By Roger Hudson, Russ Weakley and Peter Firminger
Most recent version of article: 30 January 2005
Originally presented at OZeWAI 2004 Conference, La Trobe University, Melbourne, Australia – 2 December 2004.

Introduction

Web accessibility and the notion of universal design are laudable and for many disabled people have resulted in significant benefits. Well made sites allow people with a range of physical disabilities to access goods and services and participate in activities with an ease that was denied them in the pre-web world.

However, the needs of the largest disability group in our community, those with cognitive disabilities and learning difficulties, appear to have slipped through the cracks to a large extent when it comes to website accessibility.

Our accessibility journey started in 2002, when we first worked together on sites for the Australian Museum. It followed a pretty standard path. We did multiple assessments with people with different vision impairments and a few with people who had other physical disabilities. When it came to considering individuals with cognitive and learning difficulties however, we did little more than give lip service to their needs.

We were not alone in this. Many people associated with web accessibility have followed the same route, including, if we may be so blunt, the W3C Web Accessibility Initiative. Although somewhat simplistic, a comparison of the following two Web Content Accessibility Guideline Version 1 checkpoints [1], which relate to the use of images on the web, maybe indicative of the situation.

“Provide a text equivalent for every non-text element”

Checkpoint 1.1

“Supplement text with graphic or auditory presentations where they will facilitate comprehension of the page”

Checkpoint 14.2

WCAG 1.1, which is fundamentally about improving access to websites for people with vision impairment, is a Priority 1 checkpoint (ie “A web content developer must satisfy this checkpoint”). However, WCAG 14.2, which suggests the inclusion of images to help make pages more comprehensible (an issue of particular importance for those with cognitive disabilities), is a Priority 3 checkpoint (ie “A web content developer may address this checkpoint”).

The aim of this paper is to offer some ideas on how websites might more effectively meet the needs of people with cognitive disabilities and learning difficulties. The paper will look at three issues:

• How the presentation of page content can be modified to make it more accessible.
• Design of site navigation systems.
• Tailoring content to the needs of different audience groups.

Note: We recognise the web can bring considerable pleasure and aid to people with different, and in some cases quite profound, cognitive disabilities. The focus of this paper however, is primarily on improving the web for people who have the functional capacity to independently access and use sites than contain some text content.

The Peepo project website, which was recenlty closed, [2] provided a wide range of resources and ideas to enable people with severe learning difficulties to browse and use the web independently.

Description and Action

Much of the published material relating to cognitive and learning difficulties appears to be written by researchers and clinicians and is not directly concerned with the issue of website accessibility. Also a significant amount of the material that does deal specifically with web accessibility for people with cognitive disabilities, concentrates mainly on describing and analysing the situation [3] and/or promoting an awareness of the different types of disabilities and the need to do more to address them [4].

The article, “Designing for Users with Cognitive Disabilities”, [5] provides a good overview of the subject and makes a number of suggestions for improving site accessibility in this regard. The article, which was written in 2001, concludes with a look to the future and suggests, “Web designers need to specialize the products for the cognitively disabled according to their needs“.

The Lisa Seeman article, “Designing Web Content for People with Learning Disabilities” [6] also provides practical suggestions and useful examples of ways to improve the accessibility of sites.

Two WebAIM articles earlier this year, drew attention once more to the importance of increasing our knowledge of the web-needs of people with cognitive disabilities and doing more to address them. The first article, “Cognitive Disabilities Part 1: We Still Know Too Little, and We Do Even Less”, [7] suggested a number of ways web content could be made more accessible for people with cognitive disabilities.

The second WebAIM article, “Cognitive Disabilities Part 2: Conceptualizing Design Considerations”, [8] sought to describe the most common difficulties individuals with cognitive disabilities have to overcome and suggests these could be presented in the following categories:

• Perception and processing
• Memory
• Problem-solving
• Attention

With this paper, we aim to build on this earlier work. Like the WebAim articles, we are primarily concerned with the problems people with cognitive and learning difficulties might have when using the web and offering a few practical suggestions on how these problems might be addressed.

