Plain behaviour modification:
rewards and withdrawal of rewards



In the learning theories of Skinner and Hull, the most fundamental operation is reward. In applying these theories to human psychology, both Skinner (1953) and Wolpe 1978) suppose that we can interpret many subjective experiences and cognitive habits, as well as obviously goal-seeking actions, in terms of histories of internal and private reward and punishment. In some cases this seems obviously right. Watching the World Snooker Championships is strangely satisfying, and even the sight of coloured balls dropping neatly into their pockets seems to elicit a special internal frisson. We remain glued to the set only because we like it (in spite of Skinner’s free-floating disapproval of all spectator sports), and all the other things we do can in some sense be said to be rewarded—up to a point.

There are numerous objections to the all-embracing claims of these theories. We can at least entertain the fantasy of a peak moment of pleasure which is so breath-taking and overwhelming in intensity that it produces instant amnesia for itself and its build-up, and therefore has little effect on our future habits.


Although this seems a far cry from our day-to-day reinforcements of watching snooker and getting paid, the correlation between the emotional intensity of goal outcomes and subsequent knowledge is not always strong. Boring routines of practice may sometimes be more beneficial for both academic and musical skills than occasional celebrated successes, and also punishments can be better aids to memory than pleasures. And 1 can remember many things about academic meetings and conferences of such unrelieved tedium as to destroy all faith in the connection between subsequent retention of knowledge and emotional evaluation of any kind whatever.

Despite these and other quibbles, there is still a case that, in practical terms even if in no others, the manipulation of rewards remains a very powerful tool for the purpose of modifying overt behaviour. This carries most weight, again in practical terms, for the behaviour of those least likely to have alternative internal resources for the direction of their own behaviour, such as an objection to all rewards as a matter of political principle. In other words the manipulation of rewards has most effect on the behaviour of laboratory animals and children, and adults with the minds of children. Some regard all deliberate giving and withholding of tangible rewards as unethical, although civilized life as we know it would soon grind to a halt if everyone acted in this high-minded way. There are certainly many sensitive issues in the question of how to deal with persons who are severely handicapped in one way or another, especially if they are confined to institutions, and legal rulings as to the rights of patients have limited some behaviour-modification programmes, though more often in the United States than elsewhere (Stolz et al., 1975). But the spectre of the behaviour modifier as invariably herding the mentally retarded into lavatories with a cattle-prodder and restricting access to films, fun and freedom, with the same goals as managers of the Gulag Archipelago, is unjustified as well as unjust. The horrendous problems of the individuals who may be treated by behaviour-modification methods are not of the making of the behaviour modifiers, many of whom are extremely dedicated and sensitive individuals, trying to surmount otherwise unsurmountable deficiencies in those in their care (Kiernan and Woodford, 1975, pp. 258, 295).


Behaviour modification with the severely retarded

Consider the problems to be faced in a school such as that discussed by Barton (1975).This is a school in a medium-sized subnormality hospital in a large northern city, attended by children for whom places cannot be found in other schools for the educationally severely subnormal (ESN(s) schools). In the 5 classes for the 70 children the mean intelligence quotients (IQs) ranged from 5 to 31 (where the average is 100), and 25 per cent of the children had mental ages measured in months for chronological ages measured in years. Some 38 per cent could not feed themselves, 61 per cent could not dress themselves, 52 per cent were not toilet-trained, 57 per cent had no speech and 36 per cent did not walk. The age range was 3—17 years. Physical difficulties such as seizures and sensory and motor impairment meant that any progress in such circumstances was likely to be slow, but behaviour- modification programmes designed for individual children suggested that it was not unobtainable. The general point is that if such children can be trained in basic self-help skills such as the following then their chances for the future, in terms of going to other schools and developing further social and manual skills, are immeasurably improved.

Toilet-training: six children of an average age of 13 1/2 and with an average IQ of 10, who made up the next to the highest class, but were all incontinent, were given a toilet-training ‘package’ of the kind proposed by Azrin and Foxx (1971). This involved (1) taking them to a toilet every half-hour and giving them extra fluids, sweets, biscuits, praise and lots of attention when they successfully used it; but also (2) strapping them in a chair for half an hour, away from the class, if they had an accident (a ‘time-out’ procedure, used as a punishment). The results of this were slightly erratic, since two children were absent due to illness, but in the other four, accidents per week declined fairly dramatically.

