Reaction time test: method

What it measures

Simple visual reaction time: the time from the onset of one expected visual signal to one motor response. Responses slower than 500 ms are counted as lapses, an index of sustained attention.

Procedure

A grey circle is shown. After a random foreperiod it turns green; the page starts the clock when it asks the browser to show the green circle. The player responds with a pointer-down (touch or mouse) or the Space or Enter key. The reaction time (RT), rounded to 1 ms, is shown for 750 ms, then the next foreperiod starts. One practice round (skippable) is followed by 10 counted rounds. A response during the foreperiod is a false start: the page shows 'too early', the round is used up and not repeated, and no RT is recorded for it, so a run can have fewer than 10 RTs. There is no response deadline and no running counter on the stimulus. Responses while the feedback is shown are ignored. Input device class (touch, mouse, keyboard) is recorded.

Parameters

Scoring

Primary metric: mean RT in ms over valid counted responses (RT >= 100 ms), lapses included; lower is faster. Also reported: fastest and slowest valid RT, number of lapses (valid RT > 500 ms) and number of false starts (taps before the signal plus RTs under 100 ms). Each tap is labelled early, slow or ok. No percentile, norm or 0-100 score is computed on the web.

Valid-run rules

Taps before the signal and RTs under 100 ms are excluded from the mean. If no valid RT remains, the mean is empty and no score is shown. Leaving the tab or switching windows during the run stops it, and nothing is scored. Every scored run is stored with its input device class (touch, mouse, keyboard), page language, scorer version and runner-config hash. Quality flags (for example many runs from one connection within an hour, a missing visitor id, an automated browser, impossibly fast answers) and earlier attempts by the same visitor, account or connection are recorded with the row. Such runs are kept, but they are left out of any comparison with other players. Flags never change the player's own result.

Differences from the classic task

The standard PVT lasts 10 minutes (3 minutes for the brief PVT-B), uses foreperiods of 2 to 10 s and a running millisecond counter as the signal, and is run on dedicated hardware. This version has 10 trials, 1.5 to 4.5 s foreperiods and a colour change, so it is a short simple-RT measure, not a full vigilance test. Timing uses the browser clock and includes screen and input-device delay.

Limitations

Browsers and devices add delay and jitter: 58 to 133 ms in one lab check of web apps on phones and laptops, and touchscreens are slower than buttons. This device delay is larger than typical age differences, so scores are comparable mainly within one person on one device and browser. Ten trials give a noisy mean. The setting is unsupervised. The result is not a measure of intelligence, driving fitness or any health condition, and it has no clinical use.

Languages and review status

Language-independent stimuli; instructions and feedback in all site languages.