Information Studies


Pete Williams: Doctoral research: Optimising web page design for people with learning disabilities

Pete Williams: Doctoral research

Comparing the effectiveness of different website designs in facilitating information retrieval for people with learning disabilities

Image ld

Introduction and aims

Much literature emanating from academia (e.g. Tarleton, 2004; Cameron and Murphy, 2002), organisations from the learning disabilities field (Proud, 2008; Lawton, 2006) and government (DH, 2001; 2005), including legislation (e.g. HMG, 2010) emphasises the need for social inclusion, informed choice and the active involvement in society of people with Learning Disabilities (LD). These aspirations can only be achieved by the provision of accessible and relevant information – now increasingly available in electronic form. However, even material written especially for this constituency may not be readily accessible because of navigation and retrieval problems. However, there is little empirical evidence regarding what measures actually aid website use, and some of it is conflicting. Nielson, (1996) and Bohman (2010) for example, recommend avoidance of the need to scroll. By contrast, other commentators (e.g. Sevilla et al 2007, Keates et al, 2007) urge the use of images, video etc. Pages containing such content, however, tend to be longer and therefore do require scrolling. Thus, the aim of this project was to determine which web page design factors facilitate successful usage of the Internet by people with LD.


Two stages were involved. The first drew upon previous work by the researcher (Williams and Minnion, 2007; Williams and Nicholas, 2006; Williams, 2006, 2008). It consisted of qualitative usability studies of a particular web site (detailed below) , and examined  page attributes that affected effective negotiation of the medium. It focused on different aspects of websites (‘menu’ lists, layout and navigation) and on different activities (interaction, information retrieval and expressing preferences). This work led to the development of tasks suitable for undertaking by people with low literacy skills (Williams and Hanson-Baldauf, 2010; Williams, in-press). The page attributes elicited were examined quantitatively in stage two by comparing web pages of different layouts. For this stage, task-times were used as a measure of difficulty and statistical analyses undertaken..

The initial web site

For stage one of the project, a website was examined that was developed by researchers at the University of East London’s Rix Centre, working with Newham Borough Council. It consisted of a number of pages on the theme of transition – the change from education to supported employment and general adult life. These were made by people with LD themselves, in sessions conducted at their colleges or day centres guided by Rix Centre staff.

Methods were also explored to elicit preference judgments, taking into account communication and articulation problems characteristic of this cohort. A three category (‘smiley-face’) rating scale was devised (and later modified) and short interviews undertaken for this purpose. In preference interviews, the tendency to agree with an authority figure, ‘acquiescence bias’, was minimised by developing tactics to avoid polar (‘yes/no’) questions, such as by asking for interface comparisons (‘Which one do you like best…?’ rather than ‘Do you like this?’)


The study was undertaken at various locations, including the special needs unit of a college of further education, an adult education class in a community centre, a self-advocacy group and adult day centres. Participants had either ‘Entry’ or ‘Level One’ in literacy, as used in UK Further Education colleges (Moser, 1999). As such, they could notionally:

  • Read and understand simple text (those at the top, Entry Three level, up to six sentences in one paragraph)
  • Get the main idea from a simple graphical or tabular source, (e.g. safety signs)
  • Understand and use a simple list.

Participants had no physical disabilities. Crucially, none were visually impaired. In total 22 people participated in stage one and 104 in stage two of the study, ranging from 17 to 63 years in age. Of the latter, 94 undertook sufficient number of tasks to be included in a quantitative analysis.


Stage One

The following attributes were shown in Stage One of the study to affect usability: menu position, text size and images, as follows.

Images: Many participants showed difficulties in interpreting the meaning of images, and the more literate participants relied on the written caption to ascertain menu topics depicted. Indeed, some of the images used could be interpreted in a number of ways even by ‘mainstream’ users. For example, that originally chosen to represent ‘Support’ was a close-up of a handshake, which could have meant friendship, some kind of agreement, or even a ‘goodbye’. It was thus decided to investigate the issue of images by examining whether even well-chosen ones aid information retrieval in terms of time spent on accessing content.

Another issue was that of the space taken up by images, which often make pages extend below the level of the screen and require scrolling. This was a big issue for some participants – the least able appearing to not realise that ‘invisible’ content existed and some not being able to scroll effectively.

