Language:English
Menu
Language: English

How is reaction time measured?

How is reaction time measured? A signal appears at a known moment, the person answers as fast as they can, and the time in between is the reaction time. Everything else is about doing that precisely: knowing exactly when the signal really appeared, catching the very start of the answer, and repeating it enough times to get a fair middle value. A calibrated laboratory set-up measured an average of 230.8 ms for adults pressing a mouse button, 213 ms once its own equipment delay was taken out[1].

Take the reaction time test

The parts of one measurement

  1. A wait of unpredictable length, so that the person cannot learn the rhythm and press early.
  2. The signal: a light, a picture, a sound or a vibration, whose real moment of appearance has to be known.
  3. The answer: a button press, a key, a touch, or a push against a sensor, whose real moment also has to be known.
  4. Repetition: many tries, with early presses and very late ones set aside, and a typical value taken from the rest.

Each step can go wrong in its own way. A fixed wait invites guessing. A screen that shows the signal late, or a button that reports the press late, adds time that is not the person's. And a single try says little, because nobody reacts at the same speed twice.

Which reaction time?

Researchers measure several kinds, and the kind decides the number. A simple reaction has one signal and one answer, and is the fastest. A choice reaction gives several signals with an answer for each, such as the 4-choice task in a large British survey[2]. A go/no-go reaction asks you to press for one signal and hold still for another; in a long study in Baltimore it slowed with age about three times as fast as a simple one[3]. A colour word test such as the Stroop test is a choice reaction with a conflict built in.

So a figure for reaction time always needs its task beside it. Our reaction time test measures a simple reaction to a light, the sound reaction test a simple reaction to a beep, and the Stroop test a choice with a conflict.

In the laboratory

The first reaction time measurements were made in the nineteenth century by Francis Galton, who recorded simple visual reactions below 190 ms in young people[1]. Modern laboratories use computers, but they check them: they measure the delay of their own screen and button and take it into account.

The large study that our test compares you with did exactly that. Its 1,469 volunteers saw a target and pressed a mouse button. The raw average was 230.8 ms; with the measured delay of the equipment taken out, it was 213 ms. The authors also measured how long each person took simply to start a finger movement, and so could estimate the time needed to notice the target: about 131 ms[1].

Laboratory software is also very precise. A comparison of experiment software, run over more than 110,000 trials with a measuring box, found the best laboratory programs timing stimuli and responses with a precision under 1 ms[4].

Some laboratories go further and record the brain or the muscles. In one study, brain recordings showed people deciding whether a flashed photo held an animal within 150 ms, before any button was pressed[5].

At a sprint start

Athletics measures reaction time for a practical reason: to catch false starts. Sensors in the starting blocks record the push of each runner's feet against the time of the gun. The rules of World Athletics treat a reaction of less than 0.1 second, as measured by such a certified system, as a possible false start[6].

Researchers have used the same kind of equipment in more detail, with force sensors in each footplate synchronised with the starting signal and recordings of the leg muscles. They found that the muscles of trained sprinters could respond to the sound in under 85 ms, and the first muscle signals in under 60 ms[7]. A button press is a much slower thing to time than a muscle signal.

In a browser or on a phone

A web page cannot calibrate the screen and the mouse in front of you. It can only time what it sees: when it asked for the signal to be drawn and when it was told of your press. Two studies measured how far that is from reality by pressing keys and touch screens with a robot. On ordinary laptops and phones, browsers recorded every reaction as slower than it was, by an average of 57.6 ms to 132.9 ms depending on the device and browser[8]; on desktop and laptop computers, the second study found 76.5 ms to 82.3 ms[9].

The good news is that the delay of one device is fairly steady. Most of the software in the large comparison timed responses in a browser with a precision under 10 ms[4], so differences between your own runs on one device are meaningful even when the absolute number reads high.

Phones add the most. In a study in which young adults answered beeps, vibrations and lights on a phone, the fastest single signal, a loud beep, still took 405 ms on average, well above laboratory figures for a light[10].

How our reaction test measures

  • The screen turns green after a random wait between 1.5 sec and 4.5 sec, so you cannot learn the moment.
  • The moment of green is taken from the screen's next frame, and your press from the time the browser gives the event itself, not from when our code got round to it.
  • A press faster than 100 ms counts as a guess and the try is repeated; one slower than 1,000 ms counts as a lapse.
  • After a practice try, 5 tries count (20 in precise mode), and your result is the middle one.
  • If you switch to another tab or window during a try, that try is dropped.

The test also measures your screen's refresh rate and shows it with your result, so you can see what the screen adds. It allows for the delay the robot studies found beyond the laboratory's own, 44.1 ms to 115.1 ms with a keyboard or mouse and 39.8 ms to 52 ms with a touch screen, and shows the comparison with the study as a range. On a computer that range is wide, which is one more reason to compare your own results on one device rather than chase one number.

