A 2019 study showed that people who use autocorrect type faster, but other methods may be more accurate.
People who enable autocorrect and who use both thumbs type faster on a smartphone. These key findings are from a report, How do People Type on Mobile Devices? Observations from a Study with 37,000 Volunteers, released in October 2019 by a team of researchers from Finland, Switzerland, and the United Kingdom. The researchers analyzed the results of a web-based typing test combined with participant-reported information.
The authors “classified the participants’ typing into four different technique categories: autocorrect, word prediction, gesture keyboard, and plain typing.” The fastest results–at just over 46 words per minute–were from participants who used autocorrect alone. Every other technique–or combination of approaches–resulted in slower typing speeds.
SEE: 2020 Tech conferences and events to add to your calendar (free PDF) (TechRepublic)
1. How to enable autocorrection
To enable keyboard autocorrection on iOS and iPad OS 13, tap Settings > General > Keyboard, then move the Auto-Correction slider to the right to enable it. Android device settings vary, but as an example, on a Pixel 3a with Android 10, tap Settings > System > Languages & Input > Virtual Keyboard > Gboard > Text Correction, then move the Auto-correction slider to the right (Figure A).
Autocorrect fixes many obvious errors. For example, on both iOS and Android, when you type “nwe hpoe” with a system that uses English, the system will change your words to “new hope.” The researchers noted that even though prior research indicated that “autocorrection can be detrimental to performance because of [the] high cost of erroneous corrections,” their analysis found that “participants using autocorrection have the highest performance in our dataset.”
2. How to utilize Smart Compose or speech recognition
Google’s Smart Compose–first available in Gmail and later in Google Docs–offers phrase prediction instead of word prediction (Figure B). The researchers mention that “a recent study showed decreased performance rate for heavy use of word prediction” and that the issue would need “more detailed analyses to better understand the usefulness of ITE [Intelligent Text Entry] in different contexts and for different users.”
Smart Compose is on, by default, for everyone who uses G Suite. When a phrase displays in Gmail on either Android or iOS, swipe across a displayed phrase, and the system types it in. Phrase prediction offers another potential area for further research.
Speech recognition may be as much as 2.9 times faster than typing–at least for short messages–with input speeds of up to around 150 words per minute. Those speeds were reported by a team of US-based researchers, Comparing Speech and Keyboard Text Entry for Short Messages in Two Languages on Touchscreen Phones, published in 2017.
SEE: How to use speech recognition to improve productivity on your smartphone (TechRepublic)
To try voice typing on a smartphone, tap the microphone icon displayed on your iOS or Android on-screen keyboard (Figure C), allow system access to the microphone (if prompted), then talk.
3. How to determine when to disable autocorrect
People who use technical words or terms may want to turn autocorrection off–or consider using an external keyboard (Figure D). The word “auto-correct” itself demonstrates a problem with the feature. As quoted above, the researchers use the term “autocorrection,” yet display the term “auto-correction” on the site about the paper. In Android 10, Google labels the feature “Auto-correction,” while Apple, in iOS 13, lists it as “Auto-Correction.” So which is it? One word, a hyphenated word, or a hyphenated word with two capital letters?
SEE: How to choose the right mobile keyboard for faster typing on your iPhone or iPad (TechRepublic)
Company, brand, product names, and technical words all present problems for autocorrection systems; few accurately help you input names, such as HP Chromebook 14b x360, Lenovo 300e Chromebook (2nd Gen), and iPad Pro 12.9,” or terms, such as CNAME. As of late 2019, autocorrect and voice recognition systems struggle with such unpredictable patterns of numbers, punctuation, and capitalization.
Both autocorrect and speech recognition work well for most standard language messages or texts but may capture technical terms with less accuracy. To convey technical details, you might tap out terms, letter by letter–either on-screen or with an external keyboard.
What has your experience with autocorrection, speech recognition, or Smart Compose on mobile devices been? Do you use these features for certain types of communications? What input method do you prefer when you need precision? Let me know, either in the comments below or on Twitter (@awolber).