Understanding the Searching Process for Visually Impaired Users of the Web (NoVA)

Background to the NoVA project

It is recognised that in order to participate fully in today’s society it is vital that citizens are not excluded for any reason, whether by virtue of birth, belief, aptitude or circumstance. Exclusion takes many forms and must be countered in many different ways. Funded by Resource and undertaken by the Centre for Research in Library and Information Management (CERLIM) at Manchester Metropolitan University, the NoVA project is concerned with countering the exclusion from access to information which can all too easily occur when individuals do not have so-called ‘normal’ vision. Our domain in this project is digital library services, and our concern is that all such services should, in their entirety, be as accessible to blind and visually impaired people as to anyone else.

The NoVA project builds on work previously undertaken by CERLIM during the REVIEL project[1] in which the REVIEL project team analysed a number of electronic resources such as subject gateways and online catalogues for accessibility. Findings revealed that while most sites were partially accessible and at times adhered to accessibility recommendations such as those produced by the World Wide Web Consortium Web Accessibility Initiative (W3C/WAI)[2], many still included features that could be difficult for some groups of people to access. For example, the use of graphics with no alternative text to support them means that a blind or visually impaired person using access technology cannot ascertain what the graphic is – and whether or not it is important. A good example of this is the use of bullet points, which the user does not necessarily need to have read out (in cases like this the ‘NULL’ alternative text attribute should be applied so that the screen reader will skip over it).

Further experiments during the REVIEL project in which a group of blind people accessed a variety of online resources revealed that navigation and in particular searching for information, is a major problem within digital library systems. The use of frames in a web environment, for example, enables the user to perform complex selections across a number of categories. Research into user seeking behaviour in a web environment reveals that people generally do not read the whole of a web page but often glance over a page, picking out the sections most relevant to their needs or interests. Sighted people are generally able to do this with relative ease, but for a non-sighted person using access technology it is necessary to navigate linearly within one frame at a time, and he or she may need to backtrack a long way (again in a linear manner) in order to reach the desired point (and then maybe track forward again).

Although much work is continuing to make interfaces accessible - witness, for example, the work of the World Wide Web’s Web Accessibility Initiative, there is little current work on how blind and visually impaired people navigate interfaces, and in particular on how a serial or linear method of searching enforced by current access technology can be mapped onto the parallelism displayed by interfaces, for example, through the use of frames. Work on accessibility concentrates on transcribing text (or replacing images etc. with text or audio) and a number of research and development projects such as TIDE ACCESS[3], Emacspeak[4] and BrookesTalk[5] have developed systems to help enable universal access (interfaces using musical tones, text-only, customised interfaces). The problem however may be much deeper when taking information seeking behaviour and the use of interfaces into consideration, as these often assume visual capabilities which blind and visually-impaired people may not possess.

Research and development work on computer accessibility in the past has had little input from human computer interface (HCI) specialists[6] although work undertaken at the University of Manchester Computer Science Department has provided a useful analogy between how sighted and visually impaired people travel in everyday life and their virtual ‘travels’ through cyberspace[7]. By comparing real life travel tasks with virtual ones the study suggests that as sighted people’s real life travel tasks will differ from people with sight problems, then the same is true in a virtual world. Despite this interesting work, communication between HCI experts and disability access communities needs to improve.

Aims and objectives of the project

The overall aim of the NoVA Project is to develop understanding of serial searching in non-serial digital library environments, with particular reference to retrieval of information by blind and visually-impaired people.

The objectives of the Project are:

• to develop an experimental framework for exploration of serial or linear searching and retrieval in non-serial environments
• to undertake a series of experiments with serial searching and retrieval, and subsequent use of digital content
• to map serial/non-serial approaches so as to develop understanding of how serial searching, retrieval etc. can be optimised in non-serial environments
• to report on findings and to make recommendations for digital library system design.

Work in progress

The research team has undertaken an analysis of documentary evidence in areas of accessibility, human computer interface design and information seeking behaviour relating to both the sighted and non-sighted community. To date the evidence collected has revealed that although there are a number of studies conducted on the information needs of blind and visually impaired people[8], [9], [10], [11], there is very little literature available on the information seeking behaviour of blind and visually impaired people. Therefore it is hoped that the results from the NoVA experiments will help to address this gap.

Development of the experimental framework is currently in progress. The project team aims to use this framework to test the information seeking behaviour of sighted, blind and partially sighted people using a number of digital library resources (for example OPACs, search engines, online services such as BUBL) as well as a selection of commercial sites (for example Marks and Spencer). The framework will include a set of structured and unstructured tasks. In both cases the users will be given a choice in how to undertake the task. For example they might choose the Search option or they may prefer to browse the site first for appropriate links. Data will be logged using Lotus ScreenCam software[12] which simultaneously records key strokes and verbal dialog so that the search terms used and search path chosen will be recorded together with a verbal description. The verbal description will be used to verify what is recorded on screen and provide a useful insight into why a search path was chosen. More intensive logging of task processing by visually impaired users will be undertaken. This will include follow up questions relating to the task and specific features used during the process.

Future work

Future work of the project will include testing of the framework. A sample of sighted and visually impaired users will be asked to perform a number of tasks to test it. Results will be used to compare differences (and similarities) between the two groups in the searching process in order to map out serial and non-serial approaches to searching. It is hoped that this will help develop an increased understanding of how serial searching is optimised in non-serial environments and be used to inform human computer interface designers in order to improve access to web-based search facilities. It is anticipated that throughout the duration of the NoVA project, CERLIM staff will continue their liaison with experts in the field of access to information by visually impaired people including representatives of the Royal National Institute for the Blind (RNIB), the National Library for the Blind (NLB) and Disability and Information Systems in Higher Education (DISinHE).

References.

Author Details