The article on input lag looks at screens, mice and refresh rates in more detail, and every study our tests compare you with is listed on the page of studies behind the numbers.

Questions

How is reaction time measured in a laboratory?

With a calibrated screen and button, a random wait before each signal and many tries. The delay of the equipment is measured and taken into account: in one large study the average went from 230.8 ms to 213 ms once it was removed.

Are online reaction time tests accurate?

They are consistent, but they read high. Robot tests found browsers recording reactions too slow by an average of 57.6 ms to 132.9 ms on everyday devices, while timing them with a precision under 10 ms in most software.

How is reaction time measured in sprinting?

By sensors in the starting blocks linked to the starting gun. A reaction of less than 0.1 second is treated as a possible false start.

Why does my result change from try to try?

Because people never react at exactly the same speed twice, and because the moment of each press lands differently against the screen's refresh. That is why tests use several tries and take the middle one.

Sources

  1. Woods, D. L., Wyma, J. M., Yund, E. W., Herron, T. J., & Reed, B. (2015). Factors influencing the latency of simple reaction time. Frontiers in Human Neuroscience, 9, 131.doi:10.3389/fnhum.2015.00131
    n = 1,469; Community volunteers in Rotorua, New Zealand, aged 18 to 65 (mean age 45.8, 40% men)
    Abstract: "Experiment 1 examined a community sample of 1469 subjects ranging in age from 18 to 65. Mean SRT latencies were short (231, 213 ms when corrected for hardware delays) and increased significantly with age (0.55 ms/year)". Table 2, column Experiment 1: "N 1469", "SRT 230.8", "SRT SD 26.8". (read on 2026-10-01: https://www.frontiersin.org/journals/human-neuroscience/articles/10.3389/fnhum.2015.00131/full)
  2. Der, G., & Deary, I. J. (2006). Age and sex differences in reaction time in adulthood: Results from the United Kingdom Health and Lifestyle Survey. Psychology and Aging, 21(1), 62–73.doi:10.1037/0882-7974.21.1.62
    n = 7,130; Adults in the United Kingdom Health and Lifestyle Survey
    Abstract: "The authors reanalyzed data for 7,130 adult participants in the United Kingdom Health and Lifestyle Survey" "The authors modeled the age differences in simple and 4-choice reaction time means and variabilities" "Simple RT shows little slowing until around 50, whereas choice RT slows throughout the adult age range." (read on 2026-10-06: https://pubmed.ncbi.nlm.nih.gov/16594792/)
  3. Fozard, J. L., Vercruyssen, M., Reynolds, S. L., Hancock, P. A., & Quilter, R. E. (1994). Age differences and changes in reaction time: The Baltimore Longitudinal Study of Aging. Journal of Gerontology, 49(4), P179–P189.doi:10.1093/geronj/49.4.p179
    n = 1,265; Community volunteers in Baltimore aged 17 to 96
    Abstract: "This study analyzed auditory reaction time (RT) data from 1,265 community-dwelling volunteers (833 males and 432 females) who ranged in age from 17 to 96." "Repeated testing within participants (longitudinal analyses) over eight years showed consistent slowing and increased variability with age." "Beginning at about age 20, RTs increased at a rate of approximately 0.5 msec/yr for SRT and 1.6 msec/yr for DRT." (read on 2026-10-06: https://pubmed.ncbi.nlm.nih.gov/8014399/)
  4. Bridges, D., Pitiot, A., MacAskill, M. R., & Peirce, J. W. (2020). The timing mega-study: comparing a range of experiment generators, both lab-based and online. PeerJ, 8, e9414.doi:10.7717/peerj.9414
    Abstract: "Among the lab-based experiments, Psychtoolbox, PsychoPy, Presentation and E-Prime provided the best timing, all with mean precision under 1 millisecond across the visual, audio and response measures." "Online studies did not deliver the same level of precision as lab-based systems, with slightly more variability in all measurements." "For response times (measured using a high-performance button box), most of the packages achieved precision at least under 10 ms in all browsers, with PsychoPy achieving a precision under 3.5 ms in all." "The results, from over 110,000 trials, highlight the wide range of timing qualities that can occur even in these dedicated software packages for the task." (read on 2026-10-06: https://pubmed.ncbi.nlm.nih.gov/33005482/)
  5. Thorpe, S., Fize, D., & Marlot, C. (1996). Speed of processing in the human visual system. Nature, 381(6582), 520–522.doi:10.1038/381520a0
    Abstract: "Here we use a go/no-go categorization task in which subjects have to decide whether a previously unseen photograph, flashed on for just 20 ms, contains an animal. ERP analysis revealed a frontal negativity specific to no-go trials that develops roughly 150 ms after stimulus onset. We conclude that the visual processing needed to perform this highly demanding task can be achieved in under 150 ms." (read on 2026-10-06: https://pubmed.ncbi.nlm.nih.gov/8632824/)
