Understanding Safety / General

Defining and Understanding Error

Ask 100 people what their view on error is and you will receive 100 different answers. Error is a very subjective concept. What constitutes an error for one is not the same for some other. What everyone will agree though, is that error is an action (or inaction), in a way different than what it should be. It is a deviation from the correct. Errors are therefore a form of unsafe acts.

Errors can be classified in many ways. James Reason, as a leading researcher in human error, states that every error can be categorised under these four basic elements: the intention, the action, the outcome and the context.

Referring to the intention we examine whether the error was intentional or unintentionally. Then if it was intentional, one needs to examine the outcome. Was it planned or was the initial plan something different which failed to materialise. If we conclude that the error was intentional and the desired outcome was indeed achieved, which resulted in deviation from SOPs, then we talk about violation.

As per literature, action-based errors could be classified as:

  • Omissions: planned action not done at the correct intended time
  • Intrusion: unintended actions which are part of a different activity
  • Repetitions: actions already performed being repeated unnecessarily
  • Wrong objects: right actions performed but for wrong objects
  • Misordering: right action performed in wrong sequence
  • Mistiming: right action performed in wrong timing
  • Blends: unintended merging of two actions meant for different goals

Contextual based classification includes the following:

  • Anticipation and preservation: a similar event triggers the action of a later act in the sequence
  • Priming: repetition of prior sounds or repetition of the same answer in different questions, results in repeating the same response for inappropriate questions/ challenge
  • Interruptions and Distractions: unscheduled interruptions of a task, losing the correct sequence of acts.
Outcome based categorisation, examines the severity of the outcome. It is vital to understand though that severity is not linked to incompetence, as it is likewise wrong to judge an error based on the morality of the outcome. The outcome categorisation has four groups:
  • Free lessons: a clear opportunity for learning where an error did not lead in any catastrophic situation, any losses or any potentially unsafe circumstances
  • Exceedances: a captured event where the limits have been broken but without a negative impact or major consequences
  • Incidents: cases where the error lead to significant damage or injuries with minor financial/ technical loss
  • Accidents: catastrophic events with adverse consequences and substantial loss/ damage to equipment and/or life, environment etc

The table below summarises the above classification

Type Volume Cost
Exceedances Very High Very Low
Incidents Moderate to high Low to moderate
Accidents Low to very Low Unacceptably high

To objectively judge and error/ unsafe act, it is best to focus on the intention. Unsafe acts based on intentions can be group as below:

Unsafe Acts Diagram

The category of unintentional errors formed by the action slips and memory lapses, can be subdivided to three groups based on the underlying factors. Recognition failures, memory failures and attention failures.

Recognition failures is a result of misidentification of objects, situations, messages and the related expectations of these. Similarity and familiarity, as well as strong habits, form the majority of contributing factors. Another similar factor is the non-detection of a signal or a problem. Inexperience, pre-occupation, interruption, speediness and lack of training can result in non-detection errors. We also have the group of wrong detections, which involve misinterpretation and false detection of a non-existing issue.

Memory failures can result in either stage of information processing, that is input, storage and retrieval. Input stage is related to the short-term memory, storage is related to the long-term memory, whilst retrieval is related to either of the two memories when trying to recall something.

Attention failures are a result of the limited resource of attention. Strong habits intrusions form the majority of slips. This is nothing more than executing a perfect sequence of actions for the wrong activity. Familiarity and constant (unchanging surroundings/ environment) are the contributing factors. Likewise, interference errors result from blending of two currently active tasks or two elements of the same task.

Intended unsafe actions are split into mistakes and violations.
Mistakes are categorised by their defining source, the underlying reason for performing an action. These can be skill-based, rule-based or knowledge-based.

Skill-based mistakes is simply the lack of required skill to perform a task, either because of poor training, lack of skills or insufficient practise.

Rule-based mistakes are a result of the human information process models, which tend to find and match patters of actions, events, situations etc. In many unplanned situations we strongly strive to find a familiar pattern (from previous experience, training etc) to help us solve the problem. Hence, we can misapply an otherwise good rule in a wrong situation, apply a wrong rule or fail to apply a good rule.

Knowledge-based mistakes happen purely when faced with novel situations, where our training, past experience and all problem-solving skills we have learnt, are not suitable. This leads to a trial and error situation, where hopefully we get to the learning part before the catastrophe.

Rule-based and knowledge-based mistakes share some common biases. These are:

  • similarity (confirmation bias)
  • frequency bias
  • bounded rationality
  • reluctant rationality
  • irrationality

Examples to help identify the differences:

  • The pilot selects heading 100o instead of 110o (Slip)
  • The pilot performs 10,000ft checks late (Lapse)
  • An engineer uses the wrong wiring connection (Rule-based)
  • A junior ATC fails to vector an aircraft away from a thunderstorm (Knowledge-based)

Violation come in a variety of forms, but similarly to mistakes they are performed at the skill-base, rule-base or knowledge-base. Violations are a result of illusions; of control, of invulnerability, or of superiority. Further to this, the underlying factor in usually a lack of trust in the expedience and efficiency of current procedures. Therefore, the easiest way to trap violations is to increase the awareness of the benefits of compliance, rather than increasing the penalty cost to violators.

Another way to classify violations is by type:

  • Routine - the most dangerous in terms of latent conditions for accident causation, as it becomes the normal way of doing things in a workplace
  • Situational - depends on the conditions existing in terms of time pressure, workload, inadequate procedures/ tools, working condition
  • Optimising - aim to optimise efficiency and gain time, by acting inappropriately fast or bypassing/ short-cutting procedures
  • Exceptional – one off, a special circumstance which required violating of a procedure for the sake of safety or an unforeseen event requiring immediate action.

Detection and Trapping

Detecting errors differ upon the level of standards by which you assess the performance, as error is basically an action which deviates from the plan (acceptable performance).
Slips are the easiest to detect as we usually know our intentions and we can immediately spot and trap the deviation. Memory lapses can be sometimes obvious but sometimes it can take long before realising we have forgotten something.
Mistakes are more difficult to spot, as we do not have a planned outcome to compare with, and we do not know the ideal pathway.

Overall one should remember that errors are not intrinsically bad, as they are necessary for trial and error in novel situations (novel thinking out of the box). Bad errors are not a quality of bad people or poorly trained errors. It is frequent to see the worst mistakes being done by the best people in their attempt to push the limits of the performance. Finally, we have to appreciate that errors are not totally random. Even though the though the theory of Practical Drift states the opposite, the majority of errors tend to share a common base and they recur the same basic form.