Behavioural Interventions In Autism: holding the middle ground

In this post I resume my review of early intervention in autism. Previously I have offered an overview of interventions, based on the therapies which I used with Dylan, and categorised these into four groups: sensory; dietary and medical; behavioural; and educational. This post considers behaviourist approaches to early intervention.

The middle ground

I was chatting to the father of an autistic child recently about the behavioural programme he is using with his son. When I told him that I had used behaviourist and non-behaviourist approaches when Dylan was a child he expressed surprise: There isn’t usually any middle ground, he observed. People seem to be either passionately in favour or passionately against.

 He is right: behavioural interventions more than any other incite strong feelings among parents and professionals. When Dylan was diagnosed in 1996 there was growing interest, in England, in therapies based on behaviourist philosophy. Some families embraced these programmes while others chose to have nothing to do with them. Twenty years later the behaviourist pioneers have established schools and services which are still used with passion and commitment by some families. The lobby against such programmes, however, continues to be equally vociferous. How could I lay claim to both lands? To explain the view from the middle I first need to consider the ground it lies between.

Behaviourism and constructivism

Behaviourism is the label given to theories of learning which are based on changes which can be observed. The Russian Psychologist Pavlov demonstrated, in the 1920s, that it was possible to train a dog to respond to a given stimulus (in this case to salivate when a tuning fork was sounded) a process known as classical conditioning. Pavlov’s Dog became the foundation for experiments to determine whether children’s behaviour could be shaped in a similar way.

In the 1930s Edward Thorndike and B.F. Skinner developed the concept of operant conditioning to describe intentional rather than involuntary behaviour, i.e. learning that involves a decision rather than a physiological response such as salivation. Operant conditioning is based on the idea that consequences determine whether or not a person will repeat a behaviour; a system of reinforcers and punishers are used to shape a child’s actions away from inappropriate and towards target behaviour. Reinforcers strengthen behaviour and can be either positive (doing something which leads to a reward) or negative (avoiding something unpleasant). Punishers, meanwhile, are used to weaken behaviour and involve a penalty or the removal of privileges.

Other theorists, meanwhile, emphasised the social rather than behavioural dimensions of learning. Also in the 1930s, Lev Vygotsky described the way in which knowledge is built collectively among learners, a process referred to as social constructivism. According to constructivist theory, learning is personal and active; our knowledge builds on what we already know and is influenced by the communities in which we learn. Learning, Vygotsky believed, is collaborative and dialogic; our interactions are experiences from which we emerge changed. This notion of knowledge as shared and co-constructed emphasises the role of groups in educational processes.

As a great deal of school-based learning is social there is considerable potential for autistic children to encounter barriers in the education system. Even before an autistic child starts school, however, the consequences of not engaging in social learning are evident; Dylan did not learn to dress, feed himself, use the toilet or play with toys through his interactions with others. Because behaviourist strategies are based on clear consequences in response to observable behaviour, rather than on the emotional protocols involved in social learning, they are felt to be a potentially more effective approach to educating autistic children.

 Parent and professional (I)

I have found it useful when discussing the behaviourist-constructivist spectrum with students to draw a distinction between education and training. Training, I suggest, is something which a person needs in order to carry out a procedure; we train people to answer the telephone in a call centre, use a piece of equipment or file papers. Education, by contrast, requires people to think critically about skills; instead of simply filing the paper in the correct drawer, we might re-write, re-categorise, disseminate or destroy the paper. Education is thus about promoting thinking skills and enabling us to act creatively in response to a particular set of circumstances.

There is debate about whether preparation for certain roles requires education or training. In my own field of education, for example, some people believe students can be trained to teach as you might learn a craft, while others believe teachers must be educated. You can tell a lot about someone who works in the sector by their language. If they refer to ‘teacher education’ then they are probably committed to promoting critical reflection on practice and to encouraging students to question their assumptions about schooling and society. The discourse of teacher training, by contrast, focuses on achievement against a set of teacher competences. I have always considered myself an educator rather than a trainer and drawn on constructivist rather than behaviourist approaches in my own teaching. My professional practice was therefore to have a significant impact on my response to intervention programmes when Dylan was diagnosed.

 The professional: non-directive therapy

Christmas 2013 028My professional position led me, initially, to reject behaviourist approaches; I opted, instead, for a therapeutic response as far removed as you could get from behaviourism. Non-directive play therapy emphasised the importance of following your child’s lead; it required me to enter Dylan’s world in order to understand his experience and win his trust. Mirroring Dylan’s behaviour involved me spending time alongside him; I would sprawl on the floor copying his actions or run in circles around the park with him, hands flapping.

