By: Roberta Brancato Daversa (May 2001)

Good science typically forms a basis for further research and refinement of methodology. This holds true for Lovaas’ early behavior modification research. Today, the application of the principles of behavioral science is the framework for several different ABA models of autism intervention.

The efficacy of ABA in the treatment of autism has achieved growing acceptance among families with autism, intervention professionals and educators. Along with the decision to pursue ABA intervention there is often an understandable confusion about the various models. This paper will provide a brief look at the origins of ABA in autism treatment, a survey of the UCLA Young Autism Project (Lovaas, 1970) and an overview of other ABA models. It will hopefully provide insight on the current scope of ABA in autism and contribute to the understanding of the distinctions among the five models discussed:

In 1938 American behaviorist B.F. Skinner's Operant Conditioning theory was published in his first book The Behavior of Organisms: An Experimental Analysis. In it he examined the basic mechanism of behavior change (learning). Simply stated, the consequences of the behavior influence the likelihood that the behavior will occur again.

Skinner's laboratory research demonstrated that a behavior would increase when it is followed by a reward such as a highly desired item. Behaviors will decrease in frequency when they are followed by punishment (social disapproval, loss of privileges).

At that time, psychology was a relatively new science. It was an offspring of philosophy and science, born in 1879 when the German scientist Wilhelm Wundt established a laboratory to explore human consciousness and mental processes. This event signaled the beginning of a scientific approach to the mind and human behavior, and a formal discipline to pursue it (Ashcraft, 1998:16).

In the early years of psychology, the science was dominated by interest in mental states and research methods of introspection originally developed by Wundt. By the turn of the century however, a small number of scientists became concerned with more concrete areas of behavior such as response conditioning. American psychologist John B. Watson was one of them and firmly rejected mentalistic approaches in psychology. In 1913 he published a paper that is now known as the "behaviorist manifesto". In it he advocated for a psychology without the terms consciousness, mental states, mind, content, introspectively verifiable, imagery, and the like (Watson, 1913 in Ashcraft, 1998:18.) Thus, by calling upon other psychologists to also reject introspection and interpretation of mental states as scientifically valid, and to develop psychology as the science of observable and quantifiable behaviors, Watson began the era of behaviorism which was to dominate American psychology through the mid-1950s.

In the 1960s, behaviorists began applying Skinner's theory to the development of teaching methods. Some, like Ivar Lovaas at UCLA, designed programs specifically for children with autism. Until then autism treatment was generally based on a psychodynamic model . . . offering some hope for recovery through experiential manipulations. By the mid-1960s, an increasing number of studies reported that psychodymanic practitioners were unable to deliver on that promise (Rimland cited in Lovaas, 1987). These failures prompted some professionals to abandon Kanner's position that children with autism have potential for normal intelligence and led to an emphasis on organic theories of autism that offered little or no hope for major improvements through psychological and educational interventions (Lovaas, 1987). Meanwhile, applied behavioral approaches were firmly suggesting otherwise. Children with autism were making treatment gains within behavior modification programs and these findings were making their way into professional journals (Lovaas, 1987). These two occurrences, the increased reports of the inefficacy of psychotherapy and the documentation of favorable outcome with behavioral intervention, made the 1960s pivotal for the study and treatment of autism.

Thirty years of research demonstrated the efficacy of applied behavioral methods in reducing inappropriate behavior and in increasing communication, learning, and appropriate social behavior. (Satcher, D. 1999)

Ivar Lovaas began using behavior modification in autism treatment during the 1960s.

A follow up study at around age 12 (McEachin, Smith & Lovaas, 1993) confirmed that the children in the experimental group maintained their gains over the control groups. The 9 children who achieved apparent normal functioning were given extensive evaluations by naïve professionals indicating that 8 of them were indistinguishable from average children on tests of intelligence and adaptive behavior. Thus, behavioral treatment may produce long-lasting and significant gains for many young children with autism (McEachin ,et al., 1993).

By applying operant learning principles to autism intervention, Lovaas left behind the disease treatment approach and replaced it with the teaching of observable appropriate behaviors. The locus of intervention was changed from treatment to teaching (Lovaas, 1981:xi).

Teaching goals were broken down into behavioral categories and presented according to the following schedule (Lovaas, 1987):

Other key innovations of the UCLA project were:

Lovaas hypothesized that the intensity of instruction for these children, 40 hours a week of direct one to one with additional carryover by parents (to extend teaching to almost all waking hours), would approach the influence of the natural environment on a typically developing child. The average child spends most of his time learning from his environment; he remains engaged with his surroundings in a way that a child with autism cannot. . . The average, or common environment that does so well for he average child does not fit the needs nor provide the structure necessary to be a good teaching / learning environment for these exceptional children (Lovaas, 1981: xi). One of the goals for the experimental group was to design a program in which the child with autism was always learning. We hypothesized that construction of a special, intense, and comprehensive learning environment for very young autistic children would allow some of them to catch up with their normal peers by first grade (Lovaas, 1987).

