Screen Reader User Survey #6
The results of our 6th WebAIM Screen Reader User Survey have been published at http://webaim.org/projects/screenreadersurvey6/.
We received our largest response ever with 2515 survey respondents. Thank you for participating!
ZoomText and Window-Eyes Data
The most surprising data is certainly the significant increase in the usage of ZoomText and Window-Eyes.
Reported usage of ZoomText as respondents’ primary screen reader skyrocketed in the last 18 months from 1.3% in January 2014 to 22.2% in July 2015. ZoomText is used as a primary screen reader as often as both NVDA (14.6%) and VoiceOver (7.6%) combined!
Window-Eyes increased as a primary screen reader from 6.7% to 20.7% in the last 18 months. As a result, there were significant decreases in the usage of JAWS (from 50% to 30.2%) and notable decreases in usage of NVDA (18.6% to 14.6%), VoiceOver (10.3% to 7.6%), and System Access (7.7% to 1.5%) as primary screen readers.
There are currently 5 different screen readers that are commonly used by more than 25% of respondents. This is both exciting and challenging.
Really?
I’ve already heard incredulity among peers in the web accessibility field regarding these numbers. Great care should be taken in discounting the validity of this data. 2515 respondents is not a small sample.
Perhaps the most prevalent criticism of the WebAIM surveys has been that they are not representative enough. An opt-in survey will never be fully representative of all screen reader users. With that said, this survey had broader distribution than ever before. It reached small schools for the blind. Many heard about it via word of mouth. We believe the responses to be more representative of the overall screen reader user population than ever before. The fact that in this survey we see a reduction in reported proficiency is consistent with the notion that our sample has shifted from prior years.
We speculate that much of the shift in the screen reader market numbers can be attributed to this broader sample. Respondents that reported lower proficiency were more likely to use ZoomText.
Those of us in the web accessibility field tend to interact more with screen reader power users and users that are more connected both with technology and with our technical field. This would naturally cause us to believe that the overall market reflects the demographics of those screen reader users. These survey results suggest that those of us in the web accessibility field may need to reconsider our perceptions of screen reader users – the typical user may be different than our own interactions and experience suggest.
What is driving these changes?
Window-Eyes became freely available to Microsoft Office users shortly after the last screen reader user survey closed in January 2014. AI Squared (who develops and markets ZoomText and Window-Eyes) has made significant marketing efforts in the last year. These factors certainly contribute to their increased usage.
Additionally, with a much broader dissemination of the survey, more existing ZoomText users completed the survey. This simply suggests that ZoomText users were probably underrepresented on previous surveys. On the other hand, AI Squared did promote the survey heavily, though it was also prominently advertised on mailing lists and publications for other screen readers.
It should be noted that ZoomText functions as both a screen magnifier and screen reader. We recognize that some respondents to the survey may only use ZoomText for magnification and not for the screen reader functionality. Only 13.4% of ZoomText users reported being blind (though 92.3% reported low vision). However, we don’t believe this consideration should at all change how we consider ZoomText as being a very commonly used access product. In fact, because the screen reader functionality of ZoomText is often used as a supplement to magnification, this strengthens practices that would ensure an equivalent visual and screen reader experience.
What does this mean?
We must recognize that ZoomText and Window-Eyes are significant players. These tools are often not part of web accessibility testing methodologies. They should be. We often consider VoiceOver with Safari as a testing platform and not Window-Eyes or ZoomText when in fact Window-Eyes with IE, ZoomText with IE, and ZoomText with Firefox are all more common combinations than VoiceOver with Safari.
This recommendation for increased consideration of ZoomText and Window-Eyes is strengthened by the fact that those who use these screen readers tend to be less proficient and also to use only one screen reader, whereas more advanced users tend to use multiple screen readers. In other words, lack of consideration for ZoomText and Window-Eyes support can have a notable impact on many users – and these are the users that are less likely to adapt to insufficient support.
As usual, very interesting reading, especially because of the evolution of the market that this survey now allows us to track.
One aspect I would find very interesting for future editions of the survey is whether attempts could be made to broaden the geographic reach, e.g. by working with local blindness organizations (or interested users) to translate the text of the survey into additional languages, and to disseminate the links even further to non English-speaking countries. In particular, this would give the option of investigating trends on a more regional basis.
For example, from my own experience I would expect the prevalence of mobile screen readers to vary widely between regions, to a large extent reflecting different use patterns of the Mobile Internet across the world.
