© David Sponder, L.E.P., Board Certified Behavior Analyst; DIR Certified Floortime Advanced
Executive Director, Sponderworks Childrens’ Services

Criticism of the Current State of the Field and the Advice Given to Parents

There is a lot of confusion among parents regarding selection of a program that is right for their family, especially when they are told that ABA is the only research proven method for intervention.  First of all, it’s misleading and it points up a problem with how self-referenced the field of ABA can be.  The Journal of Applied Behavior Analysis (JABA), the Journal of the Experimental Analysis of Behavior (JAEB) and other journals published by behavior analysts tend to lack references to sciences that provide the theoretical and research support for the use of developmental, relationship-based methods or any methods referencing typical development (developmental psychopathology based) or; neurodevelopmental methods.

The field of ABA could be more of an open system, and there is increasing evidence that it is indeed becoming more open to the similarities and complements of other methods and theories.  But it has a long way to go, guidelines currently in use have had a stagnating effect on innovation and the average parent is unlikely to meet an unbiased, widely experienced person (i.e. in more than one method or system of intervention).  The field as a whole is much too dismissive of other theories of behavior and intervention.  Subsequently there is an embarrassing and appalling lack of understanding of natural childhood development (especially emotional development) used in the behavioral treatment of developmental disorders.  We endorse the basic principle of Developmental Psychopathology: that to understand abnormal development and behavior, one must first understand normal development.  We also believe in Edward O. Wilson’s (who pioneered the field of Social Biology) ideal of “conscilience” among scientific disciplines.  Essentially, conscilience occurs when various and seemingly unrelated fields (e.g. systems theory and behavior; genetics and behavior, etc.) find common underlying truths.

By virtue of an extensive survey of professional training programs across the country, I have found that the overwhelming majority of Master’s Degree or Ph.D. programs in Behavioral Analysis or Autism feature few if any requirements for understanding typical cognitive, language, or emotional development; few if any requirements for understanding neurological or physical development; no requirement for the study of systems theory; and no real study of interventions developed by developmentalists.

As a result, the current state of the field of FAB is remarkably uncsonscilient with other fields outside of behaviorism, therefore, it is fair to ask how much a professional knows about what he or she appears to reject.  (For an excellent critique along these lines, I refer you to “Shortcomings of Functional Behavioral Assessment; Functional Behavioral Assessment: Deeper Understanding of Behavioral Handicaps or Trojan Horse? by John Stewart, NCSP)

Here again, I cite the Autism Society of America’s policy on treatment options:

ASA Parent’s Choice/Options Policy

The Autism Society of America promotes the active and informed involvement of family members and the individual with autism in the planning of individualized, appropriate services and supports. The Board of the Autism Society of America believes that each person with autism is a unique individual. Each family and individual with autism should have the right to learn about and then select, the options that they feel are most appropriate for the individual with autism. The ASA believes that all individuals with autism have the right to access appropriate services and supports based on their needs and desires.  We believe that the decisions should be made by both the parents and the individual with autism in collaboration with professionals.  Services should enhance and strengthen natural family and community supports for the individual with autism and the family whenever possible. The service option designed for an individual with autism should result in improved quality of life. Abusive treatment of any kind is not an option.

We firmly believe that no single type of program or service will fill the needs of every individual with autism and that each person should have access to support services. Selection of a program, service or method of treatment should be on the basis of a full assessment of each person’s abilities, needs and interests. We believe that services should be outcome based to insure that they meet the individualized needs of a person with autism.

Our prescription based approach is meant to deal with the problem of no single program fulfilling every need, but it takes a whole lot of extra effort to develop expertise in both behavioral and developmental methods and few people do it.  People can advise you about what they know, but unless they have really studied and implemented a program, they have no basis for advising you against it.  Ironically, newer forms of ABAsuch as pivotal response therapy feature principles and methods developed and tested long ago by Developmentalists, such as following a child’s lead and developing pivotal responses.  And in the most cited research articles and books about the newer methods, there are few if any references to the developmental literature or researchers that first tested the methods.

What is a Functional Analysis of Behavior

The whole debate becomes a little ridiculous when you take a closer look at the guiding principles of ABA.  This is because any therapy that uses interaction and modification of the environment to promote developmental change or behavioral improvement can be analyzed functionally, and can be shown to follow ABA principles if the program is well designed and understood by the interventionist.  Below, I will offer a Functional Analysis of interaction principles used in the various methodologies.  But first, it would be helpful to understand what a Functional Assessment/Analysis of Behavior (FAB) is.  I will try to keep this non-technical.

