Letter from Sharon A. Reeve, PhD, BCBA, Vice President of the Association for Science in Autism Treatment (long letter)
May 18, 2006
Time magazine has a long-standing track record of publishing stories that promote awareness about autism. For this, the magazine should be applauded. However, in your May 15, 2006 issue, the article “A Tale of Two Schools” by Claudia Wallis perpetuated a number of myths, biases, and misrepresentations about certain autism treatment approaches.
The article begins by saying that parents of children with autism “enter a dizzying world of specialists, therapists and, alas, purveyors of snake oil. Getting the right help quickly is paramount, but it is hard to make good decisions when you are in a panic or fighting despair.” These statements seemed to suggest that the article was going to help parents sort out the “wheat from the chaff” in autism treatment. Instead, the article wasted an opportunity to publicize an evidence-based approach to autism treatment, applied behavior analysis ( ABA), by putting it on level footing (but sometimes beneath) with another approach, Floortime, which lacks empirical peer-reviewed outcomes research supporting its use. The end result of this article is that it fails to clarify important concerns surrounding effective treatment for autism. For example, the description under the article’s title states that “Early intervention in a well-run program is the key to helping autistic children reach their potential. A close-up look at two approaches.” This statement implies that as long as the intervention is put in place early and is “well-run,” then it doesn’t matter which approach is used. Unfortunately, this assumption, albeit incorrect, is fairly common in discussions of autism treatment. Rather, the more important issue should be which “early intervention” has outcome data to support its use. A large and continually increasing number of published peer-reviewed studies have demonstrated that children who are part of an ABA program are able to make substantial gains across a wide range of skills and abilities.
Although the article does mention one of the landmark studies demonstrating the effectiveness of ABA for autism treatment, that by Lovaas in 1987, the article then asserts that questions have been raised about the ABA model in recent years, presumably as a result of Smith’s failure to fully replicate Lovaas’s findings. Not mentioned in the article, however, is that a recent 2005 study in the American Journal On Mental Retardation by Sallows and Graupner, who used the same methods as Lovaas did in his 1987 study, showed the same degree of gains in children’s IQ and skills as those observed by Lovaas. This repeatability of effects further strengthens the assertion that the research evidence for ABA’s use in programs for autism is sound. Perhaps Wallis’ article would have been more accurate if she asserted that “recent years have brought further evidence about ABA” rather than questions about its effectiveness. The article also states that the children in the Lovaas study in 1987 passed first grade, it failed to mention the follow-up study conducted by Tristram Smith and his colleagues in 1993 that demonstrated that out of the nine children who had the best-outcome, eight maintained their treatment gains. Treatment gains were also not merely described as passing first grade but by these children being “indistinguishable from their peers.”
The article further states that “ABA derives from the classic work of psychologist B.F. Skinner, who showed--mostly in animals--that behavior can be altered with carefully repeated drills and rewards.” There are two problems with this statement: one it perpetuates the myth that ABA is similar to animal training. Since the 1960s, thousands of studies in ABA have been conducted with people, both typical and developmentally delayed. The research findings in these studies form the core of ABA’s knowledge base as it pertains to explaining how we can best learn. Although it is true that Skinner and his colleagues did much of their early work with animals, such studies were conducted to identify the most simple and basic principles of learning, all of them unknown and uninvestigated at the time. Such a tactic is similar to how biologists who wish to understand the basic processes of life began by identifying the functioning of simple components, like cells, before moving on to investigating the workings of more complex systems. In fact, Skinner’s later work focused almost exclusively on research in the development of more effective teaching methods for our education systems.
The second problem with Wallis’s statement is that it simplifies the complex discipline of ABA by erroneously distilling it down to two parts: “drills” and rewards. Investigations in ABA as to how we learn have concerned questions such as: do peers make better teachers than adults? Is it better to teach expressive language before or after teaching receptive language? Do children learn better when they have more choices for rewards? Do video models work better to teach skills than live models? Can children learn to be more independent using activity schedules? What is the optimal number of steps to teach conversational skills? There are thousands of additional questions. The point is that ABA seeks to uncover principles about how we learn and to then apply those principles to socially significant skills and behaviors.
The article states that “robotic behavior” is one shortcoming critics attribute to ABA.” Yes, this “shortcoming” is frequently expressed by “critics” of ABA programs. However, there are no peer-reviewed studies that have ever lent any support to this characterization of ABA. One of the hallmarks of any quality ABA program is to ensure that teaching involves learning how to engage in activities that should ultimately be fluid and natural. It is true that certain skills may be somewhat less fluid early in the learning process, just as they would be for any person. Do you readers remember the early stages of their learning a new language or learning to play an instrument? Do they remember how “robotic” their behavior appeared at first? Such “robotic” behavior, when observed, seems to be more a function of an early level of skill learning than it does an outcome of an ABA program.
The article states that an “inability to use trained skills outside school” is one shortcoming critics attribute to ABA.” This is the old “failure to generalize” myth about ABA. No empirical study has ever shown that ABA somehow decreases the likelihood of skill generalization. In fact, one of the hallmarks of any quality ABA program is to ensure that skills to be learned are always taught so as to maximize their use under new situations outside of the school. Skills are taught with varying materials, in varying settings, and with multiple teachers so as to promote generalization of skills. ABA recognizes that the usefulness of any skill would be severely limited if it did not generalize to new situations, persons, etc.
