Taking an online reaction time test is easy. Interpreting the number is where almost everyone goes wrong, because the figure your screen and mouse contribute is a large fraction of the range that separates fast people from slow ones.
A simple reaction test has one signal and one response: you wait, something changes, you click. There is no decision to make, which is why it produces the fastest times a person can record and why it is the baseline every other timed task is compared against.
The delay before the signal is randomised deliberately. Without that, anticipation would beat reaction, and the test would measure your rhythm rather than your reflexes.
A choice reaction test — several possible signals, each with its own response — is always slower. Hick described the relationship in the early 1950s: response time rises roughly with the logarithm of the number of alternatives.
Figures around 200 to 250 milliseconds for visual reaction are widely quoted. They are a reasonable rough guide and a poor benchmark, because measured values shift substantially with the equipment doing the measuring.
Anything much under 150 milliseconds on ordinary hardware almost certainly means you anticipated rather than reacted. Nerve conduction, processing and muscle activation together set a floor, and display latency sits on top of it.
The number worth tracking is your own average over five or more clean trials on one device. That figure is real and it moves.
A 60 Hz display can add up to 16 milliseconds of latency before the signal reaches your eye. Input lag from the mouse or touchscreen adds more, and browser timing adds more again.
Touchscreens are the worst offenders, which is why the same person reliably records slower times on a phone than on a desktop with a wired mouse.
This is why comparing a score from one site or device against another tells you almost nothing, and why every honest reaction test tells you to compare against yourself.
Sleep. Loss of sleep slows reaction time reliably and substantially, and it is one of the most sensitive markers of fatigue there is. A bad score is often a sleep signal.
Caffeine shortens reaction time in the short term — one of the more robust findings in the area.
Age slows it gradually and consistently in every large study that has looked, though alertness and practice both shift it back somewhat.
Practice does relatively little to the underlying speed. What improves is consistency and anticipation, which is exactly why an average of several trials is worth more than a personal best.
Hover your finger rather than resting it on the button — contact primes a press before you have decided to make one.
Look at the centre of the screen. Detecting a change in your peripheral vision costs a few milliseconds.
Run at least five clean trials and use the average. Discard false starts rather than counting them.
Test at a similar time of day when comparing. Reaction time follows a daily rhythm large enough to mislead you.
Around 200-250 milliseconds is the commonly quoted range for visual reaction. Because hardware moves the measured value by tens of milliseconds, your own average on one device is the only number worth tracking.
Touchscreen polling and display latency add real milliseconds. It is a hardware difference, not a difference in you.
Only modestly. Consistency and anticipation improve with practice; the underlying speed barely does. Sleep and alertness move it more than any amount of training.
No. Reaction speed and reasoning ability are only weakly related, and a fast reaction test result says nothing about how well you think.
Because you are anticipating rather than reacting. The delay is randomised precisely so that anticipation loses — wait for the change rather than for the moment you expect it.