How attention is measured: reaction time, flanker and visual search

§ 1The classic attention tasks

Each task isolates a small part of what we call attention. None of them measures attention as a whole, and none is a diagnosis on its own.

§ 2Reaction time and alertness

The psychomotor vigilance task is one of the simplest and most sensitive tasks: a signal appears after a random wait and you respond as fast as you can. The standard lab version lasts 10 minutes; a 3-minute version, the PVT-B, was developed by Basner and colleagues (2011). Researchers look at response speed and at lapses, very slow responses that become more frequent when people are tired.

Short versions are useful for tracking yourself over time, but they are not interchangeable with the full lab test: Antler and colleagues (2022) found that the 3-minute and 10-minute versions agreed poorly.

§ 3Flanker and visual search: costs, not raw speed

In the flanker task, introduced by Eriksen and Eriksen (1974), you report the direction of a central arrow and ignore the arrows around it. When the neighbours point the other way, you are slower; that difference is the flanker effect. In visual search, finding one orange bar among blue ones is fast however many bars there are, while finding a target defined by colour and orientation together takes longer with every extra item (Treisman and Gelade, 1980; Wolfe, 1998).

These tasks report differences: the cost of conflict, the cost per item. On phones, that turns out to be an advantage.

§ 4Why phones differ from labs

Labs use equipment with known timing. A phone adds delay between your touch and the moment the app registers it, and between the app drawing a stimulus and the screen showing it. Pronk and colleagues (2020) measured this with a robotic finger and a light sensor: on Android, reaction times were overestimated by about 70 ms on average. Across devices the extra delay is roughly 50–100 ms, and it differs from one phone to the next.

• Absolute reaction times on a phone cannot be compared with lab norms or with results on other people’s phones.
• The delay is fairly constant on one device, so your own results on the same phone can be compared over time.
• Differences between conditions, such as the flanker effect, cancel most of the delay, and so do accuracy and memory span.

§ 5Why one result is not enough

Short tasks are noisy. Performance changes with sleep, time of day, practice and motivation. Difference scores are robust to device delay but vary more from one session to the next than raw speed does (Hedge, Powell and Sumner, 2018). Results become more meaningful when you repeat a test at a similar time of day and look at your trend and your usual range rather than at a single score.

§ 6A note on “brain training”

Practising an attention task makes you better at that task. A major review by Simons and colleagues (2016) found that brain-training programmes improve the trained tasks, sometimes closely related tasks, and show little evidence of improving everyday cognition. A rising trend on a phone test is a fact about the test, not proof that your attention in daily life has changed.

Questions and answers

Can a phone test diagnose attention problems?

No. Short phone tasks are not clinical tests and cannot diagnose ADHD or any other condition. If you are worried about your attention, talk to a qualified health professional.

What is a normal reaction time?

In lab studies, simple reaction times of rested young adults are often around 250 ms. On a phone, measured times are longer because of device delay, so compare with your own earlier results instead.

Why is my reaction time different on another phone?

Each phone adds its own delay, roughly 50–100 ms, from the touchscreen, the software and the display. Compare results only on the same device.

Does practising these tests improve attention?

It improves your performance on the tests themselves. Evidence that brain-training games improve attention in everyday life is weak.