Text size: Whilst study participants tended to prefer larger font, and some guidelines emphasise the importance of making it possible for people with LD to increase font size (e.g. BSI, 2006), as with images, the space taken up by even a small body of large-font text pushes the page length down below the visible screen. There is also, of course, the potential problem of readability. In print form larger text sizes have been shown to be more readable than smaller sizes (Mills and Weldon, 1987; Rudnicky and Kolers, 1984). No significant research had been undertaken on text size in a digital environment.

Menu position: Clearly, access to content needs to be as simple and accessible as possible and so the presentation of menu entries is of great importance. The preliminary qualitative work undertaken suggested that, for LD users, the horizontal option might be the optimal choice, because:

  • The entire menu is visible on-screen (discounting horizontal scrolling, heavily frowned upon by accessibility experts) (Colorado and Eberle 2009);
  • The direction in which the information presented is compatible with the eye movement of someone reading, so it may be easier for someone with LD to read menus from left to right rather than vertically.

In addition to these findings, also noteworthy was the difference in behaviour exhibited when undertaking set tasks and free browsing respectively. For the former, participants tended to stay on the required page and spend a fair amount of time absorbing its contents. However, when free browsing (i.e. having been asked to ‘have a look’ at the site), most participants tended to activate a large number of pages rapidly and apparently without imbibing much, if any, information.

Stage Two

Following Stage One of the project, a new website was created, and included various design modifications suggested by usability testing, below, shows the ‘home’ page. The grid menu design was present only for this ‘portal’ page - vertical and horizontal arrangements were tested for the information pages themselves.

As mentioned, this phase of the study sought to examine the effect of different layout designs. By using JavaScript this was possible to undertake without resorting to producing multiple copies of pages, each exhibiting a particular layout. Java coding enabled toggling between text size (‘large’ or ‘small’), menu position (vertical and horizontal) and images (visible or not) as required. This gave eight interface designs, as shown below:

Interface no.    Contents pos.  Images?  Text size

Interface 1       Horizontal       No       Small

Interface 2       Horizontal       No       Large

Interface 3       Horizontal       Yes      Small

Interface 4       Horizontal       Yes      Large

Interface 5       Vertical           No       Small

Interface 6       Vertical           No       Large

Interface 7       Vertical           Yes      Small

Interface 8       Vertical           Yes      Large


By testing information retrieval performance on each of these interfaces, in terms of task-time,  it was possible to address the question of which attributes of websites are the most important in designing for accessibility – taking into account the effect each variable has on the overall site appearance and functionality. Results suggested that menu position was the most significant factor, followed by text size. Surprisingly, small-text was imbibed more quickly than large (for this non-visually impaired cohort). Images did not appear to aid understanding or facilitate quicker access to information.

A major finding was that to seek specific information, participants exhibited only ‘serial access’ to content. That is, information was imbibed sequentially until the required comment was reached (unlike ‘linear’ access, which includes skimming and ignoring inconsequential words or phrases). This contrasted with the ‘random access’ behaviour adopted when browsing sites, where rapid page changing suggested attention only to images and other non-text content.

Serial access limitations may explain why images did not aid information retrieval speed (despite prior selection by the cohort to reflect appropriate content representations). They were ignored until reached ‘serially’– and therefore did not help signpost text. The fact that small-font was more effective than large may have been because the latter required more lines and eye movements to negotiate. As mentioned, the main task-time predictor was menu position. Here, the more effective arrangement was horizontal, because entries were always visible, and the lateral arrangement suited ‘serial access’ behaviour. By contrast, the body text of the page laterally juxtaposed with vertical menus proved distracting. Lateral serial access behaviour may have made it difficult for participants to ignore the menu entries in the side column.

Preferred site designs were not those that yielded fastest retrieval times (except for menu position).  Participants greatly preferred sites that were liberally populated with images, and had a large text size – both attributes which tended to produce the slowest retrieval times. This discrepancy may be due to preferences being judged on aesthetic rather than performance considerations and on the basis of random rather than serial access practices.

