From a technical perspective, this is related to electronic survey accessibility, which in turn could easily be seen as a real world instance of accessible web form design plus accessible navigation; and therefore covered by a subset of WCAG 2.0. However, it’s not as straightforward as that.

I’d initially thought e-assessment was a tool mainly used in colleges and universities – my bias being partly due to working in that sector, and also my impression that the tertiary education community seemed to be responsible for much of the research and development into e-assessment, such as that funded in the UK by the Joint Information Systems Committee (JISC). But I’m now aware that there’s a much, much wider scope of use – in schools, by professional standards bodies, organisations assessing employee capabilities, lifelong learning.

Despite this diversity of use, there are common constraints which can affect how accessibility of e-assessment is approached.

Balancing accessibility and fair competency assessment

E-assessment is about objectively measuring whether someone has sufficient knowledge/skill to meet a certain level of attainment. All candidates should have an equal chance of being objectively assessed, and accessibility barriers should not obstruct a disabled candidate from being able to demonstrate their competency. This can lead to problems when trying to figure out for example how to provide appropriate text alternatives for a graphic that forms part of a question, or choosing between drag-and-drop or radio buttons as a question type.

The e-assessment author should know what knowledge and skill is being assessed, so that should be at the forefront when thinking about accessibility. Are someone’s powers of visual interpretation of a photo, diagram or video being assessed? If so, is it reasonable to exclude someone who can’t see by not describing the graphic or providing audio description for a video? Is manual dexterity a critical part of the skill being assessed? If so, is it reasonable is it to exclude someone who has a tremor, or is unable to use their hands by utilising a drag-and-drop style answer selection? These are questions that have to be answered by the assessment author before effective accessibility solutions can be applied.

The problem occurs when the method of assessment requires a capability that isn’t necessary for the skill being assessed. So, in the above example, example keyboard-inaccessible drag and drop questions are rarely justifiable.

Assessment Environment

The assessment environment also presents accessibility challenges. In theory, e-assessment presents many accessibility wins, by supporting flexibility of delivery – alternative formats become easier to generate based on personal preferences – and easing maintenance tasks. In practice, flexibility can be difficult to achieve in a controlled environment. There is the question of the extent to which the interface used to present the e-assessment to candidates is accessible. Can a disabled candidate navigate through the questions, understand each question, and select and input an answer without undue difficulty? This should be assessed, and any issues acted upon, in advance.

But, also, will a candidate who needs a particular assistive technology or accessibility solution be able to use their own computer and AT? Or will they have to become familiar (potentially at short notice) with another AT that has been provided to them at the last minute, may be unfamiliar and might not be exactly suitable for their needs? The latter situation means a disadvantage for the disabled person being assessed. How do you provide the assessment environment – which might  be a special locked down browser – with sufficient accessibility support? Can you justify refusing to allow someone to bring and use their own computer and AT on the grounds of fairness to others taking the assessment?

The inaccessibility of the assessment enviroment was the central focus of the Latif vs PMI court case, where a ruling found in favour of a blind person claiming discrimination by a professional association.

Assessment creation

We’ve been doing some work looking at the extent to which e-assessment software supports accessible assessment authoring using W3C ATAG as a reference. This work highlighted some of the potential issues that an author might unwittingly introduce, through insufficient or obscure accessibility prompting by the authoring tool. These issue can be managed short term by author training and support (general accessible design and the specific issues surrounding the authoring tool they use) and a suitable quality assurance process before assessments are presented to candidates. Longer term, of course, we need improved authoring tools.

But, at a recent event on e-assessment and accessibility held by Becta, I became more aware of the complex chain of organisations involved in the supply and delivery of an electronic assessment. At a university, a lecturer is likely to be responsible for creating and delivering their own assessments; but elsewhere there are organisations which are responsible for managing and validating the assessment process. They, or a third party, might author assessments which are then provided to schools and other organisations to administer to students.

