Review: Behavior Therapy for Children with ASD

Every now and then I level some criticism in the general direction of ABA as a method of enhancing the well-being of autistic children. And generally the motivation to do so comes from what I consider quite fraudulent claims made by ABA proponents that they are, ‘scientific’ in either principle or method (they claim both), effective for everyone and/or more effective than any other teaching method out there. None of those claims stand up to scrutiny and yet they are all of them commonly cited. That ABA therapists readily claim them is quite congruent with the aspirations of that field. That those claims are broadcast with what appears to be autistic support is a bit more worrisome. I really believe that autistic self-advocates have to be very careful about what they are seen to be endorsing. “Behavior Therapy for Children with ASD” by Zosia Zaks contains most of those fraudulent claims thereby authenticating methods which really should be overhauled and preventing the development of teaching methods, which might, for a change, start out with autistic perceptions and learning styles as a foundation. It’s the latter point I worry about most. ABA, because of its non-scientific methods and narrow base just is not capable of incorporating new material into the repertoire, especially material that is not a good fit for the standard reinforceable step-wise approach of shaping, chaining and so on.

For the record I’m no expert when it comes to ABA, which got me thinking about what constitutes being an expert in any field really. Among my rambling cogitation, I came to the conclusion that an essential element is that the subject is one’s life’s work. Unless the person lives breathes and works in that area, I doubt that they could develop the necessary depth of approach to the subject matter to be a reliable expert. Not, mind you that this necessarily confers truth value to their ruminations. There are experts in phrenology and homeopathy, for example. I also don’t believe that an academic qualification necessarily confers all that much gloss to the expertise. It might and in some fields may even be considered necessary but it is not in my opinion always a necessity for expertise. History has too many exceptions. Michael Faraday comes to mind.

Looking around the myriad Hub bloggers it appears we have just two ‘experts’ in ABA, Janna Hoskins and Interverbal, both with the necessary immersion. This is or was their life’s work. They really know about this stuff. Not that there aren’t a number of gifted amateurs out there. There are; it’s just that ABA isn’t their life’s work. I also think that knowing a good deal about behaviorism in general and ABA in particular is something of a necessity if you’re into autism advocacy and it doesn’t much matter which PDD label applies to you and yours either.

And back to the claims made in this paper in the order in which they appear.

1. The best thing about ABA is that it works with everyone"

This statement was made not by a professional, but a parent, and one rather prone to overstatement. However, that does not absolve the author of this paper of the responsibility for allowing such a nonsensical claim to pass unchallenged. The one thing that is known about ABA is that it does not work with everyone. Every study ever done has a sizeable minority of children for whom this intervention did precisely nothing in any domain; language living skills or IQ. There are very few (I think the number remains at 1) scientific studies of the effectiveness of ABA. Smith Groen Wynn 2000, a properly randomised controlled study of ABA had 2/13 children who made it to mainstream classes without an aide but showed no improvement on any other measure. Not only does this intervention not work for everyone but it also does not work for everything. Sallows and Graupner's study had a minority of children who appeared to regress entirely. The most famous study of the ineffectiveness of ABA was Smith et al 1997. An archival study, he looked at a group of children whose IQs were judged too low for them to be a part of the original Lovaas study. The results after two years of intensive therapy were disheartening. Yet, a number of professionals cite ABA as the method of choice for those with severe cognitive challenges. The evidence says otherwise.

1. “Applied Behavior Analysis (ABA) is a scientific discipline that seeks to understand and improve behavior using data and scientific methods.”

I got motivated into having a closer look at ABA partly because the therapeutic aspects of intervention for autism were dominated by this bit of psychology and by an Ipsos Reid poll conducted in Canada for FEATBC. That poll insisted repeatedly that ABA was ‘scientific’, so naturally the intervention was steeped in the scientific method. Skeptical me said – really? And six months later having waded though swathes of the Journal of Applied Behavior Analysis and the Journal of Experimental Analysis of Behavior and the writings of a number of behaviorist luminaries, I concluded, in a few thousand words, that the trouble with a therapy framed around a naïve inductivism operating in a form of Jamesian Pragmatism is that it’s a long way from what a scientist would recognise as the scientific method. This was one very weird form of Science bearing at best only a very superficial likeness and really quite removed in its experimental set up from the scientific method. Now, I am not an expert in this, but Professor John A Mills, formerly of the University of Saskatchewan is and he said quite a few things about the pretensions of behaviorism among which:

