How does the Vigilance Test work?
For 3 minutes you watch an empty box. After a random wait of 1.5 to 5 seconds a counter starts in it, counting milliseconds; stop it as fast as you can by tapping anywhere or pressing Space. Your time stays up for 1 second, then the next wait begins. That is all that happens, and that is the test.
A press before the counter starts, or within 100 ms of it, is a false start: too soon to be a reaction. If you don't press within 10 seconds the counter stops by itself and counts as a lapse. The run ends at the first response after 3 minutes, never in the middle of a count, and a bar at the top shows how much is left.
How is the Vigilance Test scored?
Your result is your median reaction time, the middle of all your responses, with your lapses and false starts beside it. A lapse is a response of 355 ms or slower, or none at all. The slowest 10% shows how bad your worst moments were, and a chart splits the run into 30-second stretches, so you can see whether you slowed down as the minutes passed.
Each count is timed from the animation frame that shows it to the moment your finger or key goes down, using the timestamp your browser gives the press. Leave the page and the test pauses: the count in progress isn't scored, and the clock waits for you.
What is the psychomotor vigilance test (PVT)?
The PVT is a standard test of how alert you are, used in sleep and fatigue research: a simple reaction to a counter at random moments, where the slow responses matter more than the fast ones. It grew from the portable reaction-time task David Dinges and John Powell (1985) described for long, sleepless shifts; the standard version runs for ten minutes.
Ten minutes is long, so Mathias Basner, Daniel Mollicone and David Dinges (2011) tested a three-minute version, the PVT-B, against sleep loss in 74 healthy adults. It kept most of the full test's sensitivity once the lapse threshold was lowered from 500 ms to 355 ms. This test borrows the PVT-B's length and lapse threshold. In a second study, Basner and Dinges (2011) found lapses and response speed told rested and sleep-deprived people apart best, and mean and median times much less well: count your lapses as well as your median.
Can the Vigilance Test tell if you're tired?
It can show you slower, or with more lapses, than you usually are, which is what tiredness does. It can't say why, and it gives no diagnosis. Compare yourself with yourself: same device, same time of day. A single run against nobody else says little, and there is no published average for this version on phones and laptops.
What can affect your Vigilance Test result?
Sleep and time of day most of all; then your device. A touch screen, a wireless mouse or a slow display add a delay to every response, and that delay is part of your median. Lapses are less affected, because a lapse is hundreds of milliseconds slower than usual. For five quick rounds instead of three minutes, try the Reaction Time Test; to keep watch for one letter among many, the Continuous Performance Test.