Introduction
The Qwerty keyboard was designed for users to type slower, so keys on a typewriter wouldn't get stuck in place.
But now that we are no longer using typewriters, the Qwerty keyboard is still the standard! Alternative keyboards haven't been too successful either because keyboards with a high learning curve are unappealing and not adaptable.
My teammates and I conducted a longitual study to see if changing the placement of two keys on the mobile Qwerty keyboard would result in a faster typing experience. We decided to implement the changes on a soft keyboard because hard keyboards have been around longer than soft keyboard. We figured that it would be easier for users to learn this new change because soft keyboards are still somewhat new. Our goal was to compare the performance differences between Qwirky and Qwerty on learning curve and the words-per-minute.
DESIGN
We switched the letters t and f because th is the most common pair of letters in the English language. This is called a digraph.
If you look at your keyboard, we are interested in the series of fgh in the home row.
So, when we switch the spots of t and f , t will be closer to h . t and f aren't too far away from each other, so we figured this learning curve shouldn't be that bad. We decided to leave a key between t and h for better accuracy on the mobile keyboard. We don't want fingers jamming into each other.METHODOLOGY
We recruited 14 participants to be a part of our study. We wanted participants to be active users of soft keyboards, so each participant chosen were students who also attended the University of Toronto. Each participant completed 7 sessions, for one session per day.
Participants had to type a total of 28 phrases. The first 14 phrases were typed either on the Qwirky or Qwerty keyboard, depending on which group the participants were assigned to. The other 14 phrases were typed with the other assigned keyboard. There was a 1 - 2 minute break after the 14th phrase and before the 15th phrase.
This experiment was a within-subject study with two factors: the keyboard layout (Qwirky and Qwerty) and the number of sessions (7 sessions). Therefore, our independent variable has two levels: The Qwirky keyboard and the Qwerty keyboard. We compared these two levels to see which keyboard was more efficient so it was only natural for our study to be within-subject. Our two dependent variables were the WPM and accuracy. WPM was one of the most basic determinants for the efficiency of a keyboard.
An Android mobile phone application (app) called Qwirky was created for this study. This app (minimum SDK: API 21 Android 5.0 Lollipop) was installed on an LG Nexus 5 with a 2GB Snapdragon 800 RAM, a 4.95 inch display (70.8% screen to body ratio), and a 1080x1920 pixel screen resolution.
RESULTS
TLDR; Users initially had an average of 18 words per minute (WPM) and an accuracy rate of 82%, this then peaked at 32 WPM and 91% accuracy on or before the fifth session. However, participants experienced a decline for both categories after the fifth session, making the data inconclusive and necessitating further study. The data shows that participants who started each session with Qwirky showed an accuracy learning curve that was statistically significant. However, it was not statistically significant for participants who started with Qwerty. We hypothesize that this is due to priming.
CONCLUSION
Our results showed that although participants showed an increase in WPM and accuracy for the first four sessions, both categories (Qwerty and Qwirky) decreased starting from the fifth session. This may be due to fatigue and an attitude of indifference in performance towards the end of the study. Thus, the data is inconclusive. We failed to reject the null hypothesis. Further experiments (ideally with more sessions) and a better keyboard layout are necessary to improve the study. For the complete paper and study on our Qwirky keyboard, please click here.