Web Writing

Most websites are text based, and so the words are good place to begin. The Web Content Accessibility Guideline 14, referred to earlier, suggests the words should be clear and simple so that they may be more easily understood.

Before considering written content for the web, we would like to make it clear that we appreciate that some people who access the web are unable to read and so discussion about the written content of a page is not likely to be pertinent to them. Conversely, some people with cognitive and learning difficulties read very well. It is clear however, that there are many people who can read, but whose reading ability maybe compromised for a variety of reasons.

Content that is well written, and more importantly written for the medium of the web, will be easier for everyone to access including people with cognitive and learning difficulties. A detailed discussion of writing for the web is not the focus of this paper however, the key points to consider include:

• Make sure the information contained within each web page is well organised.
• Keep it short. People don’t ‘read’ web pages in the same way they read printed documents. Website visitors rely heavily on skimming and scanning techniques to find areas of interest quickly.
• Use the “inverted pyramid” style of writing adopted by most newspapers. Start with a summary or short overview of the issue and the outcome, and then provide the supporting information and background details.
• Break the information up into small chunks, with one key idea per paragraph.
• Present related points in a list rather than a long paragraph.
• Use meaningful headings and subheadings.
• Keep it simple. Site visitors should be able to understand what is written on the first reading without having to stop and re-read sections.
• Check for spelling and grammatical mistakes as they can make the content harder to understand and diminish the integrity of a website.
• Provide definitions/explanations of technical terms, abbreviations and acronyms.

These suggestoins for improving the effectiveness and accessibility of written material on websites are reflected in the Web Content Accessibility Guidelines. We would like to mention three other specific issues that relate directly to web users with cognitive and learning difficulties.

• Line length. All readers find long lines of text hard to read and for people with reading problems they can become a significant barrier. As screen resolutions increase, it is possible to get more and more characters into a line at any given font size, however the optimal font size for ease of reading varies from reader to reader. As a result, it is difficult to be definitive about what is the best line length, but as a general rule, lines should not exceed 70 – 80 characters and the text should have a left and right margin.

• Rivers of white. Many web users with reading difficulties have problems with text that is both left and right justified. The uneven spacing between words in fully justified text can cause “rivers of white” space to run down the page making reading difficult and in some cases impossible for some people.

• Alternatives for non-literal text. Some text such as colloquialisms and metaphors can have a literal meaning that does not match the intended meaning within the context of the document. For example, “it’s raining cats and dogs”. People with some cognitive disabilities (eg autism/Asperger syndrome, Semantic Pragmatic Disorder) can become confused by these texts since they tend to focus on the literal meanings of words, which in the example above is pretty meaningless for what has rain got to do with cats and dogs. Lisa Seemen suggests the W3C Ruby Technology [9] offers a potential solution to this problem since it allows literal translations to be provided for non-literal text [10].

Readability

The WCAG checkpoint 14.1 advises, Use the clearest and simplest language appropriate for a site’s content. [Priority 1] To achieve this, the content developer must have a clear understanding of the target audience or audiences for their material and the ability to determine the language level, or readability, of the document. While an awareness of the audience for a site, their needs and abilities is very important, it is beyond the scope of this paper. We feel however it is important to comment on the second requirement, determination of the readability of a document.

A large number of people in Australia have literacy problems. In 1996, the ABS found nearly 20% of Australians aged between 15 and 74 had very poor literacy skills, that is, they could be expected to experience considerable difficulties in using many of the printed materials encountered in daily life and at work [11]. The possible causes for this lack of literacy (in the English language) are varied, for example different language background, missed school, brain injury as the result of an accident or stroke, learning difficulties etc. Regardless of the cause, literacy problems are likely to reduce the individual’s ability to access and understand the written content on a website.

Readability tests might be considered controversial in some areas, such as the assessment of educational material, but they can be very useful tools for web content developers. Readability tests measure those features of a text that can be analysed mathematically, for example the average number of words in a sentence or the number of syllables in the words. The results are used to indicate how easy or difficult a document is to read, and usually provide an estimate of the reading grade or age required to understand the content [12].

The different readability formulas however, cannot measure how comprehensible a document is and so will not tell us if an individual will be able to understand the ideas it contains.

