Human error (and violations)

Introduction

People get things wrong all the time. Even the most straightforward task, such as reading numbers from a display, is not error-free.  An error is something that gives us information about how near our goal we are. Errors are at the heart of our ability to adapt to, and master, new situations. On the other hand, some errors have such dire consequences compared with their learning value that they simply need to be prevented wherever possible.   

Before 1947, investigations of military aviation accidents had concluded that pilot errors were the cause of crashes. But then two psychologists, Fitts and Jones, looked more closely at what pilots did in the cockpit. They realised that the design of the instruments and controls was producing misreading’s and actions that had never been intended by the designers. The pilot errors were not random events. Rather they resulted from understandable, regular and predictable aspects of the designs they were faced with. What is more, the errors occurred much more often than accidents did. Significantly, disasters and near misses usually occurred only when these human errors occurred in combination with other factors or other circumstances.  So, whilst it is important to understand human error, it is more important to understand the underlying causes which may relate to job/workplace and organisational factors such as equipment design, fatigue or the quality of written procedures.   

Relevance to rail

Railway systems rely on the actions of staff, for example, train drivers, signallers, maintainers and train dispatchers. Railway systems also transport passengers. All of these people may commit errors of various types and it is therefore key to consider error management and design railway systems to be resilient to human error. 

Approaches and models

In this section we introduce human error types, before discussing violations or intentional rule breaking. In each section some pointers on how to manage errors and violations are also presented. It should be noted that there is currently debate around the use of the term human error and that performance variability may be a better and less value-laden alternative (see Read et al reference below). 

Human Error  

The most common categorisation within a HF framework is the Generic Error Modelling System (GEMS). This is based on the belief that different errors occur depending on three types of performance:  

• Skill-based performance – where we routinely perform highly practised activities with little conscious effort, e.g., replacing a rail clip or setting a route on a signaller’s panel
• Rule-based performance – where we have more mental involvement and apply previously learned rules to tasks, we have usually been trained for, e.g., planning a route in a signal box before setting it, talking a driver past a red aspect, or evacuating a station
• Knowledge-based performance – where we have even more mental involvement, often in novel situations, e.g., attending an accident scene, or counselling a staff member 

Errors are not random events. Rather, they are a consequence of what normally goes on in our mind, arising because of inattention, incomplete knowledge, sparse sensory data, mis-perceptions, forgetting something, problems in our relationships with colleagues, friends and family, and so on.  In turn many of these factors are shaped by the operational context in which we work, including the safety culture, the management culture, our working conditions and the fitness for purpose of the tools we work with.  

In order to reduce errors, the challenge for organisations is to create environments in which people can make errors without dire consequences. This means using the most cost-effective combination of the following approaches to error management: 

• Technical design – is used either to ensure that people cannot make certain sorts of errors (e.g., by installing signal interlocking), or else to help users to review their decisions before enacting them (e.g., a dialogue box that asks, ‘Are you sure you want to delete the selected file?’).  
• Procedure and process design – well designed procedures and processes can help to manage human error, for example limiting the memory burden on staff or by putting related tasks together to reduce errors. 
• Training – is used to ensure that people are rehearsed in their skills and knowledge and therefore less likely to make errors and are better able to recover from errors when they do make them. However, thought needs to be given to the type of training required. More training on skills, rules and knowledge is of little benefit to people who commit deliberate violations which are discussed below. Violations are better dealt with by showing people the consequences of their actions.
• Staffing – is used to ensure that the right people are placed in the right jobs. More importantly, it ensures that people are recruited who can be trained in the skills and responsibilities to the level that will be required of them.
• Culture – is developed by the organisation, through its leadership, management and teamwork so that people work in a supportive, blame-free atmosphere. As a result, everyone develops a responsible approach to managing the detection and correction of errors, reducing their consequences and preventing their re-occurrence. 
• Conditions – are considered with the aim of identifying and reducing the error-making consequences of motivation and morale, stress, workload and shift work. The reduction is achieved mostly through effective design, training and management (including self-management). 

Violations – Intentional Rule Breaking 

Sometimes people break a rule because they don’t know it exists, or they don’t understand it well enough, or they fail to recognise that a situation demands it. Or perhaps they simply forget that the rule exists. In all of these cases, rule breaking falls into one or another of the GEMS categories of human error discussed above. 

Sometimes, however, people break a rule deliberately. This means that the rule-breaking is not really an error, but a violation.  Overwhelmingly, people do not break rules maliciously, but for entirely rational reasons. In general, violations result from the conflict between an organisation that is attempting to control the behaviour of the workforce, and the individual who is attempting to carry out their task as easily as possible.  

To reduce rule breaking, we must make sure the rule is necessary. Before trying to persuade people to follow a rule you should first see if you can simplify the task or remove opportunities for error, and therefore the need for the rule. Here are some key pointers: 

• Make sure the rule is credible. If the main purpose of a safety rule is to protect an organisation rather than the safety of the individual, its credibility will suffer. It must be clear that the focus of the rule is on safe behaviour, not compliance. Credibility will also suffer if it is out of date or out of line with new procedures – it needs to be reviewed at appropriate intervals. Not doing so will increase the number of routine violations.  
• Make sure the rule is understood. People must be aware of the rules themselves, how they fit with related rules, the hazards that they are attempting to avert, and the consequences of not observing them. This means you need to pay serious attention to rule description, training and dissemination.  
• Make sure the rule is practicable. If the wrong method is easier, or if the right method is impractical, people will use the wrong method. You need to make it possible for staff to plan their work to take the rule into account and provide the equipment necessary to perform work according to instructions. You also need to ensure that targets are achievable without short cuts.  
• Make sure the rule is consistent with both organisational and team goals. When the goals of a work group conflict with the goals of the organisation this may give rise to informal ways of doing things that encourage infringements of the rules. The supervisor, being close to the work group, may share the norms of the group and therefore support such infringements. The aim of the organisation should be to foster informal norms that do not go against its goals.  

• Make sure the rule is rehearsed right after training. Failing to practice new rules soon after training – either operationally or via simulation – simply wastes the training resource. Either the rule itself will be forgotten, or its perceived importance will be reduced. 
• Make sure the rule is enforced. Rules must be supported by effective monitoring of the work practices and enforcement. You need to apply sanctions consistently and fairly when non-compliance occurs. Increasing the costs of violating will increase compliance.