elkraneo’s Twitter Archive—№ 4,430

Most people don't know about the accessibility features on their phones, and some may not think of themselves as having a disability, discarding features that could still help them 🧵 #A11y #Accessibility
On twitter.com 2022 Oct 20 Mood 0

…in reply to @elkraneo
Accessibility features can be seen as mechanisms that allow user interfaces to be personalized, and a system that recommends them based on how a person uses a device can demonstrate to be very helpful
Permalink On twitter.com 2022 Oct 20 Mood +5 🙂

…in reply to @elkraneo
The paper When Can Accessibility Help? shows that this is the case. Based on a survey with 100 participants, the researches designed detection strategies and prototypes to collect insights from 20 older adults machinelearning.apple.com/research/recommending-accessibility
Permalink On twitter.com 2022 Oct 20 Mood +2 🙂

…in reply to @elkraneo
From the survey, 4 detection strategies were identified:
Permalink On twitter.com 2022 Oct 20 Mood 0

…in reply to @elkraneo
1. Statistical: identifies differences in users’ behaviour over time
Permalink On twitter.com 2022 Oct 20 Mood 0

…in reply to @elkraneo
2. Near-Miss: monitor features that rely on a preset threshold or multiple conditions to be reached before triggering
Permalink On twitter.com 2022 Oct 20 Mood +1 🙂

…in reply to @elkraneo
3. Action Sequences: based on the known connection between a specific series of behaviours and a feature that might be useful
Permalink On twitter.com 2022 Oct 20 Mood +2 🙂

…in reply to @elkraneo
4. Grouped: recommends accessibility features based on the ones the user already has enabled
Permalink On twitter.com 2022 Oct 20 Mood +2 🙂

…in reply to @elkraneo
During the process, participants commonly said they would ask a doctor if they were impaired. For example, if their eyesight prevented them from reading on-screen content, they responded that they would go to an optometrist to get glasses
Permalink On twitter.com 2022 Oct 20 Mood -3 🙁

…in reply to @elkraneo
But even though human experts are able to match access technology with a higher degree of certainty, there are many reasons why this method may not lead to the widest adoption of accessibility features (e.g.,time/money, lack of knowledge).
Permalink On twitter.com 2022 Oct 20 Mood -2 🙁

…in reply to @elkraneo
With the detection strategies in place, prototypes recommenders were built and the final user study was conducted. 77.8% of participants owned a smartphone, and on average, they owned their smartphone for 2.3 years
Permalink On twitter.com 2022 Oct 20 Mood 0

…in reply to @elkraneo
The final results showed that awareness of accessibility features was much lower than in the baseline study. (10.5% of older adults knew what “accessibility features” were, compared to 90% of baseline participants)
Permalink On twitter.com 2022 Oct 20 Mood 0

…in reply to @elkraneo
Participants, even if some didn't identify as having a disability, observed that the recommenders pointed to potential useful features under the accessibility menu (where they would not have thought of looking)
Permalink On twitter.com 2022 Oct 20 Mood 0

…in reply to @elkraneo
Almost all participants (89.5%) were open to receiving recommendations, with most preferring low frequency surfacing methods (e.g., home screen, email, or a message) that did not interrupt their current task
Permalink On twitter.com 2022 Oct 20 Mood +2 🙂

…in reply to @elkraneo
The categorization of the 50+ accessibility features present on iOS and the idea of data-driven methods of identifying accessibility recommendation triggers (in a privacy-preserving manner) is inspiring
Permalink On twitter.com 2022 Oct 20 Mood +3 🙂

…in reply to @elkraneo
And even if some features require a system level implementation, many learnings can be applied already per app basis. Many thanks to the authors 🦾
Permalink On twitter.com 2022 Oct 20 Mood +2 🙂