In spite of this important limitation, readability tests can provide a useful indication of how easy a document is to read. For example when the content of the CentreLink, “Services for People with Disabilities” page [13] was checked, the readability of the content was found to be about equivalent to that of an academic paper, which would require about eleven years of schooling to understand.

On the other hand, a similar test of the Services South Australia page, “Finding a job, am I entitled to welfare assistance?” [14] revealed readability at about the level of a popular novel, which would take about 6 years of schooling. No prizes for guessing which one a person with reading disabilities is likely to find the most useful.

A variety of readability tests or formulas are available to help web content developers determine the reading level of their documents. These include the online Juicy Studio Readability Test [15].

Guardianship Tribunal

In 2003, the Guardianship Tribunal of NSW sought to develop a new website. A high level of accessibility was a key consideration since the Tribunal serves a diverse audience including legal and health professionals as well as people with physical and cognitive difficulties, their families and carers.

This project provided us with an opportunity to work together once more and to put into practice a few accessibility ideas we had been working on.

During the development process, we designed and tested a variety of tools to enhance the accessibility of the site. The strip of “accessibility tools” at the top of nearly all pages on the Guardianship Tribunal site contains three tools (or utilities) that give the user some control over both what information will be presented on the pages, and the way that information is presented [16]. We will return to the issue of tailoring page content to meet the information needs of different user groups later. Now, we would like to look at how recent advances in the use of cascading style sheets (CSS) can give users greater control over how that information is presented.

Two of the accessibility tools on the Guardianship Tribunal site allow the user to increase or decrease the size of the text and to increase the line spacing between links in the navigation menus. With these however, we were only just beginning to scratch the surface of what might be possible.

CSS Controlled Presentation

The use of CSS to break the nexus between web page content and presentation, which had become entrenched over the years, has greatly enhance the potential of websites to meet the needs of people with cognitive and learning difficulties.

CSS can now be used to give site visitors control over the way page content is presented, including:

• Increasing line height, or the space between the lines of text, which may make text easier for early-stage readers to read.
• Increasing the size of the ‘clickable’ area for links within a paragraph of text, which should help people with spatial difficulties and/or fine motor skill problems.
• Mouse over highlighting of text paragraphs or table rows with changes in colour and/or underlining, which should help people who have difficulty negotiating lines of text. Currently, Internet Explorer does not support the hover pseudo-class on anything other than the “a” element. It is possible to generate the effect with JavaScript.
• Change the background colour of the page. While we don’t want to get into the whole contentious issue of the relationship between page colours and dyslexia, it is clear from observing many people using the web, including some with cognitive disabilities, some prefer the contrast between the text and the background to be high, whereas others find a lower contrast level (often with muted background colours) easier to read.
• Colour inversion. Although most pages present dark text on a light background, as an extension of the previous point, some users prefer to invert these colours since they find light text on a dark background easier to read.

Suggestions and examples are available showing the use of CSS to aid readability and clickability [17] of site elements.

Allowing Users To Control Presentation

As we have seen, CSS allow web developers to provide users of their sites with the ability to change the presentation of written material in a variety of ways. A balance needs to be found however, between providing options and swamping people with choices. Too many options can cause cognitive overload, as anyone who has shopped for breakfast cereals for ten year olds in an unfamiliar supermarket is probably only too well aware.

Providing web users with greater control over the way material on a website is presented is very much in keeping with one of the fundamental tenets of the web. Browsers already give users the ability to do some of these things, for example increase font size (assuming the page is made correctly), set their own style sheet, and some circumstances change or remove background colours.

Many users however are not aware of these functions and some web-savvy users who are aware of the range of browser controls are often reluctant to use them. For example during a recent evaluation session, a highly experience web-user who requires a headwand commented frequently that she found the text size a little too small. When asked if she ever uses the browser to increase font size, she replied, “It’s not that easy to do and most times doesn’t work so I don’t bother anymore“. Last year, the same person evaluated the Guardianship Tribunal site for us and commented very positively on the accessibility tool that allowed her to set the font size directly from the page.

