This is a submission for the RDWG Symposium on Text Customization for Readability. It has not been reviewed or accepted for publication. For more information about RDWG symposia and publications, see the RDWG Frequently Asked Questions (FAQ).
This paper explores some of the similarities between guidelines for making text easy to read for different audiences, including people who have difficulty reading because of visual disabilities, dyslexia or other reading difficulties, or who do not read well.
It is based on work to improve the usability of large information-rich web sites, including The Open University and the US National Cancer Institute’s cancer.gov.
Web sites with both large amounts of content and widely diverse audiences have both advantages and challenges in addressing the individual needs of their visitors. They are often organized in content management systems, which allows consistently accessible templates to be used, but the text is usually written by a large number of authors, over long periods of time. As a result, the quality of both the text itself and its presentation vary widely.
One solution has been micro-sites – sections of the site aimed at specific audiences. This strategy has several usability and management challenges:
On the other hand, web designers may fear that creating a single site that works for a wide audience will be bland and be a mediocre experience for all.
In our work at The Open University, we had three primary student audiences: young adults, older adults, and people speaking English as a second language. We were concerned that their needs were so different that it would not be possible to create one site that worked for all of them. The issues we identified included difference in writing style as well as how the text is presented.
We decided to start with by reviewing guidelines for designing for these audiences, especially those developed through empirical research. After a literature search, we decided on three sources for guidelines:
After using these sources to create our own list of guidelines, we created the content for a new site and conducted usability testing of a prototype. The participants in this test had widely varying ages, educational level, socio-economic status, and familiarity with higher education.
It was difficult to find guidelines that were both based on user research and which addressed similar aspects of web design. Guidelines were also written at different levels of granularity: some provided conceptual guidance, while others were very specific and detailed.
The first step in this project was to review the guidelines and come up with our own, consolidated set. At first, the guidelines for the three audiences seemed very different, focusing on the specific characteristics. For example, we learned that each group has its own characteristic reading style.
But, when we dug a little deeper, we noticed that these very different user behaviors often led the three sets of researchers to similar guidelines. For example, in this case, they all suggested keeping paragraphs and sentences short, avoiding long, dense blocks of text.
Even though the reasons are different, the guideline is the same: short paragraphs and short clear sentences to break up “walls of words” into meaningful chunks of text.
Many of the guidelines looked at how the text was written, but some concerned the default presentation of the text. Even there, we found the same guideline for different reasons. For example, all of the guidelines urged a “larger” default font size (as small as 10pt for teens up to 12-14pt for low literacy audiences).
Some of the guidelines contradicted popular beliefs. For all of the audiences, the sources suggested informative headings were helpful when readers scanned a page quickly, images worked best when they were meaningful rather than simply decoration, and content was more effective when it was written directly to the user.
When we tested our consolidated guidelines with a diverse audience of potential Open University students, we saw many of the behaviors predicted in our literature review. We also observed that content written and presented to address these behaviors was successful in helping participants find, understand, and use the information.
One of our concerns, especially for a university web site, was that participants would find information written in this style too simplistic or bland. They did not. Instead, more proficient readers appreciated the ability to quickly and confidently scan the page, while those with lower English reading skills were able to more accurately find answers to key questions about the university.
This work suggests that making text easier to read makes a site more usable for everyone, rather than only helping those with a specific disability.
This approach may be applicable to thinking about text customization. If there is a set of customizations that are useful for many different needs, they could be approached as part of the general preferences set. Perhaps they can join the long history of features like curb cuts: originally intended to remove a disability barrier, but evolving to be used by all. Future research might include:
This paper draws on work done at The Open University with Caroline Jarrett and Ian Roddis, and on the work of the Design to Read project, led by Caroline Jarrett.