Peer-reviewed research

How do People Type on Mobile Devices? Observations from a Study with 37,000 Volunteers

Kseniia Palin et al., Proceedings of the 21st International Conference on Human-Computer Interaction with Mobile Devices and Services (MobileHCI 2019), 2019

Read the original

Read on October 16, 2026

What it is: A web-based transcription study in which 37,370 volunteers typed sentences on their own phones through a typing test site. It reports typing speed by age, gender, language use and finger use, and includes a U.S. subsample matched to population gender and age groups.
How much weight it carries: Peer-reviewed, with a very large sample. Participants were self-selected web volunteers who chose a typing test, so they skew toward people who practice typing. The U.S. subsample was drawn to match population gender, age and phone operating system, which makes it the better comparison for the general public.

Facts taken from this source

The authors compare their phone data with desktop data gathered by the same method, and find average mobile typing about 15 words per minute slower.
“Average performance is about 15 WPM slower in mobile typing.” (Comparison with desktop keyboards (text))
Participants aged 10 to 19 typed the fastest of all age groups on phones, with a mean of 39.6 words per minute.
“Participants of age between 10 and 19 typed the fastest (M = 39.6, SD = 14.3)” (Results: Age)
Participants aged 20 to 29 averaged 36.5 words per minute on phones.
“10-19 39.6 (14.3) p<.001 38.0 (14.8) p<.001 20-29 36.5 (12.6) 39.0 (13.3) 30-39 32.2 (10.8) 34.3 (13.8) 40-49 28.9 (9.2) 27.3 (9.0) 50-59 26.3 (9.9) 24.8 (9.1)” (Table 4: Performance by demographic group)
Participants aged 30 to 39 averaged 32.2 words per minute on phones.
“10-19 39.6 (14.3) p<.001 38.0 (14.8) p<.001 20-29 36.5 (12.6) 39.0 (13.3) 30-39 32.2 (10.8) 34.3 (13.8) 40-49 28.9 (9.2) 27.3 (9.0) 50-59 26.3 (9.9) 24.8 (9.1)” (Table 4: Performance by demographic group)
Participants aged 40 to 49 averaged 28.9 words per minute on phones.
“10-19 39.6 (14.3) p<.001 38.0 (14.8) p<.001 20-29 36.5 (12.6) 39.0 (13.3) 30-39 32.2 (10.8) 34.3 (13.8) 40-49 28.9 (9.2) 27.3 (9.0) 50-59 26.3 (9.9) 24.8 (9.1)” (Table 4: Performance by demographic group)
Participants aged 50 to 59 averaged 26.3 words per minute on phones, the lowest of the age groups in Table 4.
“10-19 39.6 (14.3) p<.001 38.0 (14.8) p<.001 20-29 36.5 (12.6) 39.0 (13.3) 30-39 32.2 (10.8) 34.3 (13.8) 40-49 28.9 (9.2) 27.3 (9.0) 50-59 26.3 (9.9) 24.8 (9.1)” (Table 4: Performance by demographic group)
Participants under 10 years old were the slowest age group on phones, with a mean of 24.3 words per minute.
“participants of age below 10 – the slowest ( M = 24.3, SD = 13.2)” (Results: Age)
In the Palin study, participants who used autocorrection averaged 43.4 words per minute, faster than all other participants.
“Participants that used autocorrection were faster (M = 43.4, SD = 14.4) than all other participants” (Results, intelligent text entry)
Across all participants, autocorrection use had a moderate positive correlation with typing speed (r = 0.237).
“autocorrection has a moderate positive correlation with WPM (r = 0.237)” (Results, correlation analysis)
Men and women typed at about the same speed on phones, about 36 words per minute each, in the full sample.
“They both typed at about 36 WPM with only the SD of WPM being smaller for female typists.” (Results: Gender)
The phone typing study's data were collected between September 2018 and January 2019.
“The data were col- lected between September 2018 and January 2019. Table 1 summarizes the demographic background” (Section 3, Data collection)
The phone study's age analysis excluded 0.2 percent of participants, who were older than 60.
“Note that this analysis excludes 0.2% of participants older than 60.” (Section 4, Age)
In the full phone sample, 65 percent of participants were female and 31 percent male, and the U.S. subsample was 51 percent female.
“Gender 65/31 % f/m 4% n/a 51/49 % f/m” (Table 1: Summary of demographics, Gender row)
In the full phone sample, women averaged 36.0 words per minute and men 36.1, a gap of 0.1 words per minute with a Cohen's d of 0.003.
“female 36.0 (12.5) p=.002 34.6 (12.3) p<.001 male 36.1 (14.4) d=.003” (Table 4: Performance by demographic group, gender rows)
In the Palin study, people who reported gesture (swipe) typing averaged 32.2 words per minute overall, against 34.8 for people who used no intelligent text entry feature.
“none 34.8 (12.6) 2.3 1.2 42.6 (14.7) 2.2 1.2 P 32.8 (12.1) 2.3 1.2 35.4 (15.1) 2.2 1.1 A 43.4 (14.4) 2.4 1.2 46.1 (14.4) 2.3 1.2 G 32.2 (13.4)” (Table 5)
In the Palin study, gesture typing and word prediction were both linked with fewer keystrokes per character.
“Using gestures and word prediction reduces the amount of keystrokes” (Results, correlation analysis)
In the Palin study, native English users were the fastest group, at 37.8 words per minute, and people who never type in English were the slowest, at 25.6.
“Native users of English were the fastest typists (M = 37.8, SD = 13.6) ... the slowest (M = 25.6, SD = 12.4)” (Results, English experience)
Native speakers of English were the fastest group by English experience, with a mean of 37.8 words per minute on phones.
“Native users of English were the fastest typists (M = 37.8, SD = 13.6)” (Results: Language use)