Best regards,
Marcus Gröber
(co-developer of Nuance TALKS)
Is there a way to discern from the survey how many of Winodw-Eyes users are using the version that is free if you have Microsoft Office 2010/2013? It would be interesting to see if any left NVDA when that offer became available.
Sean –
It’s difficult to really know for sure. 62% that use Window-Eyes as their primary screen reader indicated that they purchased it. Much higher than the 33% for JAWS users. 8.1% indicated that they downloaded Window-Eyes. Of course they may have indicated that they purchased it because they purchased Office in order to get the Window-Eyes license.
Hey Jared,
I tend to agree, the larger sample rate (still slightly skewed, but getting much better) is likely one of the reasons why the numbers seem ‘strange’ this year. I think that a lot of us “professionals” also tend to overlook the Low Viz community too often: witness Wayne Dick’s on-going efforts to bring Low Viz issues to the fore (and a HT to Denis Boudreau for taking up that clarion). We already know that the low vision community is roughly 6 times larger than the blind/non-sighted community, so why should this be such a shocker? Denis recently sited the following: “285 million people are estimated to be visually impaired worldwide. Of which, 39m are blind and 246m have low vision. I did the math: there are precisely 6.30769239 times as many people who have low vision than there are people who are blind.” (http://www.creativebloq.com/web-design/accessibility-beyond-screen-reader-91516540) Why then are we ‘surprised’ when their numbers start to actually show up in your survey? (It could be argued that ZoomText isn’t a screen reader per-se, but that would be a silly argument, as it is a tool that low-vision users use, and it includes both magnification and reading, amongst it’s many features.
I also recall having dinner once with a woman in California, who’s job was to provide on-site training to new non-sighted users who for whatever reason were just starting out with a computer/screen reader combination. The gig was a contract with the State of California, and I recall her telling me then that the state “recommended” (read: funded) the use of Window-Eyes (over JAWs), and so that is what she taught each day. These weren’t ‘power-users’ but rather users at the other end of the spectrum – new users grappling with new technology for the first time. Think they have time for an on-line survey about screen-readers? (Na, me neither)
Finally, Window-Eyes itself has seen a ‘second chance’ – yes, for the longest time it was the screen reader that was sneered at (certainly in North America by “experts” – and I’ll admit to being guilty of that too), but the combination of it being acquired by AI Squared (a solid company with the right kind of commitment, and some rock-star engineers too), the ‘partnership’ with Microsoft (remember iOS/Mac fans, most blind users are of limited financial means, and maybe, just maybe they can’t stand in line at the Apple Store for the latest iToy – the superior functionality of VoiceOver not-withstanding), and yes, maybe AI Squared helping get the word out to more mainstream respondents helped tip the scales, but I’d rather think (which I gather so do you) that the scales have been ‘righted’, and that for various reasons over the years the survey results were slightly tipped in favor of the power-user.
However, the one thing that must be said is this: thank you WebAIM for running these surveys. We need data, more data (lots more data), and/but you guys have been solidly and consistently providing data for a long time now, and I just want to add my voice to the chorus that says – THANKS!!
Hello. Very interesting survey! I am very glad, as a Window-Eyes user, and certifier, yes, I am now Window-Eyes Certified, that it is now being given a second chance. AI Squared defenatly is doing the right thing by marketing Window-Eyes, and I hope that now with this partnership with Microsoft, and the product being made by AI Squared, Window-Eyes can now be included as one of a range of options, and I hope Window-Eyes can gaim the foothold in the market. It is certainly true that Window-Eyes has been the screen reader that has been smeared at especially by experts in North America. Luckilly I have not been one of those people who smeared at Window-Eyes. Part of the smearing stended in part by a few things, one of which was how slow it took Window-Eyes to get Aria support on the website.
I would like to congratulate Window Eyes and ZoomTech for their deserving resurgence. Also, WebAim must be heartily thanked for such statistical data that strives to be adequately representative of the actual end user base.
I am attributing ZoomTech’s leaps to the survey respondents comprising of more and more partially visually impaired people. Could Magic be included as a choice in future surveys?
Also, I have a small suggestion to make. Could the underdeveloped and underrepresented continents be made more enlightened of the WebAim survey? Coverage of these geographies would significantly increase the importance and usage of free and open source screen reading products (NVDA, Talkback, Window Eyes etc.).
I, being an Indian NVDA user, am aware about the striving efforts made by major NGOs to popularize and promote the free and open source screen reader NVDA. There is an NVDA development team as well in India that works full time on contributing code to NVDA. JAWS held dominance prior to NVDA, and Window Eyes and ZoomText at present are almost, if not absolutely non-existent as far as India is concerned. I would even say that Dolphin Supernova would be used by a base greater than that of Window Eyes and ZoomText in India. Interactions with folks from developing nations also has enlightened me about the actual quantity of NVDA users.