“Logically, one can detect four flaws in the reasoning of behavior modifiers. First, they tend to confuse mere associative with causal relationships. Second, they are guilty of the inductive fallacy. Third, they fail to distinguish between necessary and sufficient causes. Fourth, they make the error of denying the antecedent.” [p 170]

In my view, behaviorism does for psychology what Aristotle did for Science – gets you the four element model of the universe - based on careful observation. The more I read about this psychological oddity behaviorism, the more fascinating it got. Far from the promised evidence- based, steeped in science intervention, the empirical basis, such as it is rests on limited animal studies and the philosophical basis of the rationale for extending the animal studies to humans rests on trendy atheism, circa 1950. Jack Michael, the behaviorist noted for resurrecting ‘Verbal Behavior’ following Chomsky’s devastating critique admitted quite openly that the principles of operant conditioning were never tested on animals more sophisticated than one species of rat and one species of pigeon.

“First, the vigorous applied approach called behavior modification developed without the benefit of any appreciable laboratory research with humans. At the beginning (the late fifties and early sixties), it was a direct attempt to use operant, animal-laboratory methods and concepts to alter socially important human behavior. The relevance of these concepts to human behavior was taken for granted by the behavior modifiers, as it was in Skinner's

Science and Human Behavior (1953),”

In 1950 and in most Humanities department since, to be a Deist, and to espouse those values was deemed simply incompatible with intelligent discourse. The intelligentsia were atheists. Mankind operated in the same evolutionary manner as other species and inconvenient attributes such as altruism or creativity would be explained eventually. There was no such thing as personal responsibility because that implied free will and that belonged to an antiquated religious past. In this climate, behaviourism looked perfectly at home and its offspring, ABA wouldn’t have raised an eyebrow. I wonder how many therapists tell their clients that doing ABA Monday to Friday and synagogue/church on Saturday/Sunday is a contradiction in terms?

2. “ If the behavior under investigation is hurting the person, hurting others, or is in some way ineffective, behaviorists encourage, model, teach, and provide incentives for new more adaptive behaviors.”

In other words, practically any behavior can be deemed worthy of changing and since behaviorists do not think that the person is important, merely the environment their reasons for a behavior happening are restricted a priori (and therefore unscientifically) to a fixed and fairly short laundry list. For a behaviorist what people do is usually to get something, avoid something, gain attention or escape from something. So paying attention to the attributes of the actual person in front of them doesn’t happen and because of that they are more than capable of replacing adaptive behaviours with others deemed more socially acceptable but in reality more obstructive to the autistic person’s ability to learn. Diminishing a person’s capacity to cope with their environment surely undermines their learning. Take gaze avoidance, long known to be quite important in improving cognitive processing of auditory data in children. What do ABA therapists do with that?– insist on eye contact as the first step in ‘learning how to learn’ and call upon a species of group think, ‘social validity’ as the rationale for doing so.

3. “ABA has the greatest research base so far”.

Yes, but the quality is terrible. 500 or so single subject research studies don’t amount to much since they fail to control for so many variables. My take on this as I’ve written elsewhere is that the results from interventions depend entirely on the attributes of the child, not the intervention. When the results suggest otherwise, it’s a pretty clear indication that group assignment wasn’t random.

4. “[O Ivar Lovaas] ..was one of the first professionals to suggest that autistic children could learn and should not be routinely insitutionalised.”

OK, this article has now descended into dishonesty on a scale not usually offered to parents by other than snake oil salesmen. Have these behaviorists not read Kanner? Is there any indication whatsoever that these children somehow cannot learn? Of all the lies proffered to parents, this is one of the most destructive and these are supposed ‘therapists’. They need immediate remediation. As for re-writing the history of Ivar Lovaas’, his reasons for treating autistic children had nothing to do with altruism and everything to do with his language research. Rick and Pam went straight back to the institution when Lovaas had finished with them.