The article states that “ ABA was once famous for its M&M rewards, but better programs now tailor positive reinforcement to the child's preferences.” Why did the author not simply assert that “ ABA programs tailor positive reinforcement to the child's preferences.” This would make note of the individualized approach and opportunities for choices inherent in an ABA program. Why did the author feel there was a need to present the common derogatory misrepresentation of ABA therapists as “M&Ms pushers”? It is also puzzling as to why the author would include a comment about a child working for a packet of ketchup. Although the author might have felt that this was humorous, it seems to promote the idea that “unnatural” rewards are used in the ABA program depicted in the article. Any parent of a child, typical or otherwise, however, would be quick to point out that many children have some “odd” preferences for foods and activities. ABA programs initially use whatever reinforcers the child prefers early on when a child has few skills and preferences. Over time, however, a wide variety of less contrived rewards come to be preferred by the child as they are introduced into the child’s education.
The article states that proponents of ABA “concede certain weak points” about the effectiveness of ABA. Although this is a true statement, its inclusion in the article under the section describing ABA and its absence under the section describing the Floortime model seems to suggest that “weak points” are to only be found in the ABA model of treatment. Given the reliance on scientific analysis of learning outcomes in ABA, it is not surprising to find that ABA practitioners have identified techniques and outcomes that need improvement. The discipline of ABA is a dynamic one, as is any active science. The techniques used and learning principles discovered are constantly under scrutiny and are tested for effectiveness. When viewed this way, isn’t it more of a “weak point” of the Floortime approach since such a scientific analysis is not undertaken in that treatment model?
The article states that “While Greenspan has published impressive long-term results, his critics say there's an absence of controlled, randomized studies.” The absence of peer-reviewed controlled research studies that indicate the effectiveness of the Floortime approach has nothing to do with “critics.” The fact is that there aren’t any. In fact, it would be particular difficult if not impossible to conduct such studies given that Greenspan himself has often said that ABA may be combined with Floortime. If this is a common practice in Floortime programs, then isn’t it possible that whatever “impressive long-term results” Greenspan has reported are due to the ABA component?
A staff member from the Celebrate the Child School, who was described as having previously worked in an ABA program for six years, was quoted as saying that "It [ABA] does a great job with skills, but the kids lacked the ability to think on their feet, to problem solve and to engage socially." This misrepresentation suggests that children who have gone through ABA programs are somehow rigid and distant from others. Nothing could be further from the truth. Children in ABA programs actively learn to interact with a variety of situations and people. They are frequently presented with opportunities to “solve problems” and to “engage socially.” Whenever it is observed that a specific child has not yet learned the skills to fully participate in such activities, then the child and his or her teacher will work on these skills until the child is able to learn them.
A staff member from the Celebrate the Child School is also quoted as saying that “the ABA emphasis on ‘looking normal’ doesn't address the reasons for behaviors like flapping and rocking: Those are organizing strategies to cope with anxiety.” In fact, ABA’s emphasis is certainly not on “looking normal.” Rather, it emphasizes that children have the right to learn how to do the activities that typical children engage in and enjoy. ABA emphasizes the learning of skills that provide for opportunities and choices in the child’s current and future life. If a child engages in “flapping” or “rocking” for large portions of time, then it is very difficult for other skills to be learned. In addition, it is irresponsible to print the blanket generalization that such types of behavior (“flapping” or “rocking”) are “organizing strategies to cope with anxiety.” Behavior analysts do not assume what the function of a behavior such as rocking is without first conducting a functional analysis. This allows the ABA therapist to understand why a certain child might be engaging in such behavior. This individualized approach for each child is another characteristic of ABA that was lacking from the article.
It seems that the author’s intention for writing an article of this type was to provide parents with options when it comes to selecting an autism intervention. When it comes to the well being of our children, there is no denying that options are important. If a child is diagnosed with a severe illness such as leukemia, for example, then it is the responsibility of the child’s medical doctor to inform the parents about their treatment options. The medical doctor would then indicate to the parents the level of effectiveness of each treatment as evidenced in the research literature and would most certainly recommend the treatment that had the best chance for success. This information would give the parents the ability to make an INFORMED decision based on objective, accurate information. In the case of Wallis’s article, however, rather than providing parents with an opportunity to make an informed decision and to positively affect the lives of thousands of children with autism, her misrepresentation of the two autism treatment options only serve to perpetuate the “dizzying world” of autism treatment that she has initially acknowledged was problematic.
Who will be there to console these parents when years later they may realize they made an ill-informed decision regarding their child’s intervention based on inaccurate information from Time magazine?
Sharon A. Reeve, PhD, BCBA
Coordinator of Graduate Programs in ABA
Guidelines for Submissions to the ASAT Forum
There are a number of internet forums that enable members of the public, both parents and professionals, to share their experiences surrounding the diagnosis, treatment and ongoing progress of a child with autism. The ASAT forum is another such venue; however, this forum will host contributions that reflect ASAT’s mission and guiding values:
“to disseminate accurate, scientifically sound information about autism and treatments for autism and to improve access to effective, science-based treatments for all people with autism, regardless of age, severity of condition, income or place of residence.”
ASAT’s guiding values:
“ASAT is committed to science as the most objective, time-tested and reliable approach to discerning between safe, effective autism treatments, and those that are harmful or ineffective. ASAT supports all scientifically sound research on the prevention, treatment and cure of autism, as well as all treatments for autism that are shown to be effective through solid scientific research, regardless of discipline or domain.”
Please note, however that the material selected to be placed on the forum represents the views of the individual authors of each selection, and does not necessarily represent a formal position taken by the Board of ASAT