Does autocorrect make you type faster?

The evidence is weaker than the headline numbers suggest. In the largest measured phone study, people who used autocorrection averaged 43.4 words per minute ( WPM), faster than every other group ¹. The authors tested the result again in a matched U.S. subsample, and there the speed difference was not statistically significant ². So autocorrection may help, but it is not proven to, and the data give no reason to switch it off. The wider picture is in our phone typing speed guide.

Autocorrection users were the fastest group in the largest study

Across the full sample of volunteers, the link between autocorrection and speed was moderate and positive, with a correlation of 0.237 ³. A correlation of this size means people who use the feature tend to type faster, but it does not mean the feature causes the gap. People who choose autocorrection may also differ in ways the study does not measure.

Squared, the correlation accounts for a small share of the differences between people. Autocorrection use explains about 5.6 percent of the variation in typing speed between participants ⁴. Most of the difference between typists comes from other factors, so no single phone setting decides your speed.

The matched sample did not confirm the gap

The authors checked the result in a U.S. subsample of 1,475 participants, chosen to match the U.S. population more closely on gender, age and phone operating system ⁵. In that subsample, the speed difference between autocorrection users and people who typed normally was not statistically significant ².

That is the caveat that matters most. A large sample can show a pattern that a matched comparison does not confirm. The honest reading is that autocorrection is linked with faster typing in the full sample, and the matched comparison is not strong enough to call the gain reliable.

Autocorrection can fix errors you would otherwise retype

In a field study of mobile typing, the 16 people who used autocorrection had a median of 15.3 automatic corrections per person per day ⁶. Each correction is a mistake you did not have to find and retype. That is where the feature can save time.

The same logic has a limit. A correction that you did not want costs you attention to notice and undo it. The study does not measure that cost directly, so the effect on speed is an open question for each person.

Word suggestions are a different tool, and they went the other way

Autocorrection and word suggestions are often grouped together, but the data treat them differently. Word prediction, which means tapping a suggested word, averaged 32.8 words per minute, a little below the 34.8 of people with no text-entry aid ⁷. In a controlled study of mobile suggestions, the authors found that the costs of attending to and using suggestions slowed average entry time ⁸.

Autocorrection changes a word for you. A suggestion asks you to stop, read and choose. That difference in attention is one plausible reason the two features do not give the same result, but the studies do not test it head to head.

What to do

1. Keep autocorrection on, unless you find that it changes your words often enough to cost you more than it saves.
2. Check the words it changes for a few days. Names, codes and technical terms are the usual problem.
3. Add your common names and terms to your phone's personal dictionary, if it offers one, so that the correction is right the first time.
4. Ignore word suggestions for ordinary words. Use them only when a word is long, unusual or hard to spell.
5. Measure with the same test before and after, counting errors in the final text as well as speed. The method is in how typing speed is measured.

For related habits, read how to type faster on a phone. The two thumbs question covers the posture data, and the swipe typing question shows what the studies say about the other input aid many phones offer.