The place of service was moved from the institution to the child's home and community. The objective was to teach the children to function in the real world. In a previous study, Lovaas had used intensive behavioral teaching methods with children who were institutionalized. They were able to make surprising and significant gains while in treatment (1 year) but regressed upon discharge from the program. The importance of the environment in maintaining learned behaviors had been completely underestimated with heartbreaking results.

Home based intervention also eliminated the potential for the child to pick up inappropriate behaviors that he might observe from other children with autism.

All children in the UCLA Young Autism Project were under 4 years of age when they began the program. Lovaas speculated that younger children would maintain and

generalize skills more easily than older children because they would be less likely to discriminate between environments . . . it was assumed it would be easier to successfully mainstream . . . into preschool than to attempt mainstreaming an older child into a higher grade (Lovaas, 1987).

Parents and other adults in the child's community were trained to be actively involved in teaching the children. This way the parents could maintain and support the gains made by the children. All significant persons in the child's environments were actively included in the program. The parents and teachers became the primary service providers with the professionals acting as consultants, staff trainers and program coordinators.

Tasks (behavior goals) were broken down into small steps. Each "discrete trial" was taught separately using highly motivating reinforcers as consequences for correct responses (behaviors).

One of the goals was to make it more fun to tune in and learn than to spin or flap or wander around the room . . . the two fundamental goals, apart from the content being taught, are to make the child want to learn and to make the child feel that he can learn. That he is competent. (Lovaas quoted in Johnson, 1994.)

Intensive data keeping was used to document the effects of the treatment on a daily basis.

Another important contribution of the UCLA Early Autism Project was the publication of a comprehensive, reader friendly 'how to" book for parents and teachers that covers all the tools of behavioral teaching. It gives clear examples of their application, both in one-on-one teaching sessions and in the general life environment of the child. Teaching Developmentally Disabled Children, also called The ME Book, was published in 1981 (PRO-ED, Inc) and continues to give parents, teachers and other interested adults clear instruction on how to establish appropriate and functional skills and responses and eliminate problem behaviors. A series of video training tapes is also available.

The groundbreaking research by Lovaas and his colleagues raises several intriguing possibilities. First, it suggests that intensive teaching that requires young children with autism to engage actively with their physical and social environments and provides them with consistent, differential consequences can result in completely normal functioning for many . . . Second, that intensive behavioral intervention produces substantially better outcomes than other treatments for young children with autism (Green, 1996:32).

But what about those children in the UCLA project that did not achieve normal functioning? What variables could explain why some children developed to become indistinguishable from their typical peers and some did not? This might suggest a different etiology for these children or perhaps a different learning style than was addressed in the original language based program. Lovaas and others believe these children might be visual learners (Johnson, 1994). Efforts to refine and expand on the concepts learned thus far continue. Some programs, such as the Picture Exchange Communication System (PECS) or the Princeton Child Developmente Institute (PCDI), utilize more visual approaches to teaching communication. These techniques have been built up by many behaviorists working with many children over many years’ time. It is a constantly developing system. (Lovaas quoted in Johnson, 1994)

Rob Leaf and John McEachin studied under Lovaas and were directly involved with the UCLA Young Autism Project for about twelve years beginning in the mid-70s. Both received extensive training which lead to their doctorates under the direction of Dr. Lovaas at the UCLA early intervention clinic. Subsequently, their work included servicing people with autism of all ages in various settings. Their model of treatment is based heavily on the UCLA model. It is also influenced by their experiences with clients beyond the preschool level. They have drawn from the work of other professionals and from their collaborations with special education and language professionals, paraprofessional teachers and parents.

The primary teaching method is one-to-one discrete trials. Great emphasis is placed on creating an intensive environment in which the child is continually learning, in the same way that a typically developing child constantly learns from his environment. Theirs is a treatment model with an average minimum duration of 2 years. A minimum of 30 hours of direct instruction is recommended (which includes structured play times and breaks). A high level of structure beyond the therapy time is required as part of the program as is consistency of behavioral approaches across all settings . . . the child’s entire day becomes part of the therapy process and the parents become an integral part of the team (Leaf & McEachin, 1999:11). In this way, the child spends most of his time reacting to and interacting with his environment (as typical children do) rather than engaging in preoccupation with self and with self-stimulatory behaviors.