5. “ He wrote a pivotal 1987 report claiming that roughly 40% of the children in one of his programs ( 9 out of 19) became “indistinguishable from their peers”

Victoria Shea wrote an equally pivotal paper debunking that claim. She took the time to read McEachin’s PhD thesis and lo and behold, only one child scored in the normal range on all the tests. No one took any data from their teachers, though their input is frequently claimed. This reminds me of Jane Gross’s article on autistic childrens’ troubles in school post treatment. While the parents called them ‘recovered’ apart from some trifling social problems, it was readily apparent that their peers thought quite otherwise.

6. “ the question is not about the ethics of ABA, but about the ethics of some professionals who provide ABA services”

When an appeal to ethics or efficacy rests not on the therapy but on the therapist, one should be very skeptical. It seems to me that ABA proponents are far too ready to blame all the failures (both ethics and efficacy) on the therapist and all the successes on the therapy. The ethical confound of behavioral intervention is that the method has no foundation, scientific or otherwise and therefore unproven methods are being used on very small children. The neat trick is that in riding roughshod over hyperlexic preferences and/or artificially manipulating the error rate for someone who learns only by making every mistake in the book (it’s a kind of learning by learning where the boundaries are) - if that leads to damage, diminished capacity for learning or whatever, the therapy is never to blame – it’s the therapist. Is that the dream therapeutic environment or what? It works! – thank the Goddess for the therapy! It doesn’t work! – it’s the therapist! Aversives are also unethical but that is only one part of the ethical challenges of ABA and not the major one in my view.

7. “ ABA therapists use a variety of teaching strategies, including prompting, cueing, modeling, shaping and chaining...... Yet many people think the only ABA teaching method is Discrete Trial Training or DTT for short.”

Except for modeling, I thought that the rest were very much a part of DTT. Nobody ever questions that DTT, the foundational platform for ABA is really the Skinner box in action for humans. The minimalist environment is not there to save the client from unnecessary distraction. It’s there to prevent the client from exhibiting any behaviour but the one under the set stimulus control – the ‘lawful’ behaviour. But parents and beginning therapists will be told that this is the key to learning for ASD children. The ABA industry did try to cloak this learning environment in scientific garb by maintaining that ASD children had a particular problem with ‘stimulus over-selectivity’. That gave them a rationale for breaking things down into very small steps, which had the bonus that each tiny step could be reinforced. This manoeuvre is called shaping and there’s a bit of chaining in there as well usually. You can get a pigeon to do a figure of eight with these methods. Should you get human children to do the same? There is no evidence that overselectivity is the case for ASD children and some that it definitely is not the case. You don’t have to be ASD to have a problem with this. So, since there’s no empirical basis for these methods why are they used at all? Good question and the possible answer I can come up with is that psychology having left behaviorism to history’s wastebasket didn’t have a problem leaving ASD children in the same wastebasket. There was even a bonus. Psychology, even now, is considered a social science and the hallmark of the social sciences is the ‘schools of thought phenomenon’. In that schema several competing paradigms can happily coexist, each with their own theory, arcane language, praxis and adherents. Behaviorists still inhabit a chapter of the Psychological Society and still have to be accommodated professionally. So they were left the hard cases – the mentally ill and the developmentally delayed. Through the sixties and seventies, their work with the mentally ill, though heavily criticized, was allowed to continue. What brought that branch of behavior modification to a halt was not the criticism but the closure of large institutions. For those who fondly believe that the APA or the APS are going to make a real effort to reform the Judge Rotenburg Centre in line with its own policies, I would point to the history, which says that is unlikely to happen though the appearance of an attempt may occur if they can be embarrassed into it.

8. Some critics say the use of drills and rewards makes children in ABA programs robotic.....The problem may also be due to a lack of skills generalization.”

Logically, or so it seems to me, if the learning is so artificial that it makes no sense whatsoever, then nothing has been learnt at all that is in any way useful. So learning is equivalent to generalization or lack of generalization indicates that nothing has been learnt. Once again, when this question arises, we have the ready explanation that this is a therapist ‘gone bad’.

9. “ basic skills of imitating, attending, learning how to learn and liking learning”

With the greatest respect for Dr Kristina Chew, this is ABA basic speak. How does one assess ‘learning how to learn’ or ‘liking learning’ for that matter? Who says and where is the evidence that autistic children learn or can learn well through imitation? Could a child who does learn readily in this way be considered autistic at all? Because a typically developing child learns this way should we assume that this is a missing developmental step and supplying the lacking skill will remediate the problem? Perhaps it is a very inefficient way for autistic children to learn. One thing that all current interventions have in common – none are based on what we know of the perceptual styles of autistic children, including their manner of ‘attending’. In fact the NRC report of 2001 was highly critical of the ‘ne’er the twain shall meet’ gulf between research and intervention.

