How can we compile statistics of accessible apps?
Submitted by Stella on 9 March, 2012 - 13:56
Posted in:
It would be a very powerful tool if we could present Apple and the public with some statistics on how many apps are fully accessible, partially accessible (which is really the same as not accessible but at least shows an effort.) It would not be precise, as we cannot afford to test out every app. However I have found that most statistics rely on a sample so it should be feasible.
Does anyone have any idea how we might be able to do this? It would be a friendly way of possibly waking up Developers or letting Apple know it is a real issue that needs to be addressed.
Stella
#1 AppleVis users say 70& of apps are accessible
I think that any figures derived from this site would be totally unrepresentative.
To illustrate by just how much, go to the advanced search and filter for only apps that have been rated as fully accessible. This currently gives you 749 out of a total of 1328. Include apps that have only minor issues, and the number goes up to 943.
So, according to users of this site, 70% of apps are fully accessible or have only minor issues. That's certainly what the figures would appear to indicate, and 1328 seems like a large enough sample to base some broad assumption on. Personally, I think that we could safely remove the zero and the percentage figure would probably still be too high.
So, if you want figures that are meaningful and representative, don't go anywhere near this site <smile>
But, where that leaves you, I don't know. With over 500,000 apps in the App Store, it's going to be very hard to find any statistics that give an accurate reflection of the situation.
One other issue is that you would probably want to exclude games from any statistics gathering, or at least list them separately, as this category is likely to be far worse than any other, so would seriously skew the numbers.
#2 Excellent points.
After reading your response David, I agree that this site is not set up to give accurate representation. 70% really seems very high. I would want to put emphasis on apps that were fully accessible too instead of partially accessible.
I also wonder about different categories too and wonder if it would be better for me to somehow devise some kind of independent poll that will take all this data into consideration- if that is even possible.
Gads- I really wish I had paid better attention in statistics class.
Any other input from anyone is welcome to help me think about this. :)
Thanks so much!
#3 Collecting Stats
This triggers my geek-response: (I liked my stats class)! :)
To do this right:
Thing one is to establish the categories:
Take the groups Apple has from 'books' to 'weather' I think you guys are right that you can /should expect less out of games and photography and more out of, say, reference or news, so we wouldn't want to collapse the data across all categories at least at first. Within those, it probably makes sense to
determine some sub-groups that we have a reasonable chance of establishing: most popular free, most popular paid and recent releases. I think we can expect different things out of free and paid apps, and to cut across those, the recent releases might be another useful approach. Take the top ten in each subcategory on a particular day. That's
22X3X10
that's 660 apps as a decent respectible sample, but it wouldn't be too bad if you had a group of folks willing to divide up the work. You'd want (ideally) to be sure all the app evaluators had similar ideas of what 'accessible meant;
maybe you'd run a little training or something to make sure you have both reliability and validity... Maybe invite anyone who has submitted X apps to the site to join the fun. (that would also be a neat way of getting loads of apps aboard all at once.
If that sounds like way too much, you could just do the recent releases, cutting down the assessment by a third... and if 220 still seems like too much, you could eliminate a couple categories.
You could even choose and publicize a particular day when the apps would all be downloaded (useful so to ensure you don't get duplicates). It'd be enlightening.