It may be both unnecessary and unwise for developers to include the full range of potential user-controls on all the sites they prepare. User control of line height or spacing may be appropriate for all sites, however the ability to highlight paragraphs of text may only be of the highest value on educational sites or sites that are particularly designed to service the needs of people with learning and/or reading difficulties.

Once the range of options for controlling page presentation have been narrowed down to those that are most likely to be pertinent for potential users of a particular site, there is the question of where to put these controls. There appears to be two fundamental choices:

• Provide user site-controls on every page. There are obvious advantages in having the controls on all pages and in the same position on each page. In particular users will be constantly reminded of their availability and will find the controls easier to locate. In some site designs however, the presentation of accessibility controls in a strip or toolbox may require an unacceptably large proportion of the screen real estate.
• Provide a separate page with the site-controls. The obvious saving of page real estate is one advantage of putting the user accessibility controls on a separate page. Another, perhaps more important advantage is that, a separate page allows for the inclusion of a greater description of the tools and how they can be used. The most serious disadvantage is that user may not find the controls at all, and this could be very significant for users with cognitive disabilities.

The final issue to consider in regard to the control of accessibility tools is the question of how and where should the switching between options occur.

The process of page customisation, even when using CSS, can occur either on the user’s browser (client side) or on the computer that hosts the site (server side).

Client side “style switching” on the user’s browser requires using the Document Object Model (or DOM usually via the use of JavaScript) to reset the active stylesheet to a named alternative stylesheet, which has already been loaded by the browser.

The main advantage of using client side switching is that when a change is selected it appears to happen instantaneously, since the alternative stylesheet has already been loaded. However, there are a number of significant potential problems with client side switching:

• The user may have an old browser with limited support or they may have JavaScript disabled.
• Internet security systems like Norton Antivirus (through it’s pop-up blocking routines) may inadvertently break the intended JavaScript action by adding their own script block to every page viewed.
• The protective firewalls used by organizations sometimes block JavaScript files.
• It can limit the number of potential options that may be provided. When multiple style choices are presented, stylesheets for all the possible combinations that may result from these choices will need to be prepared.

A more reliable option is to do as much as possible on the server before the page is sent to the user’s browser for display. While this means the page will have to be reloaded in order to apply the change, this is not likely to be significant issue when pages are designed well with semantically correct code.

There are many ways to achieve style switching: The settings can be held in persistent or session browser cookies; in a URL query string; or, in a session variable on the server. Using persistent cookies offers one significant advantage to the user, which could be particularly helpful for people with cognitive disabilities. With persistent cookies, once the user has set their presentational preferences these will be remembered and implemented each time they use the same computer to visit the site in the future.

The actual changing of the stylesheet can also be done in a variety of ways, depending on the skills of the developer. For example:

• Changing the ID of the BODY element
• Changing the link to the stylesheet
• Using a dynamically generated stylesheet (e.g. a ColdFusion or PHP file with a mime-type of text/css) in which the values change depending on the selection.

Navigation

Website navigation that is both usable and accessible requires a number of key elements:

• Clear labels and signs so the user can find and understand the options that are available.
• Good feedback so the user can confirm their actions, see whether they made the right choice and recover from any mistakes easily
• Reliable functionality or performance so that it is easy to use and will function with different browsers and devices.

The location of navigation menus on the page is a subject of long-standing and ongoing discussion in the web community. In the early days, sites were relatively small and navigation menus were simple, often presented in a strip across the top of the page. As sites got bigger containing many different sections or sub-sites, there was an advantage in separating global or site wide navigation from the local navigation for each area of the site. In many cases this was done by presenting the global navigation horizontally at the top and the local navigation in a vertical menu down the left or right side of the page content.

As the web developed further, the demands for navigation increased and new ideas were tried. Many were good, but some reduced the ability for people from some sections of the community to find information or to use the site at all.

Many large sites, for example the St George Bank and AMP Insurance, now simplify navigation by providing navigation menus for accessing informational content that are clearly separate from the menus for accessing the various functions or utilities a user of the site might need. This separation of navigational components can greatly reduce the number of choices that need to be considered at anyone time, potentially making the site easier for people with cognitive or learning difficulties to use.