Recommendations and conclusions

Recommendations are possible that reconcile both the problems of serial access behaviour and the discrepancy between preferences and performance. In short, a major recommendation would be to undertake a ‘trade-off’ between text content, size and position. ‘Serial access’ and longer time taken to negotiate large text suggests that as little text content as possible should be used - substituting full sentences for key words and phrases. This would facilitate the use of large-text, which might in any case suit people with visual impairments better. The finding that images are apparently almost ineffective suggests the importance of ensuring adjacency with text and that much careful consideration of how to pictorially represent concepts needs to be applied .

It is fitting to conclude with a final word about the aspirations of self-advocacy and inclusion which underpinned the research. Many of the findings  outlined in this paper contradict current guidelines, so hopefully they will inform a debate currently lacking empirical evidence around how best information to facilitate these aspirations may be presented.


Most importantly, a huge thanks to the staff and students at all the fieldwork locations for their participation and interest in the project. The research could not have been done without their total co-operation and help. Many thanks also to Mr. Andy Minnion, Director of the Rix Centre, for his helpful comments and advice, and to other colleagues at the Rix Centre and at UCL.

The Economic and Social Research Council generously funded the development and testing of ‘Pete’s Easy Read’ site. The original Newham web site was developed at The Rix Centre by equally generous funding by the Social Care Institute of Excellence, BP, The Jack Petchey Foundation, and Newham Borough Council.

References and bibliography

Adam T, Tatnall A (2008) Using ICT to improve the education of students with learning disabilities in Kendall M, Samways B (Eds.) Learning to Live in the Knowledge Society Boston: Springer 218 pp.63-70

Bernard, M. and Hamblin, C. (2003) Cascading versus Indexed Menu Design. Usability News, 5.1.

Bernard, M., Lida, B., Riley, S., Hackler, T. And Janzen, K. (2002), A comparison of popular online fonts: which size and type is best? Usability News, 4(1).

Bohman U (2010) The need for Easy-to-Read information on websites in Miesenberger Klaus J, Zagler W, Karthmer A (Eds.) Computers Helping People with Special Needs, Part I:12th International Conference :ICCHP 2010, Vienna, Austria Berlin: Springer-Verlag pp634-640

BSI (British Standards Institution) (2006) PAS 78:2006. Guide to good practice in commissioning accessible websites. London : BSI

Burzagli L, Emiliani PL, Gabbanini F (2009) Design for All in action: An example of analysis and implementation Expert Systems with Applications 36 985–994

Cameron, L, Murphy, J (2002) Enabling young people with a learning disability to make choices at a time of transition British Journal of Learning Disabilities, 30 (3), p105-112

Chen SY, Ford, N. (1998) Modelling User Navigation Behaviours in a Hypermedia-Based Learning System: An Individual Differences Approach. Knowledge Organization.  25(3) pp.67-78.

Choi J, Bakken S (2010) Web-based education for low-literate parents in Neonatal Intensive Care Unit: Development of a website and heuristic evaluation and usability testing, International Journal of Medical Informatics 79(8) pp.565-575

Colorado JT, Eberle JH (2009) Web accessibility essentials for online course developers In Song H, Handbook of Research on Human Performance and Instructional Technology pp334-356

DaCosta B, Seok, S (2010) Managing cognitive load in the design of assistive technology for those with learning disabilities in Seok S, Meyen E, DaCosta B (2010) Handbook of research on human cognition and assistive technology: design, accessibility and trans-disciplinary perspectives New York: Hershey pp20-39

DeStefano, D., & LeFevre, J. (2007). Cognitive load in hypertext reading: A review. Computers in Human Behavior 23 pp.1616-1641.

DH (Department of Health) (2001) Valuing people a new strategy for learning disability for the 21st century: A white paper London: Department of Health

DH (Department of Health) (2005) Independence, Well-being and Choice London: Department of Health

DH/CNO (DH Partnerships for Children, Families and Maternity/ CNO Directorate) (2008) Transition: Moving on Well London: Department of Health.

Florian, L. (2004), “Uses of technology that support pupils with special educational needs”, in Florian, L. and Hegarty, J. (Eds.) ICT and Special Educational Needs: a Tool for Inclusion, Open University Press, Berkshire, pp. 7-20.

