Part 1
A Cognitive–Behavioural Approach to Child Behaviour Problems
Chapter One
Child Behaviour Problems: Theory and Assessment
Cognitive-behavioural approaches to child behaviour problems encompass a number of different models and no one model is sufficient to adequately describe the complexity of all possible difficulties. These models in turn generate a wide variety of strategies and techniques for the cognitive-behaviour therapist. From a practical point of view the focus is necessarily on those aspects of the client's functioning that are accessible to intervention. This is not to deny that genetic and constitutional factors will also be interacting with past learning and environmental influences on behaviour. But genetic and constitutional factors are largely outside the sphere of influence of the therapist. Nevertheless it is important to be aware of their role.
Constitutional Factors
The origins of temperamental differences amongst children are unclear but some evidence indicates a genetic component (Torgersen and Kringlen 1978). However the heritability of a propensity towards conduct disorder is not entirely clear. Children of biological parents with a history of anti-social behaviour have an increased risk for anti-social behaviour even when the child is separated from the biological parents (Robins 1979). Yet, having an adoptive parent who has anti-social behaviour, in addition to a sociopathic biological parent, further increases the risk of a child's anti-social behaviour (Hutchings and Madnick 1975). A number of studies have shown early child temperament to relate to the later development of behavioural problems. The New York longitudinal study (Rutter 1964) showed an association between temperamental differences and later rates of referral for psychiatric help. Graham (1973) found a similar association between temperamental differences and later disorder. More recently Walkind and De Salis (1982) have developed a scale to measure temperament and found that four month temperament, as reflected on this scale, related to the presence of behavioural problems when the children were aged forty-two months, with 'difficult' children developing higher rates of problems. It has never been suggested that a 'difficult' temperament will in itself cause a behavioural disturbance, but that problems may arise when a child with this characteristic is mishandled or confronted by stress. One analogy (Cameron 1977) has been to see the temperament as a 'fault line' and the family or other stress as the 'strain'. Both are needed to produce the behavioural difficulty or earthquake. In the Walkind and De Salis study, temperament was the fault line and maternal psychiatric disorder was the stress. Thus there were patterns of mutual influence between child and caretaker leading to eventual outcome.
A Reciprocal Model
It is simplistic to view children as the passive recipients of the influences of parents. The final outcome of the child's socialization and personality growth cannot necessarily be attributed to the manner in which they were treated by their parents. However, as Bell and Harper (1977) have pointed out, scientists in child development research have followed an unidirectional viewpoint on parent-child interactions for over fifty years. A variety of studies have shown a significant association between different styles of parenting and children's behaviour. For example, Wilson (1980) on different forms of parental supervision, Newson (1982) on different disciplinary styles in relation to delinquency, and Cox (1982) in relation to schoolgirl pregnancies. However, these findings were purely correlational and did not indicate which variables had a causal influence. Probably the most frequent conclusion drawn was that aggressive children are so because of the primitive styles of their parents. Bell and Harper suggest that from the data it is just as likely that aggressive children create primitive parents.
The issue of causality in parent-child interactions has been addressed by the studies of Berkley and Cunningham (1979) and Humphries et al. (1978). These studies examined the effects of tranquillizing drugs on the mother-child interactions of hyperactive children. If the child's behaviour is the result of parental directiveness, then reducing the child's hyperactivity with tranquillizing medication should result in litde change in parental behaviours. But if the parent's directiveness was in fact a response to the child's excessive behaviour, then decreasing the child's excessive behaviour should reduce parental directiveness. In fact when medication reduced the hyperactive behaviour and increased compliance, parents dramatically reduced their level of commands. These studies serve to underscore the notion that the child's behaviour can exert control over parental responses, in addition to the traditional view that parental behaviours influence child responses. Further, the selection of one person in the dyad as having the greater influence in the interaction sequence is not necessarily as arbitrary as it may at first seem.