Dressing. this was taught by a ‘prompt and fade’ technique. That is, the children were first both instructed and helped to dress themselves, but the amount of help given was gradually reduced. Social praise and congratulation were augmented with spoonfuls of sweet chocolate pudding as reward for the younger and more retarded participants. Large, loose elasticated garments were used to start with. This all seems perfectly sensible, and the


method was strikingly successful even with children with an IQ of 8. Teachers thereafter saved time by not having to give as much help as before with coats, knickers and shoes, and the children were able to demonstrate the skills they had learned at school when getting up in the morning.

Movement and awareness of the environment: it is occasionally possible to elicit some response in individuals whose behaviour appears to be only vegetative. Fuller (1949) conditioned arm raising by injecting liquid food into the mouth of an 18-year-old who had barely ever before made any controlled movements. Barton (1975) reports that an 8-year-old girl with a developmental age of only 3 months seemed to learn very rapidly to make voiced sounds for the rewards given by 10-second activation of a large vibratory pad placed under her.

Language and social communication: the elicitation of basic movement and awareness is perhaps only worthwhile if it can be the starting point for much further development. At a different level of achievement, there have been many attempts to use the methods of behaviour modification to improve thc skills needed for appropriate speech and social interaction. For, if this succeeds, it clearly provides the individuals concerned with access to other sources of information and help. It is sometimes said that language is given innately, and is acquired by all children independently of motivation and intelligence (Lenneberg, 1967; Chomsky, 1965, p. 58). For applied purposes, this is quite without foundation. There are very many particular kinds of language deficit and speech difficulty but also, in general, impairments of speech and comprehension necessarily go hand in hand with mental retardation. Barton (1975) quotes a single case of systematic improvement in the correct naming of pictures of 12 items that were originally misnamed by a profoundly retarded 15-year-old, when sweets were given for correct responses during tests.

Somewhat more subtle linguistic skills, as well as articulatory skills and object naming, are also susceptible to reward training. Stevens-Long and Rasmussen (1974) used food rewards and praise to persuade an autistic 8-year-old to use plurals correctly when he was describing pictures, and also to use longer and longer descriptions, in the form of compound sentences. Lutzker and Sherman (1974) used similar methods to promote subject—verb agreements for singulars and plurals when both retarded children


and normal toddlers who had not yet started doing this described pictures. An important aspect of these forms of training is that (if necessary by trainingwith large numbers of examples) the trainees generalize, or learn functional units, or learn rules so that, after being rewarded for saying ‘boats are sailing’ instead of ‘boats is sailing’, they can apply the rule to a new description, such as ‘girls are riding’, without having to be separately rewarded for each new instance of the rule (see pp.127-8). Similarly, in linguistic analyses of remedial language training, it is taken for granted that ‘practice and motivation’ will serve the purpose of strengthening new linguistic structures (Crystal, 1979).

For normal children, waving goodbye is often a very early and popular form of social exchange, although greeting strangers by saying hello may sometimes be inhibited by shyness. In the retarded even these basics may require contrived and protracted training. Stokes et al (1974), working with withdrawn boys in a home for the retarded, began gently pulling a boy’s arm back and forth in the form of a wave when they met him, before giving crisps or sweets. Alternatively, sweets were waved about, followed by the boy’s hand being waved, before they were delivered, as a ‘visual prompt’. Following this, more realistic greeting responses, sustained only by the rewards of social interactions with reasonably enthusiastic training personnel, were observed.

Wanted and unwanted behaviours in the less severely retarded and in special eases

The basic methods of behaviour modification are the giving of both tangible and social rewards systematically for small steps towards all kinds of skills, and the equally systematic withdrawal of reward for antisocial or disruptive behaviours. The behaviour modifier may withdraw immediate social responsiveness to these misdemeanours by parents and teachers; or he may remove the learner from most possible sources of reward (often by isolating a child from companions and toys in a ‘time-out’ area) as a consequence of these unwanted acts. The methods have been found useful in a number of .instances other than with very severely retarded children. Williams (1959) and Wolf et aL (1964) are often quoted for examples of very extended and intractable tantrums in young children of 2 and 3 years old (in these cases associated with


hospitalization for other reasons), which disappeared when the tantrums resulted in the children being left in their rooms until the tantrums subsided. Both Wahler (1969) and Callias and Carr (1975) instructed parents of older children with more generally disruptive behaviours in similar deterrent techniques, with apparently satisfactory results. In the Callias and Carr report this was part of a more general programme in which clinical psychologists instructed the parents and teachers of retarded children in behaviour- modification techniques for accelerating the acquisition of various skills (the usual self-help necessities, and also helping in the house, communication, and miscellaneous matters such as training in wearing hearing aids) and for ‘decelerating’ many other things, including, hitting, hair- pulling, biting, throwing, smearing and tearing clothes, tantrums, screaming and head-banging. Both Wahler and Callias and Carr gave special training to parents with instructions via earphones and videotaped methods of feedback, where this was necessary.