Participation from these underprivileged geographies would make WebAim’s collections and results more reflective, in my opinion. Although WebAim has apparently received inputs from the beginner segment this time, I am saddened to see the minimal responses from Asian, African and Carribean regions.
Thanks for reading.
Best Regards,
Bhavya Shah.
This is interesting, though the size of the sample would render these results statistically insignificant especially since there were no controls for internal or external variables (which I am sure everyone knows). To me any product that requires a certification is actually counter to the inclusion philosophy. Its why I hope that AT that utilizes more intuitive design becomes the norm versus where the industry currently resides.
I’m not very surprised, it has long been known that there are more people in the low-vis category compared to (full) blindness. I’ve also noticed that people stick to magnification as long as they can, as learning a screenreader can be quite daunting.
So I think it is case of congratulations to WebAim for increasing the reach of the survey!
The only concern is that Windows Eyes is (or has been?) far less capable than the other popular screen readers.
John’s comment give me hope, I haven’t tried it for a while, is it’s ARIA support improving?
One thing that has not been mentioned that may have contributed to some of the drop in JAWS users recently is the reworking of JAWS keyboard commands. In JAWS 15 many of key commands changed. Could it be some JAWS users figured, “if I have to learn to use JAWS again, maybe I’ll try something else?”
Myself, I’ve moved away from JAWS as my primary screen reader, to ChromeVox for day to day screen reader Web accessibility testing, the latter with better ARIA support, and much easier to learn to use, albeit less functional than JAWS and less likely to be used if you are blind. I do however, end up testing with JAWS before a product goes into production, which often results in having to go back and build work-arounds to accommodate JAWS less than optimal support for ARIA.
Perhaps in next years survey there could be a few questions regarding changes in preferred screen readers, something along the lines:
-Did you change your preferred screen reader over the past year (or two)?
-From which one to which one?
-Why did you leave the old one: too expense, difficult to use, does not support ARIA well, using a different OS now, poor voice quality, frustration,…
-Why did you choose the new one: less expensive, easier to use, others recommended, better voice quality, ARIA support is better…
-Comment of your reasons for switching preferred screen reader.
etc.
Thanks for this survey BTW. Its a great resource!
I will make a couple of points that haven’t been covered already. As a very long-term Window-Eyes user the long hiatus in upgrading of web support was a bit frustrating and improved PDF support is still to happen. Hence my switching back and forth between WE and NVDA. I spend much of my working day evaluating and writing HTML, CSS and JS code. Based on conversations with other blind people I know, my awareness of what is happening on websites and of a screen reader’s potential does not equate to most users. While “experts” sneered at Window-Eyes for lacking ARIA support, none of the blind friends I have quizzed know what an ARIA landmark is. That is a whole topic in itself.
The other really important point, which seems to escape many in the web accessibility field, is that a screen reader must do much more than provide access to websites. To even get to first base, it has to provide ready access to the operating system. Then there is writing and reading emails; reading, writing and editing word processor files; using spreadsheets and databases; accessing audio/video software; etc etc. Most of us (including this geek) spend more time doing things other than browsing websites when using computers. As important as it is, judging a screen reader purely on the access it provides to the web is therefore neglecting the majority of its duties. Another important consideration is how usable a screen reader’s access to applications is. People should not have to undertake something approaching an IT qualification to get value from their assistive products. Yes, I do wish more people would read documentation :). Perhaps however one reason for real users moving to Window-Eyes is that its long-awaited improved web access did not involve a huge learning curve. Clever software should make the decisions and not rely on the user having to choose from a vast list of options.
Andrew
I agree with most of the comments.
My concern is also that Windows Eyes has been until now far less capable than the other popular screen readers, but the main pb is the lack of PDF support.
WE is the “default” screen reader for many people…
I’d gladly switch to a screen reader other than Jaws, which I use at work, if Window-Eyes or NVDA could work with the database we had specially developed. Unfortunately NVDA crashes the database, and Window-Eyes isn’t able to read anything in the DB. Jaws is the only screen reader that works with this DB, which uses the Access 2003 runtime. I’m stuck using Jaws 15 because Jaws 16 and 17 both crash my system when I’m reading a PDF and alt tab away from the PDF. This has been going on for a year now, and FS says it’s an Adobe issue, and Adobe isn’t able to seem to fix it! Frustrating. At home my main screen reader is VoiceOver on an iMac Late 2009, which I love! 🙂