To redress the imbalance caused by the unidirectional view of parent-child interactions Bell and Harper propose that parents have expectations for a child's behaviour in a given situation. If a child's behaviour is excessive in terms of frequency, duration, intensity, or age-appropriateness for a given situation it is said to exceed that parent's 'upper-limit threshold'. In contrast, when a child's behaviour is deemed by a parent to be inappropriately deficient along the same parameters it is said to exceed the parent's 'lower-limit threshold'. Such expectations can differ across settings, parents, and time. In the case where the child's behaviour exceeds the upper-limit threshold of the parent because it is excessive, high-rate, or aversive, the parent is likely to respond with the 'upper-limit controls', frequendy consisting of ignoring the child's behaviour, usually progressing to restrictive commands, negative affect, and physical disciplining of the child. Where the child's behaviour remains within the parental expectations or limits, parental reactions are likely to consist of positive interactions, questioning, occasional praise, and mild physical affection; these are 'equilibrium controls'. Where the child's behaviour falls below the parents' lower-limit threshold, the parents emit 'lower-limit controls'. Such reactions include drawing the child's attention to activities, coaxing, prompting, and encouraging as well as providing provocative commands and physical guidance. These behaviours are intended to increase the frequency of behaviour that is infrequent in the child and would be seen in parents of retarded children. Bell and Harper also hypothesize that parental reactions are probably hierarchically and sequentially organized such that when initial reactions prove unsuccessful, other reactions next in the hierarchy of that set of control behaviours will be emitted. Should no behaviour within the parental repertoire serve to affect a child's behaviour, it is likely that disengagement from and future avoidance of the child will be the result.
Models in Child Behaviour Therapy
In traditional child therapy the focus has been on changing a child's behaviour via control of the child's environment. In the newer cognitive perspectives the focus is primarily on remedying the cognitive deficits or dysfunctions that have an aetiological role in conduct and emotional disorders. The behavioural parent training programme in Chapter 2 is based on the traditional behavioural models, and the basis for the strategies used may be best appreciated from the more detailed discussion of classical conditioning, operant conditioning, and modelling, which follows. This discussion is itself followed by 2m elaboration of the cognitive models which offer explanations for behaviours such as inadequate impulse control, anger arousal, and childhood emotional disorders. The cognitive models differ as to whether the target is to supply the child with a new way of talking to him/herself, remedy a deficit in thinking, or to teach a child to modify existing maladaptive thought processes. These models underpin the cognitive approach to parent-child behaviour problems described in Chapter 4.
Classical Conditioning, Operant Conditioning, and Modelling
The classical conditioning model has been at its most useful in explaining the learning of simple reflex-like behaviours. Pavlov (1927) noticed that whenever he placed meat powder in a dog's mouth it began to salivate. The meat powder was called an unconditioned stimulus (UCS) because it automatically produced the salivation response, an unconditioned response (UCR). Repeated pairing of the UCS with a neutral stimulus, for example, a ringing bell, resulted in presentation of the bell alone eliciting a response very similar to the UCR. The response produced by the bell alone (after pairing) was called a conditioned response (CR) and the bell was therefore a conditioned stimulus (CS).
Pavlov and his colleagues found that the strength of the learned association was related to the number of repeated pairings of the CS-UCS and of the intensity of the UCS. They also found that presentation of the CS alone eventually results in the CS losing its ability to cue the CR, a process called extinction. Therapeutically, this suggests that if a neutral stimulus has become a noxious conditioned stimulus, the durability of the CS will depend on the frequency and intensity of the distasteful learning experience. Further the CS can lose its power to evoke the CR and this procedure can be accelerated by strategies such as relaxation training to weaken the connection between stimulus and response. The strategy of 'systematic desensitization' is based on classical conditioning. In systematic desensitization the child is first taught deep muscle relaxation, then a hierarchy of the child's fears is constructed. Whilst in a state of relaxation the child is asked to imagine the least threatening item in the hierarchy, when this image is tolerated without distress, the child again uses the relaxation procedures and progress is made to imagine the next least threatening item and so on ascending the hierarchy. An alternative to systematic desensitization is to use another classical conditioning-based technique, 'flooding'. This entails prolonged exposure, either in imagination or in vivo, to the most anxiety arousing stimuli. The extinction model underlying flooding predicts that inappropriate, learned emotional responses can be unlearned by repeatedly presenting the stimuli conditioned to elicit them in the absence of an intrinsically aversive stimulation.
The operant conditioning model developed by Skinner (1953) draws on the earlier experiments of Thorndike (1911). He put a cat in a cage secured by a simple latch. A piece of fish was placed out of reach of the cat outside the cage. To begin with the cat tried to stretch out and reach the fish but to no avail and subsequendy gave up. Whilst moving about the cage, the cat accidentally bumped into the door and opened the latch and made directly for the fish, which it ate. The cat was then placed back in the cage, and a new piece of fish laid outside; eventually it inadvertendy freed itself and ate the fish. After a number of repeated trials the cat gradually took less time to free itself and consume the fish. Thorndike called this trial-and-error learning 'instrumental learning' to emphasize the active role played by the organism in discovering an appropriate response to the stimulus. The key feature emphasized by Skinner in operant conditioning is that behaviours are altered by their consequences. Clearly parents may organize reward of the behaviour they want in a child, but whether the prescribed reward actually constitutes a reward from the child's point of view can only be determined by whether the desired behaviour increases in frequency or not. It may be that a parent is perhaps inadvertendy rewarding inappropriate behaviours, e.g. giving attention to a child's temper tantrum and that behaviour is then set to continue.
