Concentration and vigilance tests for safety-critical roles

In safety-critical work, the most dangerous moments often come after long stretches when nothing happens. Concentration and vigilance tests measure the ability to stay alert through those stretches. Here is the science behind them, how they are used in train driver selection, and how to make decisions from them responsibly.

What concentration and vigilance tests measure

Concentration and vigilance tests measure sustained attention: the capacity to keep focus on a task over time and to detect infrequent, important signals among many routine ones. That matters wherever a missed signal has serious consequences, including train driving, signalling, control rooms, process plant monitoring and many transport and healthcare roles.

Two task families dominate. Vigilance tasks ask the candidate to monitor a display for occasional critical events over an extended period. Cancellation tasks, in the tradition Bourdon pioneered, present rows of symbols and ask the candidate to mark every target and ignore close distractors as quickly and accurately as possible. Modern Bourdon-style tests separate speed (how much is worked through) from errors, which are themselves split into omissions (missed targets) and commissions (wrongly marked distractors).^[1]

The vigilance decrement

The scientific foundation goes back to wartime research. Mackworth studied radar operators' problem of spotting rare signals and, using what became known as the Mackworth Clock, showed that detection performance declines measurably within the first half hour on watch.^[2] This vigilance decrement has been replicated many times since. An influential later review argued that vigilance is not the undemanding task it appears: sustaining attention takes hard mental work, carries a real workload cost and is experienced as stressful.^[3]

Two practical lessons follow. First, the ability to sustain attention varies between people, which makes it worth measuring for roles where it matters. Second, sustained attention is also shaped by the job itself, through shift patterns, fatigue, task design and workload. Selection is one layer of safety management, not a replacement for good rostering and well-designed equipment.

Train driver selection in Great Britain

Train driving is the clearest case where cognitive assessment is a regulatory matter. Under the Train Driving Licences and Certificates Regulations 2010, the minimum medical examination for a train driver must cover cognitive functions including attention and concentration, memory, perception and reasoning, and psychomotor functions including reaction time and hand coordination; drivers must also pass an occupational psychological examination designed to identify deficiencies in operational aptitudes or relevant personality factors likely to interfere with safe working.^[4] The Office of Rail and Road's guidance notes that these cognitive and psychomotor tests are usually completed as part of the psychological examination, carried out by a psychologist recognised on the ORR register.^[5]

The industry standard that sets out how this is done is RSSB's Rail Industry Standard RIS-3751-TOM, Psychometric Assessment in Train Driver Selection. The current version, Issue 4, was published in December 2024; RSSB said it streamlined the process, set out how long test results remain valid and allowed candidates more attempts.^[6] ORR's guidance states that psychologists should use the tests the standard recommends or show that their assessments are equivalent.^[5] Concentration tests are only one strand: employers also assess judgement in realistic scenarios, for which a train driver SJT can provide evidence alongside the regulated psychometric assessments.

Scoring and interpreting concentration tests

Report speed and accuracy separately. Research on the d2, one of the most widely used Bourdon-style tests, found that the speed score is highly reliable, omission errors are reasonably reliable, but commission errors, on their own, are not reliable enough to base decisions on.^[1] Interpretation should rest on the measures that can bear the weight.

Use enough test length. The same research showed that reliability rises with test length,^[1] so very short tasks give noisier scores. For a measure of sustained attention, duration is part of what is being measured.

Interpret in bands, with other evidence. Report results in bands from Very Low to Very High against a relevant comparison group, and read them alongside the other assessments in the process and the requirements of the role and any applicable standard.

Consider adjustments carefully. Some candidates with disabilities may need adjustments, but in a vigilance task time and duration can be central to what is measured. Seek specialist advice; see Reasonable adjustments in online testing.

Supervised retesting for high-stakes decisions

Many organisations use unsupervised online tests early in a process for efficiency. For high-stakes decisions, the International Test Commission's guidelines on computer-based and internet-delivered testing support confirming unsupervised results with a follow-up test in a controlled setting,^[7] and research on confirmation testing shows that verifying results this way can offset the damage cheating would otherwise do to validity.^[8] For safety-critical roles, the principle is simple: a result that will contribute to a decision about whether someone can work safely should be confirmed under supervised conditions, with identity checked.

That matters more now that candidates have easy access to AI tools; see Can candidates cheat psychometric tests with AI? for what that means for test design.