Improving the communication rate for symbol based, scanning voice output device users

Children with physical disabilities who are unable to speak have reduced opportunities in education, their social world, and eventually in the workplace. They are treated as deaf, as having low intelligence, or are ignored completely [17]. Symbol systems and voice output communication devices have been used to help these children communicate [4]. The symbols represent words or phrases and allow the child to indicate whole words or phrases with minimal movement. If the symbol is represented on a voice output communication device, the device will speak the word as the symbol is selected, but communication rate is low. Children with severe physical disabilities who are unable to point directly, use a switch and a process called scanning to select the symbols, and these children are the slowest of the augmented communicators [5,11]. Prediction is commonly used as a technique to enhance the rate of message selection. However the research literature gives little support for the use of prediction as an effective rate enhancement strategy. Horstmann-Koester and Levine [13,14] and others have demonstrated limited improvement with prediction amongst spelling users (Horstmann-Koester and Levine [13,14], Horstmann-Koester [12], Higginbotham [9], Anson [1], Mathy-Laikko and West [16] and Venkatagiri [23]). They find that the key stroke saving advantage of prediction is negated by the time loss involved in searching and choosing from the list of predicted words. New software was developed in response to the need for an effective rate enhancement strategy for scanning symbol users. It is unique in many ways because it represents an attempt to design a strategy which incorporates many of the findings from research into scanning and word prediction. A close analysis of the literature suggests a number of features which may be useful in a prediction product.