The application of behaviour-modification methods in these instances may be judged as severe and unappealing, but they appear to be effective in dealing with problems to which no alternative solutions are available. There is similarly a case to be made for the application of controlled rewards in the symptomatic treatment of anorexia nervosa. In this syndrome patients (often but not always girls or young women) refuse to eat and may because of this die, if untreated. When hospitalized patients are allowed access to visitors and such privileges as listening to the radio and making telephone calls only if they gain a specified amount of weight (e.g. a pound in five days), weight gains ensue (e.g. Halmi et al., 1975).

The use of tokens

A broom used in an experiment in 1965 has become, to some behaviour modifiers, an object of considerable veneration (e.g. Craighead et al., 1981). Haughton and Ayllon (1965), working in a ward for chronic female schizophrenics, gave cigarettes to one of the patients only when she was holding the broom. As a result, the broom became a valued possession, and the patient stood holding it for long periods of time. The experimenters then asked two psychiatrists to make independent diagnoses of the broom-


holding behaviour. One said that this behaviour was a stereotyped response of the kind commonly found in regressed schizophrenics, with the broom acting as a favourite toy, as might occur with a young child. The other, more imaginatively, proposed that the broom signified something more meaningful to the patient, and that to her it represented either her own child, or a phallus, or the sceptre of an omnipotent queen, or perhaps some combination of these. Behaviour modifiers believe that to the patient the broom represented cigarettes, and thus to them the broom represents both the fallibility of psychiatrists and the infinite malleability of all behaviour when rewards are dispensed by behaviour modifiers.

In the belief that extended control over rewards would lead to extended control over behaviour, tokens such as coins and chits were established in place of individual cigarettes as rewards, to be used by schizophrenic patients in the hospital as money, exchangeable at appropriate times for consumables or privileges. Examples are: personal chair, 1 token per day; choice of bedspread, 1 token per day; 20-minute walk in the grounds, 2 tokens; attending an off-ward religious service, 10 tokens; having a private audience with the ward psychologist, 20 tokens; having a private audience with a social worker, 100 tokens; choosing a television programme, 3 tokens; toilet articles such as toothpaste, comb and lipstick, 1—10 tokens; clothes and accessories such as slippers, skirt and handbag, 12—400 tokens. In order to obtain tokens patients had to earn them, either by useful self-care activities, or by on-the-ward jobs. Examples are: make own bed and clean area, 1 token; brush own teeth once per day, 1 token; act as waitress (for 10 minutes), 2 tokens; washing-up (for 10 minutes) 6 tokens; write names of other patients brushing teeth (for 30 minutes), 3 tokens (Ayllon and Azrin, 1968).

For obvious reasons such a system is referred to as a ‘token economy’. There is no doubt that when it is well run, changes in the behaviour of patients, even chronic schizophrenics, may be observed, interpretable in terms of economic pay-offs. These include the hoarding, lending at interest, and stealing of tokens — the first two discouraged in at least one instance by a policy of token inflation (at 25 per cent per month: Atthowe and Krasner, 1968). There is equally no doubt that under even the best-run systems the patients remain schizophrenic, and although occasional transitions to normal working, via half-way houses, have


been reported, there is no evidence that token economies produce ‘cures’. In the United States, the setting- up of token economies has been discouraged by a number of legal rulings which forbid such things as limitations of patients’ rights to attend religious services or to have their own chairs. Thus, many of the items and privileges used as reinforcers in the original schemes have been declared to be the legal rights of all patients. Also, it is now required that the legal minimum wages (in dollars) be paid for many of the duties which they originally performed (e.g. Wyatt v. Stickney: Stolz et al., 1975; Kazdin, 1977). In the United Kingdom, the setting-up of such systems has proved not often to be feasible within the National Health Service (Thorpe, 1975).

In view of the ethical and legal problems raised by token economies, their high administrative cost, and the very limited therapeutic gains demonstrated, they cannot be said to have proved their usefulness. However, it is claimed that similar schemes have had beneficial effects without contravening legal rulings, not only in mental hospitals but also in special schools and residential institutions for various kinds of delinquent (Kazdin, 1981; Craighead et al, 1981). At their mildest, such schemes have the benefits and limitations of various point, mark or gold-star award systems which have a long history in educational contexts (Kazdin, 1977).


End of Chaper 11 | Contents | Start of Chaper 12