Potentially, because if the navigation systems are not well delineated and do not reflect the likely needs of the user, they may just lead to a greater feeling of confusion for some users. For example, we recently evaluated the accessibility of a site with five different navigation menus on each page. One of the evaluation participants who has short-term memory difficulties as the result of an accident, frequently became lost when undertaking the evaluation tasks and often asked to be reminded of what he was supposed to be finding (or doing) mid way through a task. We have used the same evaluator a number of times in the past, and in nearly all these other cases the sites and their navigational systems provided him with sufficient feedback to retain the focus of a task until it was completed.

There could be a variety of reasons why individuals with cognitive disabilities may find the navigational systems of a site difficult to learn or use. It would seem reasonable to assume, the logical and consistent positioning of navigational elements, including the search facility, on all site pages will assist in the short term memory required to acquire information about the navigation system and to use it. The consistent positioning of navigation elements within a site in a way that that conform to general web conventions is also likely to assist in longer-term procedural memory, which is important in the performance of routine tasks [18].

Well-designed navigational feedback, including the use of breadcrumbs, can be particularly beneficial for people with cognitive problems, since it appears to provide ongoing confirmation of navigational choices and reinforcement of the overall task objective.

The use of an expanding left-hand navigation menu is now very popular. This allows the presentation of global and local navigation choices within the same column, making them easier for user to associate. We used this device with the Guardianship Tribunal site. A task-based evaluation of the completed site was undertaken using nine participants, including three with cognitive disabilities. All the participants found the navigation system easy to use, with one exception. The blind participant, who relies on a screen reader, appeared to become disorientated by the changing status of navigation menu each time he made a top-level navigation choice.

We would like to suggest another way of presenting global and local navigation choices that may be appropriate for some sites, particularly those designed for people with cognitive disabilities and learning difficulties. This could be described as the “side by side” navigation menu system, for in essences it involves presenting the global navigation in a left hand column and then generating a second navigation column next to this for those global navigation items that contain second level choices. The second level choices within this second column are presented adjacent to the global choice they relate to.

A picture is worth a thousand words, as they say, so we have prepared a few pages for an imaginary Rental Advice Board of the fictitious State of Equity.

Selecting “Residential Tenancy” from the global navigation in the left column should reveal the second level navigation for this section of the site in another column when the page loads. The “Rental Bond” choice in this menu leads to another page.

When it comes to the position of navigation menus, clearly there are many options with no obvious right or wrong way of presenting them. However, all Web users, including those with disabilities, are assisted when the site navigation is where they expect it to be and performs how they expect it to perform.

Allowing Users To Control Content

We would now like to return to the issue of tailoring page content to meet the information needs of different user groups.

Sometimes providing users with control over the way a web page is presented will not be sufficient to enhance the access for all potential users to the information it contains.

When we began to focus on how a new Guardianship Tribunal site might more effectively meet the needs of people with cognitive disabilities, we soon realised that the legal nature of some of the content could cause problems for people who find reading and understanding large text documents difficult.

We felt that the level of content detail presented on some pages (and required by the legal and health professional users of the site) could become a barrier for those Tribunal clients with cognitive disabilities who still had the functional capacity to access the web.

To help address this concern the accessibility tool, “Content: Long or Short”, was developed so that users can determine the level of detail provided in the content of each page. The Short option provides an easy to read version of the information. The different versions of the page concerned with Confidentiality are a good example of this tool. The long version is equivalent to 3 printed pages with quite a lot of legal sounding language, while the short version is just half a page with the information mainly in dot points.

When people with cognitive and learning difficulties tested the completed site, they were attracted to the short content option and found it easy to use. In keeping with many other accessibility features, we found that this option also benefited the wider community; for example, social workers and doctors were using short content as a way of quickly locating the information they required.

With the aid of the imaginary Rental Advice Board site, we would like to outline three ways we believe that web content can be made more accessible for people with cognitive and learning difficulties:

• Inverted Pyramid writing. Earlier in this paper we briefly touched on the importance of writing for the web, mentioning the advantage of using a newspaper style of writing. With this pyramid writing example, we push this point a little further by suggesting the first paragraph of the page could serve as a concise summary or abstract of the page content.
• Expanding Bullets. With this Show and hide content example, the user has the option of having all the material presented at the same time, as is the case with most web pages, or choosing to have a list of content bullets, which when selected cause the information relating to that bullet to be presented. The relevant content can be revealed either directly below the bullet list or within the list under the selected bullet.
• Long and Short Information. With this Long and short content example, the user can either choose to have the long (or full) content version or the short version where the content is also presented in plain English.