So if the original authors are not aware of accessibility issues, there is a long chain down which a request for accessibility information – or adjustments – must pass; and there is no guarantee the request will get to the end of the chain. More informed procurement processes may help to ensure that organisations ask for, and receive, e-accessibility in an appropriate way. But the short term requires people administering e-assessments to make doubly sure in advance that they are aware of any potential accessibility barriers present, and takes steps to manage their impact.

The extreme circumstances under which e-assessment takes place means that developments like better accessibility profiling of users and assessments will hopefully have a positive impact, but a contextual and pragmatic approach to accessibility is essential. Thankfully there are people on the case helping to raise awareness, including Becta, and also Techdis, with their guidelines for accessible assessments.

As awareness amongst web designers grows of the need to create accessible web content, there’s a natural frustration that despite the designer’s best efforts, people can still experience accessibility problems. In the WAI model of accessibility, this might be defined as a user agent problem. A lack of conformance by browsers with UAAG means it still seems to be extraordinarily difficult to make basic,  effective and persistent changes to web page display (for example changing text size, colour schemes) or interaction style (efficient keyboard shortcuts). Following UAAG, especially Guideline 11 and Guideline 12,  would – the theory suggests – make it easy for a user to make the necessary adjustments to meet their accessibility requirements.

The short term, but unsatisfactory, workaround to this problem, has been for web sites to provide accessibility aids as part of their content – features like text resizing widgets and alternative style sheets. Embedding accessibility features in web content has obvious and significant shortcomings – they can only every apply to one web site; so a user can’t rely on them being available across all the sites they want or need to visit.

This ‘give a man a fish or teach a man to fish?’ conundrum has been a focus of attention for many accessibility and standards advocates. In terms of raising awareness, efforts have been made to make it easier to apply accessibility changes through browser widgets (e.g. Firefox text size extension), scripts for accessibility (e.g. AccessMonkey), or user-definable access keys. There are web sites such as the BBC’s MyWeb MyWay, and video tutorials on how to make basic accessibility changes; there’s an ongoing debate on whether a site’s accessibility page is the best place to provide helpful accessibility information.

However, the paradox with all these efforts is that the people who are most likely to be affected are those who are also highly unlikely to find these solutions independently. Without assistance, these valuable tools may remain undiscovered by the people who need them most.

What’s important to realise is that this issue concerns not those with significant impairments that might require, for example, a full blown screen reader, and where an accessibility problem – and solution – is pretty easy to define.

Instead, the people who are most likely to be affected by an inability to make accessibility changes are people with relatively minor, but multiple impairments, which may fluctuate short-term but gradually decline long-term. This includes many older web users, experiencing age related capability decline – and what makes the problem difficult to solve is that the gradual nature of change, allied to an unfamiliarity with the nature of the Web, these users might be unaware they have accessibility requirements.

We’re used to assuming that in accessible design, the user at least knows they have accessibility needs, but what happens when they don’t?

There’s been a lot of work focusing on defining and storing user needs, and using them making adaptations to resources based on these accessibility needs, for example the Supple project. User profiling for accessibility is now standardised, in an e-learning context, in ISO 24751. The WAI-AGE project  is investigating the overlap and differences between ageing and web accessibility; and browser manufacturers like Opera do now seem to appreciate how difficult ‘simple’ accessibility tasks can be for many users.

This is all important work; but beyond technical problems we need to look at how people develop skills in IT use, in how they learn web browsing techniques – and what role, if any, accessibility plays. And we need to understand better how we can discover that a person’s visual, cognitive or mobility capabilities have declined to such a point that applying an accessibility solution (or changing an existing one) would help to sustain independent access.

Tracking a user’s capability changes, keeping this information as accurate and up to date as possible, and using it to make helpful changes – all in an ethical way – is something we plan to tackle in the SUS-IT project over the next three years. I’m really looking forward to working with a team drawn from sociology, psychology, information science as well as computing science to tackle this problem in a more holistic way.

And maybe at the end of it all, we can have a situation where, to use one of my favourite analogies, awareness of the text resizing functions of a web browser is as widespread as that of another accessibility aid – the TV volume control.