The primary purpose, and therefore, one of the optimal outcomes of a FAB is the determination of the motivation of a specific behavior or class of behaviors (different forms of behavior that are similar in their relationship to the motivation).  These motivations are called “functions of behavior,” the meat of functional analysis.  At SCS, we further analyze functions in terms of ‘functions of’ (unintended or indirect results of behavior: such as losing friends), and ‘functions for’ (behaviors that can be considered intentional – even if it is undesirable).

Prior to developing a hypothesis about the reasons behind a behavior, one must observe and define in measurable terms what that behavior actually is.  For instance, it is easy to observe and measure the number of times a child leaves his seat or completes work or says a 3 word sentence.  But describing a behavior as a “bad attitude” or “angry behaviors” can be a source of confusion among different observers, and failure in intervention.  The important thing in this defining or “operationalizing” a behavior is that once defined, different observers can agree that they are measuring the same phenomenon.  This is called, “inter-observer agreement,” which provides the test of how well the behavior has been defined.

Since no one can ever determine with complete accuracy what motivation seems to be operating at any given time, the determination will always remain a “hypothesis of the function of the behavior.”  Hypotheses will always range from less to more accurate, so it is vital to consider as many forms of relevant information as necessary (always including the history of the individual and the pattern of the behavior, as well as environmental, health and developmental data.

Before fully engaging in teaching replacement behaviors, some effort should be made to test the analyst’s hypotheses (hypothesis testing).  This describes the process of testing out different ways to elicit the behavior by changing aspects of the environment.  Responses to various forms of intervention can also be tested preliminarily by a similar process.

Once a hypothesis of the ‘function(s)’ of the behavior is done, the analyst looks for “replacement behaviors” for the behaviors the individual is currently using.  So if a child engages in a behavior such as a tantrum, and the hypothesis is that the tantrums are primarily used by the individual as a means to avoid certain demands, then the analyst can propose possible and more satisfactory behaviors to meet that function, or make modifications to the environment that allow the individual to solve the prom or conflict in other ways (e.g. ask for a break; engage in verbal problem-solving/conflict resolution, etc.).  The analyst also has the responsibility of proposing teaching procedures to accomplish the learning of replacement behaviors, as well as ways to ensure the individual can use them consistently when needed.

Equally important is the measurement of behavior change following intervention.  It is well known that an individual can do the same thing for a variety of reasons, depending on the context perhaps (single form: various functions), and conversely, an individual can do a variety of things for the same reason (various forms: single function).  The analysts job is to make sense of this and to develop procedures that work.  If this is done properly, it is often a complicated thing to do – as people and their behaviors can be very complex.

The basic data points of functional analyses are the stimuli that elicit responses (antecedents), the responses themselves (behaviors), and the effects the behaviors produce in the individual or the environment or both (consequences) or (ABC).  Intervention might center on modifying the antecedents or the consequences or both, and the measurement of success or failure is the change in behavior that results from such manipulation of the environment.

This functional unit of analysis: antecedent ® behavior ® consequence is rarely enough, because the results of behavior produce new stimuli that the environment responds to, and the individual responds to the responses of the environment in continuous fashion.  Therefore, behavioral interaction with the environment should be seen as dynamic.  A simple ABC is usually overly static to analyze what is essentially a dynamic process.  Unfortunately, systems theory has only recently even been discussed in FAB, and has not become common in FABs performed by behaviorists.

Consistently in the behavioral literature, the following motivations are identified almost to the exclusion of all others:

  • behavior that elicits attention or obtains access to desired events (e.g., access to toys, desired activities);
  • behavior that enables the person to avoid or escape demands or other unwanted events/activities, and;
  • behavior that occurs because of its sensory consequences (relieves pain, feels good, etc.).

We note here that the primary driving force in the evolutionary expansion of the human brain occurred due to environmental pressure to engage in complex social relationships.  Traditional behaviorists typically place behaviors in one of the three functions above, and many believe that any behavior can be subsumed under one of these functions.  In light of the primary function of social intelligence in human development and the brain, these functions are limited to instrumental interactions among people, not interpersonal or intersubjective motivations.  This has led to interventions devoid of the tools of social intelligence – dynamic appraisal and dynamic responding.  That is why after years of non-developmental intervention, many individuals merely become more knowledgeable individuals with still-disabling autism, as the core [read:] emotional deficits of autism remain unremediated.

Despite my serious criticism of the stagnant state of the field in which I work and the collective hubris of many that call themselves “behaviorists,” a proper Functional Analysis (i.e. informed by developmental data and the sciences related to human neurological, social and emotional development) of developmental interventions can show the actual conscilience of the various behavioral and developmental interventions.