Non-directive therapy, it is suggested, supports the relationship between parent and child by enabling positive interaction. Some of the pre-requisites for communication, such as turn-taking and imitation, can be promoted through non-directive approaches to play. For example if Dylan threw a toy car across the room, rather than reprimand him for throwing I would throw a toy too. The intention would be to get Dylan’s attention: hopefully I would be able to re-shape the behaviour so that we were throwing soft toys instead of cars, or throwing more gently or into a different space; the aim would be to encourage throwing in turn in order to create a communicative rhythm between us. The key to the approach is to take the child’s lead and not to direct; this is quite unlike a behaviourist approach where the aim is to promote a pre-determined target behaviour.

I have reflected a little on my experience of non-directive therapy in a previous post. One of my key observations is that it made little difference to Dylan’s behaviour but had an enormous impact on me; non-directive approaches enabled me to walk in Dylan’s shoes and thus to develop a better understanding of his world. I think that this helped to strengthen our relationship. I don’ t believe, however, that there was any measurable change in Dylan’s behaviours while I was working within this non-directive framework. On all key measures – toilet training, feeding, dressing, communication, play, accessing public spaces – life was a difficult and challenging as ever.

One day I had cause to reflect. Families with autistic children in my city had been given free tickets to the cinema by a local charity and I took Dylan along. Immersed in my non-directive therapy I followed Dylan out to the front of the auditorium when he ran off. People were still taking their seats. I wasn’t confident that I could get Dylan to sit down. I sprawled next to him by the big screen, squirming around on the floor. Perhaps he’d calm down I thought. Maybe this way I could persuade him onto a seat. Suddenly I heard a voice and looking up saw a face I vaguely recognised. Hello Elizabeth. Fancy seeing you here. I see you’ve got your hands full. A word of advice if I may: it’s all about discipline. If I were you, I’d get on with it. I couldn’t quite place her – had to ask her name. It turned out to be somebody I’d gone to school with; she had three children now, the youngest autistic. He was doing well, she told me, in mainstream school but that was all down to the strict discipline and direction the family had built around him.

 Parent and professional (II)

 What if she was right? Could my rigid adherence to an approach I felt comfortable with be unhelpful to Dylan? While it was natural that my work as an educator would influence the way I chose to work with Dylan, I was no longer just an educator; I was the parent of an autistic child. Could I accept that my preferred approach to learning may not be appropriate for Dylan?

Having children of your own can present some interesting challenges; almost inevitably your practice develops to take account of parental as well as professional experience. If your child is disabled, however, the impact on practice can be significant. For me, the interplay of parental and professional knowledge could involve more than a few amendments to my lecture notes: repositioning myself on the behaviourist-constructivist spectrum would strike at the heart of my professional values. I decided, however, that I had to at least keep an open mind.

The parent: behaviour modification

So in 1997 I attended a conference in London organised by PEACH (Parents for the Education of Autistic Children) a group championing the Lovaas approach. The programme involved working intensively with children for 30-40 hours a week on a 1:1 basis as soon as possible after diagnosis. Based on behaviourist philosophy (specifically Applied Behaviour Analysis) the Lovaas method used reinforcers as part of a programme which built step by step to support the development of communication and re-shape behaviour.

I was uncomfortable at the conference. I didn’t like watching the children who were paraded as ‘proof’ that the approach worked. The atmosphere felt more like a rally than a conference; there were celebrations and exhortations, appeals and promises. But there were a couple of presentations which I found interesting, particularly one by Nina Lovaas which focused on adaptations to the programme which might be required for more classically autistic children who are non-verbal. Although it was still early days for Dylan my gut instinct was that he had a similar profile to the children Nina Lovaas described. On the train home I decided I would trial an adapted behaviourist programme based on what I had heard.

Dolly mixtures and jelly tots

dolly took a career break, telling myself that if I didn’t I might regret it one day; maybe I could make a difference. We weren’t in a position to hire the team of aides who would usually be employed to help deliver the programme so I did it by myself. As I’d been used to working full time this wasn’t too much of a shock; I simply exchanged paid work for work with Dylan.

