Not working to rule: Understanding procedural violations at work

Abstract

This paper begins by presenting a brief synopsis of the literature regarding the relationship between rule violations and accidents. The paper goes on to report a study of UK railway shunters' motives for rule violations. Violations are defined as behaviours that involve deliberate deviations from the written rules. Preliminary investigation elicited the motives to be included in the main questionnaire survey which required 36 shunters to rate the importance of various motives for violating. Generally, violations were perceived to be the result of a well-intentioned desire to get the job done. Together with previous analysis of shunting accidents, observations of, and discussions with shunters in the workplace, the results of this study were used to develop a classification of violations which includes situational, exceptional and routine violations. Erroneous behaviours that also involved deviations from rules were recorded as a separate category, namely unintentional violations. The results of the study, together with the findings of other research in this area, are used to describe a model for the investigation of violating behaviour in an organizational setting. This model depicts the factors that promote violations at work and, as such, may aid managers and health and safety practitioners to develop appropriate preventative measures.

Introduction

In an attempt to understand and minimise the role of human error in any given behavioural domain, it is important to be able to answer a number of questions. Firstly we need an adequate description of what sorts of behaviours are leading to negative outcomes and then, if we are to offer sensible advice on possible routes to accident reduction, we need to determine why people behave in this way.

The Chernobyl disaster, culminating in two explosions, killing 30 people, contaminating land and producing serious health risks was the worst accident in the history of nuclear power plants. In response to this accident, Reason (1987)reviewed the human factors inherent in the series of events which lead to the explosions. This was an unusual accident, he writes “it appears to have been entirely due to human action” (p. 201), and as such requires that psychologists offer an explanation.

Reason highlights those actions, more and less serious, which resulted in the explosion. Of the seven operator actions perceived as causative, five involved deliberate deviations from written rules and instructions, rather than errors in the form of slips, lapses and mistakes. Reason concludes that “one of the important lessons of Chernobyl, as well as of the Challenger and Zeebrugge disasters is that by itself, a purely cognitive analysis of error mechanisms fails to capture some of the more important human contributions to catastrophic system breakdowns” (p. 204).

Violations could not be defined merely in relation to the information processes of the individual and necessitated a description in terms of the social context and the rules and procedures which regulate behaviour. Reason also noted that one could err without violating and one could violate without erring meaning that “both notions are necessary for an adequate taxonomy of aberrant behaviours” (p.206). In other words, attempts to understand safety-related behaviours requires that the study of errors, in the form of slips, lapses, and mistakes, coincides with the study of violations, defined here as deliberate departures from rules that describe the safe or approved method of performing a particular task or job.

Whilst a small proportion of the population will comply with a rule, simply because it is the rule, the majority will not, because compliance usually involves expenditure of time and effort. As behaviour is governed to some extent by an analysis of the costs and benefits involved in acting in a particular way, Battman and Klumb (1993)discuss rule compliance in terms of behavioural economics. They support the study of violations in a social rather than a cognitive context. The fact that violations persist, they write, depends on social forces, on the existing safety culture, the shared attitudes, beliefs, norms and practices of a particular organization. If violations are contingent on the existing safety culture, this provides further grounds for a distinction from errors. Errors, based on cognitive failures will be largely unaffected by the attitudes, beliefs and values of the individuals at the sharp end.

In a study of aberrant driving behaviour (Reason et al., 1990) the error classification was extended to include not only slips, lapses and mistakes, but also violations. Recognising that the previous classification had failed to account for the fact that most people plan their action in a social milieu, the research team devised a Driver Behaviour Questionnaire (DBQ), which included items varying in two respects: the type of behaviour involved and the amount of risk involved. The behavioural types were unintentional violations, violations, mistakes and slips. The risk categories included, no risk to others, possible risk to others and definite risk to others. A three-factor solution discriminated both of these variables. Factor 1 and factor 2 were both related to dangerous behaviours, yet factor 1 consisted almost exclusively of violation items, while factor 2 comprised a mixture of errors. Furthermore, factor 3 consisted of slips and lapses which were unlikely to cause danger to others.

The factor analysis suggested a distinction between errors and violations. This was supported by sex and age differences with respect to errors and violations: violations decreased with age, errors did not; men of all ages reported more violations than women, while lapses were reported more frequently by women than men. In addition skill was found to be related to violations, so that subjects who reported more violations tended to rate themselves as particularly good drivers.

