The simple reaction time test is the oldest measurement in experimental psychology. A stimulus appears. You respond. The interval between the two is your simple reaction time, measured in milliseconds.
This page is a plain-English explainer. What the test actually measures, what typical adult numbers look like, why the number drifts with age, the difference between simple and choice reaction time, and why your online score is probably slower than your real reaction time.
What the simple reaction time test measures
Simple reaction time measures the total interval between stimulus and response when there is exactly one possible stimulus and exactly one possible response. No choice is required. The participant just responds as fast as possible whenever the signal arrives.
That interval breaks into roughly four pieces:
- Stimulus detection — how long the visual or auditory system takes to register the signal.
- Decision — minimal in the simple version; just "yes, there it is."
- Motor planning — preparing the muscle response.
- Motor execution — actually pressing the key (or saying the word, or releasing a button).
The "decision" component is tiny when there is no choice. That's what makes simple reaction time a relatively clean measure of basic processing-and-execution speed. Add a choice (red square → left key, blue square → right key) and you've measured choice reaction time, which loads decision-making on top.
Where the paradigm comes from
Reaction time as a measurement predates modern cognitive psychology. The Dutch physiologist F.C. Donders proposed the subtraction logic — measure simple RT, measure choice RT, the difference is the decision time — in the 1860s. James McKeen Cattell, the first American psychology PhD, made reaction-time tests one of the centerpieces of his "mental tests" battery in the 1890s, betting that reaction speed would correlate with intelligence. (It does, weakly. More on that below.)
The modern reference for adult norms is Woods et al.'s 2015 Frontiers in Human Neuroscience paper, which collected simple-reaction-time data from over 7,000 participants and published normative curves by age and several other variables. It is the largest published RT norm dataset and the one most cited in current research.
For the relationship between RT and other outcomes, Deary's research program on reaction time and cognitive ability — and the related longitudinal Lothian cohort findings on RT and aging — are the standard reference. Faster RT correlates modestly but persistently with measures of general cognitive ability and (interestingly) with later-life health outcomes. The effect is small but consistent across very large samples.
The classic Salthouse 2000 paper on aging and processing speed is the reference everyone reaches for on the age curve.
How the modern test works
In a typical setup:
- You watch a screen (or wait with eyes closed for an auditory cue).
- A signal appears at a random interval — typically 1 to 5 seconds.
- You respond as quickly as possible.
- You do this 20 to 50 times.
- Your score is reported as either median or mean reaction time across trials, often with outliers (above 1000 ms or below 100 ms) excluded.
The median is usually the more trustworthy single number because reaction-time distributions are right-skewed — a handful of slow trials pull the mean up disproportionately. Lab researchers usually report both.
A simple reaction time test takes 2 to 5 minutes including instructions.
What "good" looks like
For healthy adults in lab-grade testing using visual stimuli and key-press responses:
- Adults 18–30: median around 200–220 ms.
- Adults 30–50: median around 215–245 ms.
- Adults 50–70: median around 230–270 ms.
- Adults 70+: median around 250–320 ms, with substantially more individual variation.
Auditory simple reaction time is typically 20–40 ms faster than visual, because the auditory pathway has fewer synaptic stages. The fastest reliable simple reaction times observed in any condition are around 150 ms in young, highly-practiced adults using auditory stimuli — that's roughly the floor imposed by neural conduction and motor execution latencies.
These ranges are from Woods et al. and similar large-N studies. Online tests will run 30 to 80 ms slower than these numbers in most browsers, due to display refresh rate, input polling rate, and assorted JavaScript and OS latency. The relative ranking between your trials is fine. The absolute number is not directly comparable to lab norms.
How to interpret your own result
Three honest framings.
RT drifts night to night, day to day, even hour to hour. Sleep, caffeine, time of day, ambient distraction, and how much you've been moving in the last 10 minutes all move your RT by tens of milliseconds. A single test session is a snapshot in a moving distribution.
RT correlates with cognitive ability — but the correlation is small. Across the literature, simple RT correlates with IQ measures at roughly r = 0.2 to 0.3. That means it accounts for somewhere between 4% and 9% of the variance in IQ. You cannot read intelligence off reaction time at the individual level. The correlation is real in large samples, but useless for any individual conclusion.
Browser tests aren't lab tests. Take your online RT, subtract roughly 30–80 ms for system overhead, and you have an estimate of your "true" simple reaction time. Use the relative RT — your number compared to your number last week, not your number compared to a published norm — for any meaningful self-tracking.
The age curve, honestly
Simple reaction time gets slower with age. This is one of the most robust findings in cognitive aging. The effect is gradual through the thirties and forties, more visible from the fifties on, and substantial by the seventies.
This is real. It's also normal — not a sign of decline, not a marker of disease, and not something brain-training apps have shown to reverse. The slowing reflects ordinary changes in nerve conduction velocity, sensory thresholds, and motor execution. It is not a measure of intelligence falling. The Salthouse work shows processing speed slows even when other cognitive measures hold relatively steady.
What practice can do: keep you near the upper edge of your age band for your specific RT task. You can absolutely get faster at the specific test with practice. The effect transfers to closely related RT tasks. It does not generalize broadly to "being faster at life."
Related Senwitt content
- The Stroop test is RT-based but measures interference rather than raw speed.
- The Flanker test is another RT-based attention measure.
- For the broader argument about reaction speed, brain training, and what does and doesn't transfer, see does brain training work.
If you want a daily practice habit that exercises speed-sensitive thinking across reading, math, code, memory, and reasoning — not as a measurement but as practice — that's what Senwitt is for.
Simple vs. choice reaction time
A simple reaction time test has one stimulus and one response. As soon as anything appears, hit the key.
A choice reaction time test has two or more possible stimuli, each requiring a different response. Red square → left key, blue square → right key. Hick's law (1952) describes the relationship: choice RT scales roughly with the logarithm of the number of alternatives.
Choice RT is what cognitive psychologists usually mean when they talk about "reaction time as a cognitive measure," because it captures decision time on top of execution time. Simple RT is the more isolated processing-speed measure.
A complete RT battery often includes both, plus a "go/no-go" variant (respond to one stimulus, withhold response to another) which adds inhibitory control.
A note on online reaction time tests
There are many. They are mostly fine for what they are.
Two things to keep in mind:
- Absolute milliseconds are not directly comparable across tests. Each implementation has its own measurement latency. A 220 ms on one test and a 240 ms on another don't mean what they look like.
- Phone vs. desktop matters. Touchscreens have higher input latency than physical keys. Phone scores typically run 50 to 100 ms slower than desktop scores for the same person on the same day.
Use online RT tests for the experience of measuring reaction time, for tracking your own trend over weeks on one specific test (held constant), or for comparing yourself against family and friends on the same test. Don't compare across implementations.