My overall impression of that time is of exhaustion (in another post I’ve shared a poem, Blackbird, in which I attempt to capture some of these emotions). One clear memory I have is of an incident one morning. I was focusing on the development of non-verbal communication through imitation (touch nose, clap, hands in the air, wave, stamp feet, blow). It might not sound challenging but for a three year old child who struggled to remain seated at a chair let alone observe and copy actions it was hard work. The reinforcers (a pocketful of dolly mixtures and jelly tots) were critical. Dylan loved sweets. Once he realised there could be a steady supply by copying me he vaguely obliged. I say ‘vaguely’ because the actions were sloppy; if Dylan lifted his hand somewhere in the region of his face for ‘touch nose’ he would get his dolly mixture. Clapping was the suggestion of hands coming together. Any attempt to observe and copy me was rewarded with a reinforcer from my pocket.

tots this particular morning when I fetched Dylan into the room and gestured to the chair he looked at me then eyed my pockets. In quick sequence he lifted his arms above his head, touched his nose and clapped his hands. I stared at him in shock. He did it again, this time more urgently. Dylan wasn’t sure which action I wanted but keen for a sweet, like Pavlov’s Dog, he performed. I could not continue with this, I thought to myself; it was training and I did not want a performing seal.

The middle way

Later I would soften this position. Although I abandoned Lovaas-inspired intervention I built some of its features into a home education programme I designed for Dylan. I also came to accept that there are areas where training is useful and sometimes necessary; toileting, eating with cutlery, dressing and accessing certain public spaces for example. I therefore drew eclectically on behaviourist and non-directive strategies depending on target and context.

dolly the educational settings, however, I resisted behaviourist approaches and while Dylan was at school I insisted on a ‘no food reward’ policy. Because Dylan doesn’t respond to social motivators, food is the only viable reward; while I am prepared to use it sparingly I am uncomfortable with the routine use of sweets. I was more comfortable using behaviourist approaches with my daughter as I could use social reinforcers (‘if you want a sleepover you’ll have to tidy your room’) in a way that I couldn’t with Dylan. The most effective use of behaviourist strategies, I think, employs such motivators; in this sense social learning is actually at the heart of many behaviourist programmes. While I have used behaviourist approaches only lightly with Dylan I would not hesitate to use them more extensively with a child for whom social reinforcers were acceptable.

This raises the issue I keep coming back to that an appropriate intervention for one autistic child will not be the right choice for another. As well as autistic children presenting differently from each other, some autistic children may present differently over time, responding to alternative reinforcers as they get older. At 20, however, Dylan is no more likely to accept a social reinforcer than he was at three; he is not motivated by anything except food and appears not to understand reinforcers based on the avoidance of something negative. As far as punishers are concerned, Dylan doesn’t understand the psychology of the withdrawal of a reward or the concept of penalty; he appears not to be concerned by the impact of his actions on other people nor does he seek the good opinion of others. The extent to which I can orchestrate consequences in order to shape Dylan’s behaviour is clearly limited by this profile.

The baby and the bath water

jelly an earlier post I suggested that choice of therapy depends on the child, parent, discourse and opportunity. While just one of these factors may be relevant for some interventions, behavioural approaches require that all of these be considered. Firstly, the extent to which a child responds to reinforcers will determine the outcomes of the programme. Secondly, the feelings and beliefs of the caregiver are crucial as the programmes are intensive and require absolute commitment. Thirdly, it may not always be possible to implement a programme given the economic as well as emotional investment required.

Finally, and perhaps crucially, the decision to adopt a behavioural programme involves parents in negotiating powerful and competing philosophical discourses. Behaviourist interventions incite strong feelings among parents and educators; this has always been the case but is perhaps more so today than 20 years ago. We now know more about how it feels to be autistic and we have more respect for the behavioural differences which can arise from autistic neurology (alternative responses to sensory stimuli for example). One of the difficulties with behavioural programmes is that they promote norms of behaviour which do not consider (indeed which aim to reduce) this ‘otherness’. It could thus be argued that behaviourist programmes are positioned within a deficit framework in that they aim to ‘fix’ autistic children.

While I understand the tendency to line up in favour or against such programmes I wonder whether we might sometimes be in danger of throwing the baby out with the bath water. Occupying the middle ground promotes eclecticism; selecting strategically from a range of approaches, and adapting these to support my personal principles and beliefs, has turned out to be the most comfortable place for me.


