Use of IT with learning-disabled populations: problems and challenges
Department of Psychology, Royal Holloway
University of London, Egham Hill, Egham
Surrey TW20 0EX United Kingdom
Tel: +44 1784 443519
Fax: +44 1784 434347
IT and learning disabilities
For some years I have been interested in the great potential of IT for improving the cognitive capabilities of those with learning disabilities. Using a computer stimulates motivation and increases self-esteem; it is a prestigious activity and gives a rare experience of control of the environment. Computers do not get bored or impatient or judgmental and do not require written or spoken answers. In theory they enable learning to be pleasurable, the learning situation to be adapted to the ability of the individual and the learner to control speed of presentation, access to feedback and so on. The teacher should also benefit from detailed records of performance.
In practice IT tends to fall short of these ideals on several counts. The deficiencies are more apparent when it is employed with those with learning difficulties, but they are by no means absent in the material available for use with children of normal ability. Furthermore, since children with learning difficulties are frequently educated in a normal setting, deficiencies in IT provision which do not impede more able children will adversely affect the former group and discriminate against them. They will also create difficulties for children with less acute problems, such as dyslexia, attention deficits and motor coordination problems.
IT and the teacher
I will consider briefly four issues. One relates to the question of whether IT is seen as a substitute for the teacher (normally after preliminary instruction) or as a supplementary aid to be administered by the teacher. In the latter case the problems discussed below may be reduced, but this type of use requires intensive back-up from the teacher, thus compounding the very problems which the use of IT should mitigate. The ideal is undoubtedly, even with learning-disabled populations, to develop software which can operate independently, at least after initial live demonstration, presenting clear demands (preferably without the need to read extensive instructions) and enabling responses to be made and recorded and feedback to be given without intervention from a human instructor.
An illustration of the need to consider carefully the method of interaction with the computer comes from one of the areas of research in which I have been involved. This developed from use of memory games (Wilding and Pauli, 1994), and more recently has employed analogues of simple everyday tasks (dressing a human figure and laying a place at the table) with participants with quite severe levels of handicap. The aim is to improve ability to deal with the real-life versions of these tasks and also to instil the basic understanding of interaction with a computer which is necessary to benefit from more ambitious training programs. The program, therefore, required participants to select responses for themselves, rather than informing an instructor verbally of their choice. However, responses via the keyboard were clearly impractical due to the complexity of the display and size of the keys. A computer mouse requires considerable manual dexterity, beyond the capacity of many participants. A concept keyboard seemed more promising, since it enables large areas and clear displays to be assigned to each response, but it was soon apparent that the conceptual link between action on the concept keyboard and reaction on the computer screen was hard for many participants to grasp without long training. A touch screen provided a more direct link between response and result, but this generated problems in defining a correct response in the software. Even quite young children can soon learn to make precise, demarcated pointing responses, but this proved impossible for many learning-disabled participants, due to both conceptual and motor problems. Fingers would home in laboriously on the target, then linger on it, triggering a cacophony of bleeps which had been programmed as error signals and totally confusing the learner, emphasising that error feedback is only informative when the learner can classify their own response in the way assumed by the creator of the feedback. Removal of the bleep signals did not solve the feedback problem. An adequate response and feedback system is essential to the development of programs designed to make the learner independent of guidance from a human teacher, and while the problems may be less acute with normal children, they are still present. Currently programs for the conventional classroom will normally permit only one response mode, which is just one illustration of their limitations. Even the development of better facilities for presenting and accepting speech will not provide easy solutions, certainly with learning-disabled participants.
Evaluation of software
Turning now to problems relating to the software itself, in my somewhat limited experience, educational software shows little concern to ensure that psychological theory and good psychological and educational principles are observed in its design (ensuring focus on the key features in the display, presenting appropriate feedback, ensuring that the task requirements are clear and clearly presented). Graphical impact and programming ingenuity, rather than the above factors, tend to be the dominant concerns. It is very uncommon for any data to be provided on the acceptability and efficacy of the program for the target population, still less for any independent evaluation of either acceptability or efficacy to be available. Some independent reviewing system is very badly needed.
