I’ve already written about my systems-led approach to accessibility research. This series of articles is designed to put a little more flesh on the bone with a detailed look at how OOA/OOD (Object Oriented Analysis and Design) can be used to describe computer systems and their interaction with users.
1.1 Method Selection
A little background.
Based upon experience in the Accessibility Research Unit at Teesside University, and looking at the text of papers listed in the bibliography page of this blog, the general approach to research within the accessibility discipline typically appears to begin with consideration of a specific physical or cognitive impairment, and a social context.
This consideration involves either investigating the effect of the chosen impairment upon a given context and vice versa, or investigating adjustments and alternatives to the context in order to better serve the disabled person. One example of the latter approach is Ramloll and Brewster’s (2002) work on augmenting tactile diagrams with spatial non-speech sounds.
The problem that very quickly arose for me when taking this approach, was in relation to the selection of impairments to consider. Having continued from a Master’s level analysis of the problems faced by mobile phone users with Multiple Sclerosis (a disease of the central nervous system) the intention was to consider a range of potential improvements to the interaction modalities within the user interface of mobile phones. It was also to consider those potential adjustments against a broader range of impairment, with a view towards the concept of universal accessibility. This meant that two different questions were to be asked of each interaction modality:
Q2: Who else may be supported by that solution?
The method one would expect to follow for question (Q1), is the established general approach of:
- Construct a user model of the person and the impairment(s) under investigation for the chosen operating context.
- Validate the fidelity of the model against the selected test users.
- Identify representative activities for the test users that would be impacted by their impairments in the given context.
- Identify quantitative and/or qualitative performance metrics related to the activities.
- Evaluate user performance for each activity against the given metrics.
- Perform a statistical analysis on the results.
- Reflect upon the implications of the results in the broader context of accessibility.
Question (Q2) is more problematic given that the breadth of the user model must entail all of the impairments under consideration, and that a larger model needs to be validated for fidelity. Depending upon the breadth chosen, such validation has practical limitations in terms of access to sufficient real users for the validation. Given variations in the severity of both physical and cognitive impairment, formal validation of such models is entirely impractical.
Even if such problems could be addressed, the appropriateness of the representative tasks to be undertaken across such a diverse range of users, and the associated metrics, would also need to be addressed.
The attempt to address the issues surrounding question (Q2), and identification of appropriate methods to answer it, became a focal point within my research. In essence, the research process followed the maxim: if the problem is too difficult to solve, then try solving other, closely related problems within the same ontology, and see if it helps change your perspective on the initial problem.
And that led me towards trying to describe the user and their impairment, the social context of use, and the information being communicated as unique and independent system models, and separately considering the interaction of the models.