Other posts in this series:


Portraits of educational thinkers are via wikipedia; images of sweets are via and

23 thoughts on “Behavioural Interventions In Autism: holding the middle ground

  1. Pingback: What Difference Did It Make? Reflections on autism and early intervention | Living with Autism

  2. Pingback: Sensory-Based Interventions In Autism: developing ‘material literacy’ | Living with Autism

  3. Pingback: Dietary And Medical Interventions In Autism: from flour to flower | Living with Autism

  4. I just wondered in joy (and felt that you may have left it out) about using both of the above approaches to demonstrate and to engage in copy behavior (as play) for concepts like NO and other limiters, thus exact behaviors, at the beginning do not have to be trained robotically nor understood, only strong responses to the limiter. I am thinking too, that it is more difficult to use don’ts than it is to prompt a substitute behavior. (stop yelling–reward, vs you are yelling, please use inside sounds–reward)

    I always LOVE(no i do not really!!) the removing prompts and rewards sections, especially if I mess up intermittent reinforcers OR he gets distracted by something and incorporates it as a new prompt OR he simply forgets the new, forgets months of stuff, and demands prompts that long have been extinguished or regresses. He’s very unpredictable (he probably IS predictable if I chose to notice each and every tiny input and output near him–it’s inhuman of me to expect this of me and I wish that I could stop feeling angry with him over it)

    I used floortime (which I think that is what you were speaking of prior to and up to the movie incident) to get engagement and then would sing songs or sign and move the random tossing of the toys, to tossing them into the toybox. I used the ten second tidy song and would randomly and quickly shift to a command like PLAY!!! (to promote rapid transitions and a comprehending of the word Play and the word clean or tidy) A variant of the one, two , three rule was also a good marker of transition–some mands are ready, steady, go. I do not so much understand the jargon for mand training but OT’s have stated that I use it modified within ABA framework, to allow adaptation and true choice making on Z’s part.

    Oh, something else! Works well with nonverbal and sensory overloaded humans! I refused to allow the children to dictate my world and my schedule. I recognized that repeated must watches of movies and ‘wopner is on’ events weren’t functional, nor were refusals of clothing ect. During floortime if I could get their attention for a picture to go with a sign and a whisper and repeat using it, and then go out and use it allll day everywhere I got generalization of a concept. Like to watch tv we sit in a movie chair, to watch a movie we sit in a movie chair…..thus the request in public to please sit in your movie chair meant that he would sit down immediately anywhere (not needing a specific chair or visual of said item)–sometimes I carried a flip book of images with said chair command image for newer situations. Sorry I got excited again 😀


    • Hi Elisa – some really interesting and useful ideas here – you have way more experience of such approaches than me so I’m really glad you got excited enough to share 🙂 Yes I agree absolutely that training doesn’t need to be robotic and that substitute behaviour is better than don’ts (my attempts at ‘don’t’ have so far been fairly useless). I really struggle with removing the prompt – I am getting better at not letting the prompt become too much part of a routine but it is still hard. For example I am really pleased at how I’ve recently trained Dylan to use the cinema (previously films could only be viewed absolutely alone with door shut) but I am now struggling to remove the sweets. I’ve left them because actually lots of people eat sweets at the cinema, but in other contexts it can be tricky. I used to do that ‘time to tidy away’ ten second song too! I need to be better than I am at ‘refusing to allow the children to dictate my world and my schedule’ – though D is probably as tolerant of my world as I am of his 🙂


      • I took a behavior mod course at college, trying to understand why consequences and rewards do NOT at all motivate one of the children. The instructor at first insisted that there was simply a flaw in my own method. She had to change her mind. I grin triumphantly, only as I can be absolved of guilt, but I was hoping for the ultimate fix ha!

        I wonder that I had to or chose to change rewards a LOT, so that the outcome was not simply pavlovian and dependent upon a closed system. I do not know if Dylan understands money, perhaps as I had to create: I used a modified token system with a list of rewards that I could use on a delay and immediate, for which any of them could buy, or choose to shift toward their own form of intermittant reinforcement as they learned to wait. Using this method I didn’t even always have to give a token(i used bingo chips–they liked visual sparkle and light color shifts). I could recall how many tokens to pay, when we got home, or end of day or end of week. The newer system also allowed me to limit food rewards and candies to a normal amount. I also used it to help them to get past not knowing what they wanted to eat or to drink. Not sure Dylan’s function level is here, though I am sure that you try to care what he might want.

        One child, who’s parents thought I could not help but let me try, was able to also use tokens as a way to request and then to place the token on or next to a live item, such as fried chicken. I put an image of that food into an altered PECS flip book and from then on parents had a way to know what foods were wanted!

        You might want to look into fading, extinguishing and intermittent reinforcement, as parts of behavior modification.


      • Hi Elisa – really interesting. Dylan is just beginning to understand that you have to hand coins over to buy things. He has his own purse now and I’m encouraging him to take charge of some transactions. I can see potential for maybe using a token system to accumulate and buy something for himself – a new DVD for example. After food, films are the main thing for Dylan. I can’t use screenings as a reward system as he has his screenings finely matched to days of the week with no room for a bonus – and he’s very much in charge. But he does like getting new films, after which he shuffles his screenings around to accommodate. So maybe I could set up a system for earning tokens for a new film. Will also look up the fading etc – thank you!