10. Dian Keller.....”she is in the fourth grade at a regular school with no pull outs.... very verbal, happy and healthy.”

I draw attention to this because the phenomenon is pervasive. It’s not lying, exactly – just stretching the truth in what the parent believes is a good cause would be my guess. The ‘no pull outs’ and ‘happy and healthy’ state of her daughter doesn’t quite gel with her earlier statement “whose now 10 year old daughter sees ABA professionals weekly in addition to follow-up therapy at home.” This isn’t all that blatant compared to say Kit Weintraub efforts in that direction, but it doesn’t persuade me that she’s quite as healthy as advertised.

11. “Where a non-invasive desensitization program procedure could open up a wealth of choices.....The child’s world is larger; he has fewer things to be stressed about and more things to enjoy”.

Earlier I broached the idea that aversives were for me at least, not the core ethical issue with ABA. I think it’s the lack of an acceptable empirical foundation for its practices that I find unethical and more unethical to subject a small child to ‘desensitization’ procedures of dubious and unproven efficacy in the name of ‘choices’ but not the child’s choices, obviously. Where exactly is the evidence that ABA offers ‘desensitisation’? I had thought that ABA poured scorn and vitriol on desensitization procedures employed under the aegis of occupational therapists. What if it’s merely training a small child to ignore discomfort for your benefit? In which case, what have you taught the child apart from expecting the child to learn in less than ideal conditions. Maybe the take home message indelibly etched on that child’s mind is that they don’t matter and that external appearance but not the reality is of prime importance. Realistically, this child now has a lot more to be stressed about and fewer things he will be able to enjoy because there is no proof that a few M & Ms can overcome any degree of tactile defensiveness. The other lesson in this is teaching not just the child but also the parents very poor problem solving strategies. Seems like the waste of a teachable moment to me. Here the child learns only that people are shallow. They are but he or she could also learn to approach ‘appropriate’ clothing systematically. One can never learn too much about the necessity for camouflage. However, it does not have to be uncomfortable camouflage.

12. “Certain ABA techniques such as modeling, visual schedules and step by step skill acquisition are particularly useful, since autistic children tend to be weak in executive functioning areas like planning, organization and problem solving and strong in visual learning.”

Quite possibly none of that list is accurate except the last point about visual learning and that would not be universal. Fred Volkmar wrote a really good review of what’s happening in autism and he made the point that the attribution of deficits to autism like those above tend to be based on very little data. Claims like these need to be checked. The chapter on Learning in Autism available online would be good checking source as well as anything that Dr Morton Ann Gernsbacher has on her site. It is also noteworthy that ABA is incapable of generating any of that list. That style of research doesn’t fit their paradigm. Professor Mills characterized it as ‘corrosively constricted’.

The last point made was that ABA has evolved and learnt new child-focused ways of doing things. If by evolved, they mean co-opting the parental handbook and borrowing freely from TEACCH, Floortime and NET to fit in with standard speech language pathology praxis, then they have indeed evolved but it’s a Lamarckian rather than Darwinian form of evolution. What they have done is effectively obliterated opportunities for operant conditioning to occur and I think that is a very good thing. I’m with Deci and Ryan – rewards can seriously mess with learning, albeit only to a point. Lisa Jo Rudy of autismabout.com made a comment that the good therapists are very much alike in how they do things regardless of the label on the lapel and that’s a comfort. The other source of comfort is “The Myth of the First Three Years’, which should be required reading for those expounding the short ‘window of opportunity’ drivel. Brains are malleable and stay that way for far longer than folks are implying. Interestingly the kinds of things for which a window of opportunity may exist are vision and grammar (up to puberty).

In the scheme of things, autistic advocates could make a push for just one thing intervention wise – ‘evidence please’. That applies to every intervention out there not just ABA and we should be a lot more insistent about it than we have been. This and that ‘treatment’ ™ are usually nothing more than commodified common sense that someone thinks they can sell at a premium. So the only way to prod them into doing what might be some good is to be highly critical of their foundations and insistent that they do incorporate what is known.