HMG (Her Majesty’s Government) (2005). Disability Discrimination Act 2005. London: HMSO

HMG (Her Majesty’s Government) (2010). Equality Act 2010 London: HMSO

Karreman J, van der Geest V, Buursink E (2007) Accessible Website Content Guidelines for Users with Intellectual Disabilities Journal of Applied Research in Intellectual Disabilities, 20(6) pp510-518

Keates, S., Adams, R., Bodine, C., Czaja, S., Gordon, W., Gregor, P., Hacker, E., et al. (2007). Cognitive and learning difficulties and how they affect access to IT systems. Universal Access in the Information Society, 5(4), 329-339.

Lawton A (2006) A Voice of Their Own: A toolbox of ideas and information for non-instructed advocacy Kidderminster: BILD Publications

LDAA (Learning Disabilities Association of America) (2010) Symptoms of Learning Disabilities Learning Disabilities Association of America website.

Lerner J, Johns B (2009) Learning Disabilities and Related Mild Disabilities: Characteristics, Teaching Strategies, and New Directions Singapore: Cengage Learning 11th edition

Meisenberg, G., Williams, A. (2008). Are acquiescent and extreme response styles related to low intelligence and education? Personality and Individual Differences, 44, 1539-1550.

Mills, C. B. Weldon, L. J. (1987). Reading test from computer screens. ACM Computing Surveys, 19 (4) pp.329-358.

Minnion A, Staples P, Singh R, Williams P (2008) Beyond the Road Ahead London: University of East London Report submitted to the Social Care Institute of Excellence

Moser C (Chair) (1999) A Fresh Start - improving literacy and numeracy London: HMSO Available online at: http://www.lifelonglearning.co.uk/mosergroup/ (accessed 10.08.10)

Nicholas D, Huntington P, Watkinson, A (2003) Digital journals, big deals and online searching behaviour: a pilot study. Aslib Proceedings 55 (1/2) 84-109

Nicholas, D, Huntington P, Williams, P and Dobrowolski, T (2004) Re-appraising information seeking behaviour in a digital environment: bouncers, checkers, returnees and the like. Journal of Documentation 60 (1), 24-43

Nielsen, J. (1996). Top Ten Mistakes in Web Design

Pottker H (2003) News and its communicative quality: the inverted pyramid—when and why did it appear? Journalism Studies 4(4) pp.501-511

Proud D (2008) Acting on Inclusion Inclusion Now 20 

Rudnicky, A. I., & Kolers, P. A. (1984). Size and case of type as stimuli in reading, Journal of Experimental Psychology in Human Perception and Performance, 10 pp.231-249.

Sevilla, J., Herrera, G., Martinez, B., & Alcantud, F. (2007). Web accessibility for individuals with cognitive deficits: A comparative study between an existing commercial web and its cognitively accessible equivalent. ACM Transactions on Computer-Human Interaction, 14(3) pp.1-23

Tarleton B (2004) The Road Ahead? Information for young people with learning difficulties, their families and supporters at transition Bristol: Norah Fry Research Centre, University of Bristol

Tinker, M. (1963). Legibility of Print. Ames, IA: Iowa State University Press.

Tullis T (2005) Web-based Presentation of Information: The Top Ten Mistakes and Why They Are Mistakes in HCI International 2005 Conference Proceedings Las Vegas, Nevada, USA, July 22-27, 2005.

Tullis, T. S., and Cianchette, C. (2003) An Investigation of Web Site Navigation Techniques. Usable Bits, 2nd Quarter 2003. 

W3C (World Wide Web Consortium) (2008) Web Content Accessibility Guidelines (WCAG) 2.0 Available online at: http://www.w3.org/TR/WCAG/#contents (accessed 03.09.12)

Williams P (2008) Transition and people with learning disabilities: reflections on the quality of content that emerges from the process of involving service-users in information provision Aslib Proceedings 60(5) pp474-492

Williams P (2011) Barriers to the Creation and Use of an Accessible Web Portal for People with Learning Disabilities International Journal of Education 3(2) e21 Available at: http://www.macrothink.org/journal/index.php/ije/article/view/1283

Williams P (2012) Web site usability testing involving people with learning disabilities using only images and audio to access information Journal of Research in Special Educational Needs [in press. ‘Early view’ at: doi: 10.1111/j.1471-3802.2012.01235.x]

Williams P, Hanson-Baldauf D (2010) Testing a Web information portal for people with learning disabilities Journal of Research in Special Educational Needs 10(1) pp42-51