      • lol you CAN use screenings–if you want to enough to break his pattern and to deal with –UHM Stuff, I know there are many time when we do not want to deal with the CAN parts. I think we have that right, that choice. What happens in a shop if he doesn’t have enough money or tokens to get the candy (like a dvd costs 50 tokens and if he uses 5 to purchase one lemon head he has the opportunity cost of having to wait longer)


      • oh yikes he wouldn’t wait – probably point to my bag (get your purse out mum). But it would be a good thing to work towards long term. And yes you’re right about the screenings – the routine breaks down at holiday and other times and I COULD do this if I wasn’t so attached to the routines myself lol.


  5. You have so much to teach others. Parents are in flux when their children are young, and somewhat desperate to do the right thing. When Ben was first diagnosed, I searched for adults with autism, Rainman was all I knew. I found them, and it was helpful to see that his outcome might be very different, but good. I think there is a wellspring of info from parents whose attitude has changed so much from the desperate fear of the early days. There is wisdom in hearing what you and Dylan have been through.

    The worst thing, I think, is that each school of thought, whether it is diet, behaviorism, floor-time, what ever, is often the idea that it must be performed flawlessly with no derivation from the program or the failure of the child to advance is on the shoulders of the parent. That is wrong…nobody parents like that! It’s unnatural. You just keep trying until you find something that works, and if it doesn’t work, try something else! No use persevering if it’s not working out…and why not take those things that work and mix them? Parenting is not a military exercise.


    • Elisa used a magic word…predictability.

      I always felt that Ben was so unpredictable, but when I became more predictable…he did too. I think one thing we miss in ABA is that the teacher/parent becomes more predictable to the child…


      • Ah! Vygotsky would be proud of us co-constructing knowledge through dialogue like this! I write something; Elisa uses a magic word; Rose pulls the rabbit out of the hat 🙂 That’s a really interesting observation – that children benefit from the way in which their teachers become more predictable. That sounds right to me. It is always the changes in the teacher that trigger the biggest changes in the learner, I think…

        Liked by 1 person

    • I so agree with you Rose – in fact that was exactly my reason for starting this series of posts – I remember how desperate I was to do the right thing when Dylan was small, and how much pressure I put myself under. I found some old notes recently and was horrified by their intensity and anxiety – I wanted to get my younger self back and give her a hug and tell her she was doing fine and not to worry so much. As you say it’s unnatural. I think I sacrificed a lot of the ordinary joys of parenting by trying to do that perfect flawless thing. I’m glad you think what I have to say is useful – there is nothing like hindsight of course. Lovely to hear from you!

      Liked by 1 person

      • I have spoken to several groups of OTs at a National Autism conference here. I was ever so shocked to see that a great many of the students and those with license truly believed that the little blurbs that they read of how to, meant literally that. There was no watching for efficacy and no understanding of what was to be achieved by doing this to or for a patient. I said that for me, and many other parents, past the fear and puppet phase, I want my children to have meaningful actions and abilities for daily living out into self actualization. I believe this is possible when we shift our perspective, allow for flexibility. I am smiling about predictability and its uses. Exhibiting flexibility is a part of being predictable and prevents everyone’s insistence on sameness, patterns, and OCD-like living. (there are always outliers)

        {I was approached after the event and asked where I taught. I was flattered and embarrassed. So I said I’m just a mom that when my child stopped talking and refused to engage, I didn’t let him win. I didn’t assume he couldn’t or wouldn’t.}

        Liked by 1 person

      • Elisa: I am predictably stubborn about certain things and predictably let others ride. I guess Dad was right when he said, regarding parenting, “Choose your battles!” There are very few things I am insistent upon, but I seldom waver on those. I say seldom, because if I can see harm done by my insistence…Ben is more important to me than my way. There’s always tomorrow to work on it again, if need be. I’m good at making mistakes.

        I have tried to look into ABA, a choice I never consciously made …one thing that kind of shocked me is that the most common part of the test the BCBA (teachers) fail is the ethics portion…like 80% of failures are from that portion!! I wonder if that is because they believe there is a “right” answer, and not the humane one that takes priority.


  6. Pingback: 50 Shades of Beige | Living with Autism

  7. Pingback: Peeing In A Pot: a quiet celebration | Living with Autism

  8. Pingback: Education-Based Intervention In Autism: Dylan’s home learning programme | Living with Autism

  9. Pingback: Sensory-Based Interventions In Autism: developing ‘material literacy’ | Living with Autism

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s