Functional Analysis of Methods Used in Autism Treatment, including “Developmental” Methods

DTT: Discrete Trial Training features the most basic form of functional analysis – the linear relationship between antecedents, behaviors, and consequences.  Using highly controlled presentations, the ABC elements are laid bare and the results are easily measured.  This is the true reason it is the most widely used intervention, not it’s validity as an intervention.  Because of the training method, the stimuli produce responses mainly from long term memory and do not activate the emotional brain in a typical way.  Dynamic social responses rely heavily on limbic influence on feelings, motivations, and behavior, and these are systematically ignored when behaviors are taught in a static (rehearsed and drilled) fashion.  No amount of “generalization training” addresses the fact that dynamic function in the environment is highly predicated on activation of the emotional brain, and the result of this therapy when used alone is typically a smarter autistic who is as unable to function in a truly dynamic social environment as ever.

Floortime: Floortime is the most psychodynamic of the research-based interventions for autism.  By design, the adult follows the child’s lead in the early stages of intervention.  But the point is to engage the child in a co-regulated interaction where each partner’s behavior stimulates congruent responses from the other – in continuous fashion.  Therefore, a 2 minute interaction could involve dozens of continual stimuli, responses, and consequences (called “circles of communication”).

The interventionist sees the child’s behavior as the antecedent, because the focus of intervention is to create a connection between affect (feelings, intent; motivation, and the outward expression of these as observed in emotional and symbolic communications emanating from the partners) and motor planning (formulating responses related to intent).  The adult responds by facilitating the child’s behaviors related to the [hypothesized] intent of the child, and/or providing challenges (stimuli in which the child is to formulate a response) that result in developmental growth.  The Therapist looks for opportunities (again the child’s behavior) to create extended, intersubjective responding.

It is this intersubjective component that annoys behaviorists the most, because it makes assumptions about the inner workings of another person’s mind.  But this is what we do constantly in human interaction – we are always sussing out the motivations and intentions of our social partners.  People who cannot do it are well – autistic, so Floortime addresses the core deficits of autism.

And what is a functional hypothesis of behavior anyway?  It seems that here, the only difference between the way a Floortimer and a Behaviorist analyzes behaviors is that the Floortimer will acknowledge the intersubjective nature of human interaction and be unafraid of analyzing it.  He or she sees emotion signaling as a valid domain of expressive behavior, and the ability to read emotion signaling in a normal way as a valid domain of receptive knowledge – as demonstrated by the congruence (or lack of it) in the child’s response.

In Floortime, the interventionist has to formulate hypotheses on the fly, because a major emphasis is on fluid and spontaneous responding.  The hypotheses are continually evaluated in terms of the child’s responses.  Because the interventionist must respond fluidly to the child’s behaviors – there is no time to stop the interaction to take data.  This really sets it apart from PRT, a behavioral form of child-led intervention, and unfortunately has become a source of criticism from behaviorists.  This is totally unfair because developmentalists have used video/film for this purpose for more than 40 years.  This allows functional analysis at the micro and macro levels.

One difference in the method is that the interventionist is looking for a congruent response, one that facilitates the ongoing interaction.  The interventionist is less concerned about eliciting any specific or delimited form of behavior than a behaviorist might be.  If a child produces a response that at the moment is not consistent or fully intentional on the part of the child, the interventionist responds in ways that “shape” more efficient responses – which is exactly the same thing as the behaviorist principle of behavioral shaping.   The Floortimer always has the response (skill) they would like to elicit in mind – the “behavior” or the ‘B’ in FAB, and develops a working knowledge of how to consequate (respond to the behavior of) the child.  This consequation (a behavioral term), when it is sensitive to the child’s emotion signals is referred to as attunement.  The more attuned the interventionist, the more the child responds and the more “material” is presented by the child to the interventionist (opportunities) for behavioral shaping (consequation with the intent of forming specific forms of behavior).

Real Floortimers must be well versed in emotional development and be able to recognize emotion signals emanating from the child.  They must know the order of development of these signals, which provides the bulk of behavioral goals for the child.  If these behaviors are taught out of order, the brain recruits cognitive systems designed for stepwise thinking, not intersubjective responding.  Emotional or limbic parts of the brain are designed for this and respond most efficiently to emotional input (stimuli).

Without being able to recognize emotion signals, we could not engage in warm, dynamic, intersubjective and spontaneous social interaction – so it is reasonable to be able to teach children to respond to such signals by systematically exposing the children to and helping them respond to them.  Most emotion signals are nonverbal, so there is a great deal of emphasis on providing emotion signals (stimuli) for the child to respond to.