  6. World Athletics (2026). Book C, C2.1 Technical Rules, Rule 16 (The start), version in force from 1 July 2026. World Athletics Book of Rules.https://worldathletics.org/about-iaaf/documents/book-of-rules
    C2.1 Technical Rules (document "C1.1 & C2.1 - Competition Rules & Technical Rules", dated 01 JUL 2026): "When a World Athletics certified Start Information System is in use, the Starter and/or an assigned Recaller shall wear headphones in order to clearly hear the acoustic signal emitted when the System indicates a possible false start (i.e. when the reaction time is less than 0.100 second)." (read on 2026-10-06: https://worldathletics.org/about-iaaf/documents/book-of-rules)
  7. Pain, M. T. G., & Hibbs, A. (2007). Sprint starts and the minimum auditory reaction time. Journal of Sports Sciences, 25(1), 79–86.doi:10.1080/02640410600718004
    Nine sprinters starting from instrumented blocks
    Abstract: "The simple auditory reaction time is one of the fastest reaction times and is thought to be rarely less than 100 ms." "Reaction time in nine athletes performing sprint starts in four conditions was measured using starting blocks instrumented with piezoelectric force transducers" "Five of the athletes had mean reaction times of less than 100 ms in at least one condition and 20% of all starts in the first two conditions had a reaction time of less than 100 ms. The results demonstrate that the neuromuscular-physiological component of simple auditory reaction times can be under 85 ms and that EMG latencies can be under 60 ms." (read on 2026-10-06: https://pubmed.ncbi.nlm.nih.gov/17127583/)
  8. Pronk, T., Wiers, R. W., Molenkamp, B., & Murre, J. (2020). Mental chronometry in the pocket? Timing accuracy of web applications on touchscreen and keyboard devices. Behavior Research Methods, 52(3), 1371–1382.doi:10.3758/s13428-019-01321-2
    Four devices from 2015 and 2016: a MacBook Pro, an ASUS laptop, a Samsung Galaxy S7 and an iPhone 6S, pressed by a robot
    Abstract: "In controlled circumstances, as can be realized in a lab setting, very accurate stimulus timing and moderately accurate RT measurements could be achieved on both touchscreen and keyboard devices, though RTs were consistently overestimated." Table 4 "Descriptives of RT overestimations (in milliseconds) per device and browser" (OS, Web Browser, Minimum, Maximum, Mean, SD): "Android Chrome 46.0 103.5 69.8 7.4", "iOS Safari 48.3 96.3 57.6 6.5", "MacOS Safari 93.0 163.7 132.9 8.1", "Windows Chrome 64.7 70.6 68.5 1.7", "Windows Firefox 49.8 84.9 61.9 5.7". (read on 2026-10-06: https://www.ebi.ac.uk/europepmc/webservices/rest/PMC7280355/fullTextXML)
  9. Anwyl-Irvine, A., Dalmaijer, E. S., Hodges, N., & Evershed, J. K. (2021). Realistic precision and accuracy of online experiment platforms, web browsers, and devices. Behavior Research Methods, 53(4), 1407–1425.doi:10.3758/s13428-020-01501-5
    Desktop and laptop computers with Windows 10 and macOS, pressed by a robot; and the equipment of 202,600 online participants
    Abstract: "We then employed a robot actuator in realistic set-ups to measure response recording across the aforementioned platforms, and between different keyboard types (desktop and integrated laptop)." "We found that modern web platforms provide reasonable accuracy and precision for display duration and manual response time". Table 2 "RT delay is calculated as the difference between known and recorded RT." Browser means: "Chrome 78.81", "Edge 80.10", "Firefox 82.30", "Safari 76.50"; device means: "macOS-Desktop 85.35", "Windows-Desktop 76.24", "Windows-Laptop 73.65". Results: "We found that 77% of these devices were desktop or laptop computers, whereas only 20% were mobile devices" "Based on a sample of 202,600 participants." (read on 2026-10-06: https://www.ebi.ac.uk/europepmc/webservices/rest/PMC8367876/fullTextXML)
  10. Yoshida, K. T., Kiernan, J. X., Okamura, A. M., & Nunez, C. M. (2023). Exploring human response times to combinations of audio, haptic, and visual stimuli from a mobile device. IEEE World Haptics Conference 2023 (arXiv:2305.17180).https://arxiv.org/abs/2305.17180
    n = 20; Adults aged 20 to 29 holding an iPhone 11
    "Our user study included 20 participants (13 female, 6 male, 1 non-binary; aged 20-29)." "Participants had the shortest response time for the high audio stimulus (405 ± 50 ms) and longest response time for the low visual stimulus (528 ± 105 ms)." "The mean response time is the fastest in the condition with high levels of all three stimuli (mean ± standard deviation, 320 ± 43 ms)" "The fastest mean response time recorded for a particular subject was 250 ms, and the slowest was 819 ms." "The overall mean response time across all conditions was (380 ± 52 ms)." (read on 2026-10-06: https://arxiv.org/pdf/2305.17180)