Time perception test: method

What it measures

Interval timing in the seconds range: accuracy (mean absolute relative error) and bias (mean signed relative error; producing too short means subjective time runs fast).

Procedure

4 rounds with fixed targets in a fixed order: 10, 4, 7 and 15 s. Each round shows the target; the player presses Start (button, Space or Enter); no clock is visible; the player presses Stop when they think the target has passed. After each round the page shows where the player stopped, next to the target. The instructions ask players not to count. A press earlier than 300 ms is not accepted. Without a press within 3 times the target, the round ends as 'much too late'. No practice round. Both presses are timed by the browser on the same device. Leaving the browser tab ends the run.

Parameters

Scoring

Per round: error % = (produced − target) / target x 100, with the produced time capped at 3 x the target. Primary: mean absolute error % over the 4 rounds; lower is better. Also: mean signed error % (negative = stopped early, shown as 'your inner clock runs fast'; positive = 'runs slow'; within 3% = 'even'), the closest round, and each round's produced time. No player comparison is computed by the scoring service (see notes).

Valid-run rules

All 4 rounds must be sent with the targets in the right order; a run with a press under 300 ms is rejected. Rounds without a press are kept and scored at 3 x the target. The server rebuilds every run from the raw log (and the run's seed where there is one) and computes the score itself; anything the page says about right or wrong is ignored. A log that breaks the game's rules is rejected, and nothing is stored. Stored runs can carry quality flags that are never shown and never block a result: 10 or more runs from one IP address, a missing visitor id, tab switches, and automation signals (automated browser, impossible speed, no input events, machine-regular timing).

Differences from the classic task

Lab time production studies usually use many trials per interval, no feedback, and control counting with a concurrent task. Here there is one trial per target, feedback after each round (so later rounds can be corrected), and counting is only discouraged by the instructions.

Limitations

Counting cannot be checked or prevented. One trial per target makes the score noisy. Attention, mood, tiredness and distraction change timing. Not a medical or cognitive test; it says nothing about ADHD, dementia or any other condition.

Languages and review status

Language-independent stimuli, instructions in all site languages (English plus 99 other languages).