This paper assumes that the reader has a good understanding of the switch scanning process.

Practitioners working in the field of AT constantly strive for the best clinical outcomes for their clients and to do so within the context of an evidence base. The real evidence base for applied AT is small, due to the limited research in the field, the specific and varied nature of client’s conditions and limited theoretical constructs to allow for modelling of situations. The field of HCI is relatively well established and the author suggests that a large number of the lessons learnt within the discipline of HCI can be transferred to AT.

Within HCI a number of applied models have been validated for the ‘normal’ population, these describe interactions between this population and computers. The author proposes that the nature of some of these interactions is the same for some populations of AT users and that the models should be re-validated for these populations. The author also suggests that many of the more general lessons from HCI can be directly applied to the AT field.

Although the use of models may be considered to be removed from the day to day clinical application of AT, the author suggests that validated models should form the basis for the majority of the evidence base into AT.

As an example of this hypothesis, a range of AT situations will be described and the potential application of HCI illustrated. Particular focus will be given to the application of HCI models to the topic of switch access to AT and scanning systems. The application of the ‘Model Human Processor’ to switch scanning will be extensively covered. In order to provide context to this work the results of a literature search on the topic of switch scanning will be presented.

Some application of these models within the field of AT:

1. Keystroke-Level Models for Word Prediction task, Koester & Levine (1997); Augmentative & Alternative Communication
2. Use of Model Human Processor for people with motor impairments, Keates, Clarkson and Robinson (1998)

What is quicker? Using 3/4/5 switches as a directional joystick or a Joystick Mouse (e.g. P&G Joystick)? Some devices are removing ability to use more than 2 switches. The percieved replacement is using a mouse device. Is this better or worse?

AutoScan VS Inverse Scan: Anecdotally, most people use 1 switch AutoScan – why? Inverse Scan is being dropped by AT products as an option. Is the cognitive load of AutoScan more or less that UserScan?

Next: HCI Models

Full evidence base: My CiteULike Library for switching

— Simon Judge 2006/02/23

Images © Michelle Turton