Reaction time test on mobile and on a computer, why the numbers differ

September 26, 2026 · 6 min read · Tyamka team

A reaction time test on mobile adds its own delay, and so does a laptop. What a lab found on ten device and browser pairs, and why not to compare them.

A reaction time test on mobile and the same test on a computer can give different numbers even if you react exactly as fast. A test in a browser records your reaction plus the delays of the device, such as the screen, the touch sensor or keyboard, the browser and the operating system, and those delays differ from device to device. A lab study that measured them found that on some combinations of device and browser the web page added about 58 ms to the real time, and on others about 133 ms. The phones were not the slowest in that study. The time you get in our reaction time test contains the same kind of delay.

Where the extra time comes from

Between the moment a test decides to show a signal and the moment it records your press, several things happen that have nothing to do with you.

• The screen. A screen updates in frames. At 60 frames per second a frame lasts 16.7 ms (1000 divided by 60), so a new image can wait up to that long to appear. The panel itself also takes some time to change its pixels.
• The browser. It prepares the page, draws the new frame and hands it to the system.
• The touch sensor, keyboard or mouse. Your press has to be detected, passed to the system and then to the page.
• The operating system. It sits between the hardware and the browser and schedules all of the above.

Each step is small, but they add up, and the page sees only part of them.

What a lab measurement found

In 2020 Thomas Pronk and colleagues at the University of Amsterdam published a careful test of this in Behavior Research Methods. A light sensor on the screen caught the exact moment a stimulus appeared, and a solenoid, a small electromagnetic plunger, pressed the screen or the Q key after a precisely known interval. Then they compared the real interval with the one the web page recorded. The devices were a Samsung Galaxy S7 with Android, an iPhone 6S, an ASUS laptop with Windows and a 2016 MacBook Pro.

Every combination recorded the interval as longer than it really was. The gap between the lowest and the highest combination was about 75 ms, by our subtraction. The phones were not consistently worse than the laptops. The iPhone added the least, the Android phone slightly more than the Windows laptop, and the MacBook the most.

On a given device the extra time was fairly steady from press to press. On seven of the ten combinations its standard deviation was 8 ms or less, two Mac browsers were at 8.1 ms, and Firefox on the Mac was at 16.1 ms. So one device is mostly consistent with itself, it just adds a different amount from the next device.

These were four devices from 2015 and 2016, with browsers from late 2018 such as Chrome 71. A newer phone or laptop may add more or less. The study does not tell you what your own device adds. It shows that the amount varies a lot, and that you cannot guess it from whether the device is a phone.

So is reaction slower on a phone?

Not necessarily. The idea that phones are always slower is common, but the lab measurement above does not support it. What it supports is that each device adds its own delay, and that delay depends on the model, the operating system and the browser.

How you press may also differ. On a computer your finger often rests on the mouse button or the key, while on a phone it hovers over the glass. That could change the time a little in either direction. We found no study that measured it, so treat it as a guess.

A faster screen shortens one of the steps. At 120 frames per second a frame lasts 8.3 ms instead of 16.7, so a new image waits less for the next frame. That is arithmetic about the frame, not a measurement of the total delay on any phone.

An older paper, Ng and colleagues at the UIST conference in 2012, wrote that touch systems of the time typically took 50 to 200 ms to update the screen after a touch. That is a different measurement, the whole path from finger to picture, and devices have changed a lot since, so we would not use it as a number for today's phones.

Why not to compare across devices

Suppose a friend gets 230 ms on a laptop and you get 270 ms on a phone. That tells you very little, because part of the gap, perhaps all of it, may be the devices. The authors of the study reached a similar conclusion for researchers. Comparing one person under two conditions on the same device is much safer than comparing absolute times between devices.

For your own results this means:

1. Compare only results from the same device and the same browser.
2. Do several attempts and look at the middle one.
3. When you change phone or laptop, start the comparison again.

How our test measures time

In our reaction time test the clock starts at the first animation frame after the circle is added to the page, the last moment the page can see before the pixels change. That leaves out the time the page spends updating before the frame. We measured that part in headless Chrome on a computer. From the moment React committed the circle to the frame that starts the clock, the median was 7.6 ms, and none of it is in your time.

What the page cannot leave out is everything after that frame, the screen's own delay, and on the other end the time the touchscreen, mouse or keyboard takes to report your press. So every time the game shows is longer than your real reaction. We have not measured by how much, and nothing at all has been measured on real phones yet. Pronk and colleagues timed their stimuli in a similar way, with animation frames, and found the 57.6 to 132.9 ms in the table above.

The article on the reaction time test average explains what research and a large online test report and why a lab can give a shorter time than a browser. The article on reaction time and driving covers why none of these numbers says much about braking in a car. If a time looks too fast to be real, the article on the reaction time test world record explains why sprint rules and fair tests treat anything under about 100 ms as doubtful.