Modelling
Children may learn behaviour patterns not only on the basis of classical and operant paradigms but also by observation of other people. For observational learning to occur, it is not necessary that the child actually perform the response nor that the child receive direct reinforcement or punishment. Learning can occur in the absence of response consequences but whether the response that has been learnt will be exhibited will depend on the consequences to the child. Bandura (1969) sees modelling as explaining most of the learning of children, particularly how behaviour is learnt in the first instance. It also goes some way to explain the finding that children who are abused by their parents are more likely to abuse their own children. Deficits in the interpersonal behavioural repertoire of children, such as an inability to approach, greet, or converse with others, may be attributable to following an inappropriate role model. Modelling has been used to help children overcome fears, phobias, and problems of social withdrawal, and to teach parents child management skills.
Cognitive Deficits
The works of Luria (1961), Vygotsky (1962), and, more recendy, Zivin (1979) have served to highlight that children, in the course of normal development, learn to control their behaviour by self-directed speech. Failure to learn this private self-talk may result in impulsive or angry or aggressive behaviours. Meichenbaum (1977) has developed self-instruction training (SIT) as a strategy for instilling self-talk in children as an effective regulator of behaviour. Theoretically, self instructions are effective through assisting the child to develop an internalization of verbal commands. SIT usually involves teaching a child to adopt a five-stage approach to making a problem manageable; each state is modelled by the therapist then rehearsed by the child. The first stage focuses on encouraging the child to ask 'What exacdy is my problem?'; in the second stage the child asks 'What is my plan?', it may be, for example, to list and prioritize homeworks from various teachers before making a start; the third stage involves the child asking whether they have put the plan into action; in the fourth stage the child evaluates what progress has been made; if necessary a coping verbalization can constitute a fifth stage in which the child utilizes a special routine if the plan of action is proving unsuccessful; for example 'I'll go back to the beginning and take it very slowly'. Essentially SIT teaches a child to think before acting. However, to achieve this cognitive goal Meichenbaum uses traditional behavioural strategies such as modelling and therapist reinforcement for each successive approximation the child makes to completing a stage of the SIT routine.
Impulsive children can also be seen as deficient in generating and scanning a range of alternative solutions to problems and further failing to impede inappropriate solutions; as such they are poor problem solvers. Social Problem Solving (SPS) training may be used as an adjunct to behavioural or cognitive interventions, a form of treatment in its own right, or as a way of ensuring maintenance of treatment gains. SPS has been evaluated and shown promise (D'Zurilla and Nezu 1982) with a wide range of client groups and target behaviours. Most of the research on SPS has been based on a model which includes the following five general skills or operations: a) problem orientation, b) problem definition and formulation, c) generation of alternatives, d) decision making, and e) solution implementation and verification.
Cognitive Dysfunctions
Not only can the absence of self-talk cause problems for a child but so too can unrealistic self-talk. The function of Meichenbaum's self-talk is to induce self-talk whereas approaches based on Beck's (1976) or Ellis's (1982) work are focused on modifying maladaptive thought patterns. In work with children Meichenbaum's model has been predominant. Partly, Beck's cognitive therapy for depression, developed with adults in mind, has been little used with children because of doubts that childhood depression could exist. However within the last few years a consensus has emerged that childhood depression is a distinct syndrome with symptoms that parallel those of the adult disorder (Orvaschel et al. 1980). Depending on the cognitive sophistication of the child, with suitable children, cognitive therapy for depression may be an appropriate treatment regime. As in the adult version of CT, treatment would initially focus on increasing a child's sense of achievement and pleasure, moving from a behavioural phase to a cognitive phase in which the child's dysfunctional silent assumptions are first made explicit and more realistic assumptions substituted. For example a depressed child may be operating on the silent assumption that 'If I work hard enough, always be there for my mum, then maybe she will say she loves me; if she doesn't say that it means I am not working hard enough'. This could be an example of the logical errors listed by Beck as often leading to depression and personalization; essentially, the child is thinking 'If something goes wrong it must be my fault'. Teaching the child to note such dysfunctional thought processes and produce rational responses to maladaptive thoughts is the cornerstone of a CT approach. The CB models of Beck (1976) suggest that in the development of fears, anxiety, or phobias, children, like adults, may make five possible types of errors in the processing of information and come to make a number of dysfunctional silent assumptions. Personalization is just one of the types of logical errors. The others are arbitrary inference, selective abstraction, magnification/minimization, and overgeneralization. The logical errors are not necess...