The Long and Short option could be combined with Pyramid writing, allowing the user to determine from the summary if they want to read more on the page and if so, obtain the information in a level of detail that is most appropriate to their needs.

In addition to significant advances in the use of CSS during the last two years, there has been an increasing awareness of the value of developing purpose-built content management systems, which are tailored to the specific needs of the client organisation.

While it is possible to hand code the different content versions of the pages and the links between them, in many situations this is likely to be a laborious process and prone to error. In fact, sites at the level of sophistication that allows users to choose content, almost always need to have a Content management System (CMS) to process and deliver the different content versions based on the user’s customisation preferences.

In general, off-the-shelf content management systems don’t allow for this degree of tailoring of site content by the user. Nearly always it is preferable to have someone who understands the issues, develop a customised CMS solution that is designed to meet a defined set of client and user needs. It is not unusual for a customised solution such as this, to cost less than an off-the-shelf system, since the client is only paying for the specific functionality they require.

There are very few examples of content management systems that allow users to choose between different versions of web page content. The sites for Guardianship Tribunal and the imaginary Rental Advice Board, referred to earlier in this paper, offer a glimpse of what is possible.

Conclusion

As we have seen, it is now possible to make sites that can give users control over both the content of the information on a page and the way that content is presented. The next, and perhaps more challenging task, is to convince website developers and proprietors that this is worth doing.

For the providers of community services, such as education, legal and consumer advice, the benefits of disseminating information to the widest possible audience, including individuals with cognitive and learning difficulties, are clear. These organizations usually have a keen desire to service the whole community, and within Australia, a legal requirement to make the information they provide on the web accessible to all people. Giving users the ability to control both content and page presentation will help them achieve these aims.

While the providers of commercial services via the Internet also have a responsibility to ensure their services are available to people with disabilities, this requirement is more enthusiastically embraced when it brings obvious commercial benefits.

Extending the potential customer base for goods and services to include consumers with cognitive and learning difficulties will bring some benefits to many businesses. However, we may have to wait until organizations start using these technologies to tailor both web page content and presentation to the tastes and needs of different segments of the general marketplace, before we see them being widely used for accessibility purposes.

Footnotes

1. Web Content Accessibility Guidelines 1.0
2. The Peepo Project site used pictures, letters, words and sounds to help visitors browse the web independently
3. Kolatch, Erica. 2000. “Designing for Users with Cognitive Disabilities“.
4. Seeman, Lisa. 2002. “Inclusion of Cognitive Disabilities in the Web Accessibility Movement“
5. Jiwnani, Kanta. 2001 “Designing for Users with Cognitive Disabilities“
6. Seeman, Lisa. “Designing Web Content for People with Learning Disabilities“
7. Bohman, Paul. 2004. “Cognitive Disabilities Part 1: We Still Know Too Little, and We Do Even Less“
8. Rowland, Cyndi. 2004. “Cognitive Disabilities Part 2: Conceptualizing Design Considerations“
9. W3C Ruby Annotation
10. Seeman, Lisa. 2002. “Inclusion of Cognitive Disabilities in the Web Accessibility Movement“
11. Australian Bureau of Statistics. “Australian Social Trends 1998. Eduction attainment: Literacy Skills”
12. “Everything you ever wanted to know about readability tests but were afraid to ask“
13. CentreLink. “Services for People with Disabilities” (accessed 27 November 2004)
14. Services South Australia. “Finding a job, am I entitled to welfare assistance?” (accessed 27 November 2004)
15. Gez Lemon. “Juicy Studio Readability Test“
16. Guardianship Tribunal of NSW website
17. Cognitive Presentation – eleven suggestions for improving readability and clickability
18. Lai, Gerald et al. “Neural Pathways to Long term memory“