When Jane (not her real name) was in her fifties, the supply of oxygen to the area of her brain responsible for language processing was interrupted. The stroke left Jane with aphasia, impairment in using language. Jane can understand spoken language well, but speaking and reading are very difficult. Like so many people, Jane uses the Web to access information, but the process is slow and painful. As Jane expresses it, when confronted with a cluttered page of information, “I have to read everything.” That is, Jane has lost the ability to skim text visually. Watching her work her way, word by word, through irrelevant text, looking for the information she needs, makes one sharply aware of how heavily computer use relies on rapid skimming, an ability most of us use without ever being aware of it.
In a consultation on ways that improved technology could help people with aphasia use computers more effectively, Jane had her first experience with text-to-speech software. The technology allowed her to select text on the screen and have it read aloud. While the software didn’t restore her ability to skim, it greatly enhanced her ability to understand the text she needed to read. Her excitement was visible in expression and body language as she struggled to say, “I haven’t seen these words since my stroke.”
Over 22 million people in the United States have difficulty with some cognitive functions, including memory, language, and learning issues. World demographic numbers for cognitive disabilities are uncertain, but 500 million is a plausible estimate. Information and communication technology (ICT) promises substantial benefits for these people, many of whom are computer users; a study by Microsoft indicates that 16 percent of computer users in the U.S. have a cognitive disability. Happily, many of the techniques that improve access to ICT for other groups also deliver value for people with cognitive disabilities. Here are some of the approaches that usability professionals can take to improve the usefulness of their products for these users.
Text-to-speech reading aids are valuable.
Screen readers for the blind play an obvious role in accessibility for those users. Less obviously, as Jane’s case shows, related tools can help people with cognitive disabilities process text. Many people, like Jane, have trouble reading, not because of visual problems, but because of trouble decoding the text. Hearing text read is a big help. A corollary is that attention to those aspects of application design that support screen reader access—not rendering text as a graphic for example—is important for users with cognitive disabilities as well as for those with visual impairments.
Clear, simple language pays off.
Writing that is clear, to the point, and expressed without unnecessarily fancy vocabulary makes applications easier to use for people with cognitive disabilities (and for other people, too). Unfortunately, this basic design attribute is tricky to measure or express quantitatively. Reading level formulae may seem like good tools here, but as Ginny Redish has pointed out, they can backfire seriously. There are two different reasons for this:
- First, the tools don’t distinguish the situations in which big words are actually needed from those in which they are not. Trying to replace standard technical terms with words that seem simpler, but are unfamiliar, will make text harder to understand, not easier.
- Second, reading level formulae work by correlations between various attributes of text, like sentence length and readability, but the correlations are based on naturally occurring text, not text that has been modified artificially to change the attributes. While it is true that among naturally occurring texts, those with shorter sentences tend to be easier to understand; if you take a text with long sentences and chop the sentences up, the sentences will be shorter. But the resulting text is often harder, not easier, to understand.
Usability professionals are well placed to respond to these difficulties. They are used to the reality that design guidelines are only a good start, and that user testing and attention to user feedback is essential to high-quality results. The same principles work here, too: see if people can actually understand the text that is included in your applications.
Another design feature that broadens the range of users who can use an application is glossary support. An easy way to see an explanation of a jargon term or acronym will be appreciated by many users.
Text structure and layout are also important. People who read with difficulty benefit from headings and layout that make it easier to identify what text they really need to process. This is especially important for people who can’t skim, like Jane.
Simple, clear navigation is crucial.
While data from users with cognitive disabilities is in short supply, studies that have been done suggest that the usability problems such users encounter are similar to those typical users find, but with greater impact. For example, any user may be confused by a back button that doesn’t work on a web page, but users differ in how well they are able to recover from, and cope with, the problem. A person with a cognitive disability may not be able to get back on track. Here again, usability professionals already make it their business to identify and eliminate these and similar difficulties in navigation. And, as for comprehension problems, they recognize the need for user testing to achieve good quality.
The same considerations apply to other aspects of interaction besides navigation. The canons of good design for usability, including eliminating unnecessary choice points and providing clear feedback on progress, positive or negative, apply with greater force to users for whom problem solving is difficult.
Allow users to tailor the presentation.