A criticism I have along with many others of Floortime in its purist forms is that in a given period of time, you may not have as many opportunities to work on the skills you want.  Oh, there are always opportunities for teaching, but you may or may not get the opportunity to work on the skill you intended.  In theory, Floortimers will tell you that they will respond and shape any behavior that is developmentally appropriate for the stage of development in which the child functions, and Greenspan and Wieder have done a thorough job of delineating these behaviors by stage – as observed from extensive research into neurotypical infant and child development.

But we often prefer a little more structure than simply following a child’s lead anywhere.  When we do it, we will set up the environment to promote opportunities – which often involves not using toys or objects to support interaction.

Pivotal Response Training: This is another form of child-led method, developed by disciples of Lovaas that found that DTT relied too heavily on manipulative, unnatural forms of motivating the child.  “Motivation” is considered a foundational, pivotal response underlying all others.  Therefore, they look for ways to produce natural motivation from the child.  Too bad that emotion signals and emotion-sharing are not identified as stimuli for the child to respond (although so-called “social reinforcement” such as praise can, but doesn’t necessarily or always result in emotion-sharing).  Access to essentially non-social stimuli constitutes the motivating consequence for social behaviors, which can reinforce the autistic child’s natural bias towards instrumental (and not intersubjective) forms of motivation.  There is only the hope that learning to interact in such a way will lead to the child developing a taste for emotion sharing, but there is no systematic way identified in the method to elicit the child to reference other people’s affective signals or to be able to read the non-verbal ones.

In the emotion development-informed methods such as Floortime or RDI, emotion-sharing is used as the main “reinforcer” – not mere access to a toy or activity the child likes.  Granting access to toys or activities the child already likes as a “reinforcer” can exacerbate the self-referenced nature of the individual with ASD, and turn the relationship into a series of instrumental contracts with the child.  We do not want children to evaluate their own behavior in terms of obtaining access to things; we want them to evaluate their behavior in terms of the emotional and social responses of others – so we would never develop such contractual relationships with the child.  True, the social/emotional forms of feedback are contingent upon responses from the child, which conforms to traditional notions of contingency in functional analysis, but they are not related at all to the child’s access to things (non-social objects).  Should your motivation come from wanting to be with me – or to get to your video game?  If it’s the former, you’re normal.  If the latter – you remain socially and emotionally impaired.

We use PRT sometimes to enhance how children play with toys – but not people.  If we want to teach social/emotion interaction skills (in other words – remediate the core deficit underlying all the other core deficits of autism), we modify PRT to provide emotion signals as “reinforcers” for the child.  This will not work without first facilitating emotion-sharing without physical objects involved.

Guided Participation Teaching (GPT): An oversimplified analogy for GPT might be: “behavioral methods: emotional goals.” Techniques such as Stop the Action are really forms of differential reinforcement and behavioral shaping.

Note: Guided Participation Teaching is a concept identified by Russian Psychologist Lev Vygotsky and further elaborated on by Anthropologist Barbara Rogoff.  Psychologists, Steven Gutstein and Rachelle Sheely adapted many of the principles of GPT for autism in their method known as Relationship Development Intervention or RDI.™  I was certified in the RDI method from 2004 to 2012, choosing to give up the certification voluntarily and in good standing, in order to pursue Floortime Advanced Certification.

Based on work on systems and human behavior by Alan Fogel, a change in the ongoing regulatory pattern of an interaction is considered the antecedent in which a child is to respond.  In this respect, it is the same as Floortime, except that the adult structures and teaches with specificity the initial regulatory pattern.  The Guided Participation teacher is looking for specific and delimited skills one at a time as the behaviorist does, but the difference again is the focus on emotion reading and signaling skills and the emphasis on the proper neurologically based emotional developmental sequence of skills instruction.

We’ve done DTT by using emotion signals exclusively for feedback to the child.  We’ve done GPT in situations where what we want to teach is how to use an object or perform an instrumental skill such as getting dressed.  We’ve used Floortime to make anything we do fun, so we can get kids to want to wash their hands and brush their teach from the foundations built from intersubjective connecting with the child.  And we the Stop the Action technique to deal with inappropriate behavior.

This is what is possible when you can truly find the conscilience among the methods.  So if a professional tells you their method is incompatible with the any of the others you might want for your child, you now know better.  Maybe they should read this.  And you should look for professionals who are in a position to evaluate therapies objectively – not out of bias or ignorance.