Having distinguished between errors and violations, Reason (1990)goes on to discuss the latter and offers a preliminary classification. Violations vary according to intentionality and outcome. In the case of sabotage or terrorist action, the act is intended and so is the bad outcome. At the other end of the continuum the violation (act) itself may not be intended and as such it is labelled an erroneous or unintentional violation. The focus of this paper are the violations that fall between these two extremes, those which are deliberate deviations from rules, but where the bad outcome is unintended. Reason (1990)also divides these violations into two categories, routine and exceptional. The former are habitual behaviours that may become an integral part of a skill, e.g. speeding on a motorway. Exceptional violations, on the other hand are those that occur in a particular set of circumstances and as such are rare, e.g. jumping into a vessel containing lethal gases to rescue a colleague.

The significance of violations in accident causation has been demonstrated repeatedly in accident analysis. Mason (1997)reports that violations of rules and procedures have been estimated by some industries to be a significant contributor to about 70% of their total accidents. In an examination of railway accidents in Britain between 1989 and 1992 Free (1994)revealed that violations play a considerable role in accidents to staff, e.g. personal injury and fatalities. There are several possible reasons why violations increase the likelihood of an accident. Firstly, some violations take people outside the boundaries of safe working practice, making the environment less forgiving to errors. In other words, violations circumvent one layer of defense, the rule book, which aims to ensure predictable and safe working practice. Secondly, violations can themselves be errors when the individual does not know or understand the rule. This lack of understanding is dangerous in itself, because not realising the risk, people often fail to protect themselves. Finally, violations can take people into new or unpracticed situations, in which the person is more likely to make an error.

A recent prospective study of British motorcyclists, Rutter et al. (1992)found that the best predictors of accident likelihood were age, sex and experience. In addition, the best attitudinal predictor of accident involvement was a willingness to break the law and the rules of safe riding. The authors suggest that the demographic variables predict behaviour directly, but they also have their effects upon outcome, via attitudes.

Other evidence for the role of violations in accident causation comes from work in the field of driver behaviour. Parker et al. (1995a)surveyed 1656 drivers and using the Driver Behaviour Questionnaire examined the relationship between driving behaviour and accident involvement. This paper describes results showing that self-reported violations, those behaviours that involve deliberate deviations from safe driving practice, were found to be predictive of accident liability. The self-reported tendency to make errors (e.g. misjudge the speed of an oncoming vehicle) or to have lapses (e.g. attempt to drive away from a traffic light in third gear) did not predict accident liability. Violations were found to be a statistically significant, positive predictor of accident involvement, even after the effects of exposure, age and gender had been partialled out. These findings have been replicated in a prospective study of the relationship between behaviour and traffic accidents (Parker et al., 1995b).

The fact that violations are often the precursors of accidents has obvious implications for the management of safety. Skills training, a popular response to undesirable safety statistics, will reduce errors by improving competence, but will have a very limited effect on the reduction of violations that have an attitudinal basis. Until recently, violations have not been conceptually distinguished from errors. Invariably, the managerial emphasis has been to ensure that rules are available, correct and comprehensible, the assumption being that this will guarantee compliance, because if the procedures are understood, errors will not arise. Health and safety professionals have tended, therefore, to focus on the procedures and rules themselves, expecting safe working practices to follow.

While it is essential that procedures are user-friendly, this is unlikely to be sufficient to ensure compliance. The investigation conducted by Zeitlin (1994)of why people fail to follow safety procedures, considered whether it was due to faulty communication or risky decisions. Zeitlin concludes that “if a worker places high value on his or her time, convenience, self-image, status among peers, and other factors, and if he or she estimates the probability of being injured by disregarding an instruction as sufficiently low, then it is likely that the instruction will be ignored despite its clarity of presentation” (p. 179).

Consequently, the intentional component of noncompliant behaviour is worthy of further consideration. Violations that are deliberate deviations from the rules require different remediation from those that result from unfamiliarity or inexperience. Training in the rules, hazards and risks of the job may be appropriate for the latter, but less so when violations are deliberate.

The study, commissioned by British Railways and reported below, investigates the violations of one particular staff group, shunters, in detail.