Participants who never type in English were the slowest group by English experience, with a mean of 25.6 words per minute on phones.
“those who never type in Eng- lish the slowest (M = 25.6, SD = 12.4)” (Results: Language use)
On phones, 75 percent of participants typed below 43.98 words per minute, and the mean was 36.17.
“with 75% of participants having a performance below 43.98 WPM” (Results: Words per minute)
The fastest phone typists in the study reached over 80 words per minute.
“The fastest typists reached over 80 WPM.” (Section 4, Words per minute)
In the phone study, average typing speed was 36.17 words per minute, with a standard deviation of 13.22 words per minute.
“On average, participants typed at 36.17 WPM (SD = 13.22)” (Section 4, Words per minute)
The distribution of phone typing speeds has a skewness of 0.72, a positive skew, so the tail on the fast side is longer.
“It has skewness of 0.72 and kurtosis of 1.13.” (Section 4, Words per minute)
In the Palin study, people who reported word prediction alone averaged 32.8 words per minute, against 34.8 for people who used no intelligent text entry feature.
“none 34.8 (12.6) 2.3 1.2 42.6 (14.7) 2.2 1.2 P 32.8 (12.1) 2.3 1.2 35.4 (15.1) 2.2 1.1 A 43.4 (14.4) 2.4 1.2 46.1 (14.4) 2.3 1.2 G 32.2 (13.4)” (Table 5)
In the Palin study, participants first read and memorized each sentence, then typed it as quickly and accurately as possible.
“Participants were shown instructions requesting they first read and memorize the sentence, then type it as quickly and accurately as possible.” (Task and Procedure)
In a field study that Palin and colleagues cite, people were almost 10 words per minute faster after practice with a gesture keyboard than with tapping.
“When assessed outside the lab, people performed almost 10 WPM faster after practice than using tapping-based input” (Related work)
In the Palin study, people who had taken a touch typing course for desktop keyboards averaged 35.7 words per minute on phones, slightly slower than the 36.4 of those who had not.
“users who reported to have taken a touch typing course for typing on desktop keyboards were slightly slower (M = 35.7, SD = 13.4) than those who reported to not have taken such a course (M = 36.4, SD = 13.1)” (Results, typing training)
In the Palin study, each participant's task was to transcribe 15 English sentences, shown one after another.
“The task was to transcribe 15 English sentences, shown one after another.” (Task and Procedure)
In the Palin study, participants who reported using two fingers averaged 37.7 words per minute, significantly faster than the 29.2 of those who used only one finger.
“Participants who reported to use two fingers were significantly faster than those who used only one finger (M = 37.7, SD = 13.2 versus M = 29.2, SD = 10.7, p < 0.001, d = 0.66)” (Results, finger usage)
Typing with two thumbs was the fastest phone technique in the study, at a mean of 38.02 words per minute.
“this was the fastest way to enter text (M = 38.02, SD = 13.1, p < 0.001 in comparison to all other groups)” (Results: Hands and fingers)
Over 82 percent of phone typists in the study typed with two thumbs.
“Over 82% of participants typed using two thumbs.” (Results: Hands and fingers)
The authors attribute the advantage of two-thumb typing to frequent switching between the sides of the screen, which allows preparatory movements that shorten the gaps between keystrokes.
“The superiority of two-thumb input is attributed to frequent switching between the sides of the dis- play, which allows for preparatory movements that decrease inter-key intervals” (Related work (citing earlier studies))
In the U.S. subsample of the phone study, participants aged 20 to 29 averaged 39.0 words per minute, slightly above the 38.0 for ages 10 to 19.
“10-19 39.6 (14.3) p<.001 38.0 (14.8) p<.001 20-29 36.5 (12.6) 39.0 (13.3)” (Table 4: Performance by demographic group, U.S. subsample column)
In the U.S. subsample, the speed difference between autocorrection users and people who typed normally was not statistically significant.
“These differences are less pronounced in the U.S. subsample where the difference between autocor- rection and normal typing was not found to be significant” (Results, intelligent text entry)
In the U.S. subsample of the phone study, the men and women difference in typing speed had a Cohen's d of 0.22, which the authors treat as a small effect.
“male 36.1 (14.4) d=.003 37.4 (15.7) d=.22” (Table 4: Performance by demographic group, gender rows)
In the U.S. subsample, men averaged 37.4 words per minute on phones.
“female 36.0 (12.5) p=.002 34.6 (12.3) p<.001 male 36.1 (14.4) d=.003 37.4 (15.7) d=.22” (Table 4: Performance by demographic group)
The U.S. subsample used for the matched comparison had 1,475 participants.
“resulting in a subsample of 1475 participants” (Method, participants)
In the U.S. subsample, women averaged 34.6 words per minute on phones.
“female 36.0 (12.5) p=.002 34.6 (12.3) p<.001 male 36.1 (14.4) d=.003 37.4 (15.7) d=.22” (Table 4: Performance by demographic group)
Across all participants, word prediction use had a small negative correlation with typing speed (r = -0.183).
“word prediction has a small negative correlation with performance (r = −0.183)” (Results, correlation analysis)

Pages that rest on this source