Adapting IT to individual differences
Another problem is the almost universal failure to take account of individual differences in speed of information processing, attention span, visual abilities etc. Teachers are, in principle at least, adaptable to individual differences in children and adjust their instructional approach accordingly, but few computer programs incorporate the potential for adjusting parameters to match such differences, either automatically or through teacher intervention. Such rigidity is particularly serious in the case of learning-disabled participants, since they will often suffer from specific impairments such as poor vision, poor motor control, unstable attention and slow reactions. As a result, though IT is often presented as particularly beneficial for children who have difficulties with conventional classroom learning, it is more likely to be useful for bright children who find no difficulties with the standard forms. The facility to modify an otherwise satisfactory piece of software to take account of such factors would be invaluable.
A third deficiency in the average commercially available program is poor record keeping. To gain an adequate picture of the learner’s performance, more than a simple count of correct responses and general congratulations to the learner is necessary. Information on the nature of errors made, speed of responding etc helps to provide important diagnostic information which will normally be taken into account by a teacher interacting with the child. It may be, of course, that the program is no more than a teaching aid which requires the presence of the teacher, who can therefore acquire this type of information, but a proper record is nevertheless useful. In any case, as already emphasized, the goal should be the development of programs which enable the learner to proceed independently of the teacher and in these cases adequate data collection and summarising is critical.
An example: the Method for the Induction of Operational thought
One area of research (Pollicina, 1999), in which I have been involved, exemplifies some of these principles. It employs a computerised version of Paour’s (1980) Transformation Box Method with learning-disabled participants. This requires the key difference between two sequences of events to be deduced and encourages the development of general classification skills by presenting a series of such tasks, with the key difference varying from task to task. The computer program is complex, involving as it does a series of tasks, each consisting of several sub-sequences, and the recording of multiple indices of performance. The participant is instructed what to do in a one-to-one relation with the teacher, who also codes and records the orally given responses on the computer. The program is not a stand-alone substitute for a teacher and could not in its present form serve as such, even if the target population could be trained to an appropriate level of competence. However, as an evaluation and training device it has proved highly successful with a learning-disabled population of Mental Age 5 years and above, who have adequate linguistic ability.
The program was developed in Italy and is known as the COT Method (Method for the Induction of Concrete Operational Thought – Pollicina et al, 1997). It consists of 12 Learning Units, each involving a different classification (e.g. distinguishing a change from whole to broken from no change). Each unit proceeds through 5 phases: Exploration, Inductive Learning, Direct Anticipation, Inverse Anticipation and Generalisation.
The experimenter begins by choosing either Assessment Mode, which proceeds without any instructional intervention, or Remediation Mode, in which guidance, verbal descriptions and reinforcement are given, but never explicit answers. At the beginning of the Exploration phase, three objects are shown in broken and intact form (when the classification is broken v. whole) and the participant is asked to describe the elements of the display. Then the display in Fig. 1 is shown to begin the Inductive Learning phase. The participant selects an object and places it in the middle box on either the left or the right; a broken version then appears immediately below if the left side was selected and an intact version if the right side was selected. The sequence can be repeated as often as required with the same or with other objects, until the participant offers a description or rule to describe the distinction exemplified. Direct Anticipation requires prediction of what will appear, Inverse Anticipation requires deduction from an end result as to what preceded it and Generalisation requires prediction of what will happen to a new object.
Figure 1. Display for the Inductive Learning phase
Data are recorded from all phases, such as naming accuracy and time taken, trials before formulating a rule, the type of rule given, and percentage of correct anticipations. Comparison of performance before and after Remediation enables learning to be measured.
A number of improvements should be included in any future version of the COT to improve its conformity to the principles which have been outlined above and increase its potential, especially with populations of lesser ability than those tested to date. For example the salience of changes could be improved by including dynamic rather than static images, facilities for varying speed of presentation should be available in order to determine optimal speeds for different populations, and use of larger screens and facilities for direct responding would be beneficial.
Though this program falls short of the ideal of providing an activity on which the participants can work independently, it is based on a coherent psychological theory (Piagetian analysis of cognitive development), has clear goals and task structure, enables as much repetition as the participant requires and records a variety of performance indices. It is to be hoped that such criteria will become more prominent in the development of IT for all levels of ability, thus enabling the enormous potential of the medium to be more fully realised for all children and adults.