This is another point of overlap with other aspects of inclusive design. Control of font sizes, colors, and contrast is helpful for the many people who have visual impairments as well as cognitive limitations, and usability professionals are already aware of the importance of supporting this. Going further, allowing users to control the amount of detail that is presented can also be helpful. A New York Times story in 2008 reported that some users are choosing to use the mobile versions of some websites because these versions present just the most important information and controls. A design that lets users request this kind of presentation directly would be step forward.
Some historical perspective may be useful here. Old timers will recall early design guidelines that called for very limited information to be presented per screen or menu, often following a mistaken interpretation of George Miller’s classic “Magical Number Seven Plus or Minus Two” paper. Now we understand that for typical users, placing as many options on a page as will fit is actually helpful because it avoids the problem of interpreting abstract descriptors that stand for subsidiary choices (does “clothing” include “shoes?”) and because it requires fewer user actions to complete a given task. Mainstream sites like Amazon.com demonstrate this design direction dramatically by placing literally hundreds of options on a single page, when dropdowns are included.
For some users, however, this design direction works poorly. If you have trouble skimming, as many users like Jane do, you will have trouble finding the option you need among hundreds of alternatives. A presentation option that lets you suppress the rarely needed options will be helpful, but configuring your system with the presentation options that work for you is a challenge. Some user interface technology projects, like Fluid (www.fluidproject.org), include efforts to create information presentation preference profiles that can be created once, perhaps with some help, and then automatically applied to many different web pages or applications.
Rely sparingly on mental models.
This is another area of historical evolution in UI design. A good deal of early work was done on the value of letting users think their way through problems in using a system by reasoning about a mental representation of how the system works. Today this approach is rare, at least as an explicit feature of design, reflecting the unwillingness of most users to invest the work needed to develop a workable mental model.
But too-optimistic assumptions about user understanding still do sometimes arise, especially where the world of the user intersects technical topics, like security, or wireless configuration. Here it is easy for designers to forget that users have little understanding of what is really happening, and offer choices that will mean little to most users. Do you want to accept this security certificate? What is a security certificate, anyway? What does it mean to accept it? What is a DNS?
A well-designed system should not require users to possess special knowledge to successfully use it. Good design, then, means shielding users from things that they don’t understand, or, if that isn’t possible, providing adequate explanations of their choices. We’ve all seen progress on this front where security is concerned; many systems make a serious effort to explain the choices users must make and what they mean in non-technical terms. And wireless configuration is increasingly automated, requiring few if any user choices in many situations.
These considerations are even more important for users with cognitive disabilities. As for other usability problems, an issue that is an inconvenience or annoyance for a typical user can be a showstopper for someone who finds it hard to understand complex situations or to quickly pick up unfamiliar concepts.
Attention to inclusive design is increasing around the world. The recently adopted UN Declaration on the Rights of Persons with Disabilities (www.un.org/disabilities/convention/conventionfull.shtml) marks a new global resolve to provide more support for full participation by people with disabilities, including in the realm of ICT. This resolve includes attention to the needs of people with cognitive disabilities. A few years ago, it was rare for these needs to be considered when accessibility is discussed, but now organizations like WebAIM (www.webaim.org), the Rehabilitation Engineering Research Center on Advancing Cognitive Technologies (www.rerc-act.org), and the Raising the Floor Initiative (raisingthefloor.net) are organizing work aimed directly at the needs of people with cognitive disabilities.
In responding to this challenge of inclusive design, there is a pressing need for more inclusion of people with disabilities, including cognitive disabilities, in the design and development process. Usability professionals should make a point of recruiting people with disabilities for focus groups and for user test panels. Shawn Henry, in her 2007 book Just Ask: Integrating Accessibility throughout Design (online at www.uiaccess.com/accessucd/index.html) has provided an excellent guide to doing this.
Usability professionals can do more than optimize the accessibility of their projects. This article draws on the work of an informal committee, convened in connection with the recent Web Content Accessibility Guidelines revision project of the Web Accessibility Initiative by W3C. Usability professionals can help push this work forward.
In particular, the committee needs examples of websites that demonstrate the full spectrum of inclusive interaction design, from sites that fully support people with disabilities, to sites that show little or no consideration for disabled people. If you want to help further this work, please contact me.
Preparation of this article was supported by the Rehabilitation Engineering Research Center for Advancing Cognitive Technologies and the Coleman Institute for Cognitive Disabilities.