methods used.

Lingard and Rowlinson (1998) used a similar design to Duff et al. (1993) in the Hong Kong construction industry. They reported highly significant improvements in housekeeping with their intervention. However, in general, the intervention did not result in significant improvements in the areas of access to heights or bamboo scaffolding. Again, the main reason was the management commitment towards those two areas of activity. They concluded in relation to that in "behavior-based safety management programs... safe behavior can only be achieved where a basic safety infrastructure is already in place" (p. 225).

Other variables of interest, such as attitudes and safety climate, have received little attention in the literature. For example, Langford et al. (2000) carried out a research on safety attitudes in construction workers. This identified some variables that influenced the attitude of construction workers towards safe behavior: organizing for safety supervision and equipment management, industry norms and culture, attitudes to risk taking, and management behavior. However, there are no data available for the influence of attitudes on behavioral or site performance. In her degree project, Curran (2000) researched safety attitudes and the safety climate on one construction site. However, the results cannot be generalized. To our knowledge, this is the only research on safety climate in the construction industry.

In relation to management strategies, Tam and Fung (1996), for example, looked at the effectiveness of the safety management strategies of 45 construction firms in Hong Kong. They report that most safety schemes, strategies or interventions reduced the accident rates on these sites. These schemes strategies and interventions included:

• Level of management responsibility
  
• Orientation programs
  
• Safety personnel on site
  
• Safety awards and incentive schemes
  
• Post accident investigation and feedback
  
• Safety training and intensity of safety training
  
• Presence of safety committees

And from the site of the regulatory authorities, Berg (1999) reports on a safety campaign carried out in 9 regions in Germany during 1995 and 1996 by 50 inspectors from the relevant construction site regulatory authority. This same authority without documenting any methodologies or indicators of effectiveness other than re-inspecting sites, reports the results of this campaign as follows:

• A conviction that a substantial increase in safety and health within the region and on sites involved in the campaign has occurred.
  
• An opinion that the campaign helped in raising awareness of the severity of, and hence importance of falls.
  
• A conviction that the campaign identified fall hazards on sites to site management.
  
• An opportunity to provide solutions for contractors in the area of fall prevention.
  
• Confirmation of a partnership approach to safety and health with the participating firms.

In brief, construction has not been subject to much intervention research, largely because of the complexity of the industry (Ringen et al., 1996). Of the interventions carried out, behavior-based programs are the most popular when attempting to change safety performance. Although positive and significant results are generally obtained, interventions can be said to be of limited success. Management commitment is identified every time as a key element in guaranteeing the successful implementation and completion of a program. However, intervention at managerial level is rare and studies carried out, while highlighting the important role of management, are exploratory and descriptive (for example, Whittington et al, 1992), rather than intervention programs to change safety performance. Safety campaigns are directed to increase worker awareness and said to be effective in improving safety performance, but systematic research suggests that posters and campaigns have limited success (Saarela et al, 1989, quoted by Duff et al. 1993), and do not make a lasting impact on the accident/injury rate. Variables such as climate, culture and attitudes, found to be of importance for safety performance in other industries, have yet not been researched in construction in a systematic way.

Functional effectiveness of safety management systems

A second level of explanation for safety failures concerns the effectiveness of the safety management system. To quote three examples, Whittington et al (1992) comment that the construction industry characteristics that lead to the poor safety record are deep rooted and complex. They go on to state that problems at site level could often be traced back to management issues such as poor contractor selection, lack of supervision or inadequate training. The European Construction Institute (ECI, 1996) reports that one of the reasons for the industry's poor safety and health reason has been the lack of a systematic approach to the management of risk. Mohamed (1999) also reports that the construction industry in general seems to suffer from an inability to manage workplace safety and health to an extent where a proactive zero accident culture prevails.

Whittington et al. (1999) indicated how the industry and potential clients were also seen to be responding to safety management demands in a fundamentally flawed manner by way of:

• Being dealt with at a late stage of the project cycle
  
• Undue emphasis on the failure of individual workers resulting in short term measures rather than resolving underlying organizational problems.
  
• Competitive tendering resulting in a failure to address safety requirements at bidding and tendering stages.
  
• Safety issues being inadequately addressed in planning and scheduling of work.
  
• Lack of safety performance monitoring and feedback.
  
• A lack of opportunities for formal project reviews to include organizational learning and proactive safety management.

In order to reduce the level of fatalities injury and ill health in the construction industry, a number of safety management systems are available detailing the elements currently thought necessary to reduce risks. Publications like Successful Health and Safety Management (HSE 1997, BS8800: 1996; OHSAS: 1999; OHSAS, 2000) give good practical guidance on the requirements for effective safety management.

A number of construction specific safety management systems and codes of practice are documented. The HSE in the UK have produced a guidance document; Health and Safety in Construction. (HSE, 1997a). The HSE have also published a Code of Practice on Compliance with the UK Construction Design and Maintenance Regulations 1994 (HSE, 1994a). The HSA have produced guidelines for the equivalent Irish Legislation (HSA, 1995). It has also published a Code of Practice for Access and Working Scaffolds (HSA, 1999), and it is producing two new codes of practice on roof work and crane usage (HSA, 2001a). The Construction Industry Federation in Ireland (CIF, 2001), the European Construction Institute (1996) or authors as Clarke (1999) or Gibb (1995) also details health and safety management systems and risk assessment procedures that can be used in the construction industry. Irrespective of the number of guidelines available, research into the elements of the proposed management systems that are effective in the construction industry, still needs to be conducted.

However, a number of studies do link management commitment with particular findings, especially effectiveness of the intervention program. For example, Duff et al. (1993) concluded that management commitment was an important moderator of the effectiveness of the intervention carried out. Six years later, in phase two of their research Robertson et al. (1999), concluded that management commitment to the intervention program was "vitally important, impacting on all aspects of the methodology." Positive correlation appeared between management commitment and safety performance improvement on site, and between management commitment and management participation in all stages of the intervention.

Summary

While accident frequency is an objective measure in safety-related researches, measures based on compliance with procedures seems to be more appropriate for comparison between sites as is identifying change when intervention programs are implemented. For the construction sector, Duff et al. (1993) have designed and researched a compliance-based measure that has been shown to be valid and reliable. Therefore, it has been adopted for this research.

Studies suggest that a compliance with safety procedures varies widely across categories of procedure and in different sites. For example, rates of noncompliance with scaffolding requirements between 12% and 43% have been reported. Evaluations of a range of interventions to improve safety have addressed goal setting, training and other behaviorally based programs. While some of these have demonstrated success in the short term, a common conclusion emphasizes the importance of sustained management commitment and activity in ensuring the effective implementation of these programs. However, there is a lack of systematic research concerning the management of safety in construction. Also, while there have been some studies of attitudes and perceptions of construction workers towards safety and safety management, they have not addressed the link between these variables and safety behaviors and compliance with safety procedures.

Therefore, it can be concluded that the factors, at both individual and management level, which influence sustained compliance with safety requirements are not well understood. This study aims to address this critical gap in our knowledge about construction safety.

Back to Contents