Intervention begins with a month or so of establishing a positive relationship between the child and the therapists. During this time reinforcers are identified and established and much of the time is spent in play. Reinforcement is provided freely on a non-contingent basis so that the therapists and the teaching setting become reinforcing to the child. Gradually discrete trials are introduced and this direct DTT time increases steadily, with a highly structured use of contingent reinforcers, until it is the primary method of therapy. In the later stages of treatment the time spent in DTT decreases as more time is spent with incidental teaching and group training.

The stages of therapy are based on the following general guidelines:

Beginning Stage - "Learning to learn"

• Sitting
• Attending
• Compliance
• Remaining on task
• How to process feedback
• Understanding cause and effect

Middle Stage

• Communication skills
• Play skills
• Self-help
• Social skills

Advanced Stages

• Subtle social skills
• Higher level play skills
• Advanced cognitive and communication skills
• Integration of skills to everyday environments (school)

A strong program emphasis is placed on treatment of disruptive behaviors that takes more patience and skill than teaching something as complex as language (Leaf & McEachin, 1999). Disruptive behaviors are often the primary barriers to the learning process and integration with typical environments. Other detrimental behaviors that require intensive treatment are inattention, lack of participation, isolating and being off task.

Intervention occurs throughout the entire day and across all settings. The base location of the program can be center or home but home based appears to have specific advantages for young children, such as arranging therapy around nap times to enhance its effectiveness.

Autism Partnership has given special attention to the unique needs of older children with autism. Adjustments to the general model are made in consideration of previous efforts and the need for age appropriate skills. Older children often have a higher degree of disruptive behaviors that interfere with learning and that must be addressed in their program.

The typical treatment team consists of 2-5 tutors who are trained in discrete trial teaching, parents who are trained to systematically integrate teaching goals into everyday family life, a program supervisor to provide case management, team supervision and parent training 2-3 hours per week, and often a clinical supervisor to provide general supervision of the child’s program.

Discrete trial teaching is a very sophisticated teaching method and involves complex application of reinforcement and prompting techniques. Effective behavior reduction treatment also requires a high level of training. A highly qualified and experienced professional should provide program supervision. The level of supervision required depends upon several variables, including the experience and skill level of the team, the number of treatment hours and the complexity of the programs.

Leaf and McEachin present a comprehensive, reader friendly training program in A Work In Progress. In it they effectively and clearly explain behavior intervention concepts and techniques and provide essential information for parents and others who wish to use intensive behavioral therapy for their children. Important attention is given to two very critical components of effective ABA therapy, reinforcement and problem behaviors, along with detailed instruction on delivering discrete trials. The Autism Partnership Curriculum for Discrete Trial Teaching with Autistic Children, an extensive and thorough collection of systematically introduced teaching programs is also included. Examples of data collection, charts and other significant information are also provided to assist those who wish to set up a program based on the Autism Partnership model.

Additionally, Autism Partnership offers on-site training workshops for families, teachers and therapists. Program supervision and case management services are offered to families in Connecticut and other areas and include development, refinement and evaluation of ABA programs. School consultation services are also available.

The Eden Institute began as a private school for children with autism in 1975. Before beginning Eden, David Holmes had been working at Princeton Child Development Institute, an ABA school for children with autism. The influence of the work of Lovaas and aspects of his work at PCDI were carried over to the philosophy and systems employed at Eden.

Eden’s mission is to provide a comprehensive continuum of services to people with autism. Eden currently provides the following collection of services to children, adults and families with autism. Each is designed to enable children and adults to live independent lives within their communities to the full extent of their abilities:

Discrete trial training is used but less intensely than with the UCLA and Autism Partnership examples. Eden uses discrete trials training to "jump start" skill acquisition. Children at the Eden Institute receive an average of 20 hours a week in direct DTT instruction. This is supplemented with "more global" ABA techniques such as incidental and group training. (When consulting to school districts, Eden generally recommends about 2 hours of discrete trials each day, according to Eden consultant Michele Brooks.)

Visual aides are sometimes used in programming. Augmentative communication system preference is for voice rather than visual systems. Neither PECS nor American Sign Language is supported within the Eden model.

Teaching ranges from one-on-one to small group sessions with opportunities for integration when appropriate.

Teaching goals are divided into 5 different programs at Eden Institute. Children progress through each program based on their age and skill level. The 5 programs are:

Transition Program

• small group instruction
• independent work
• academics
• language
• pre-vocational

Sundberg and Partington base their approach directly on B.F. Skinner’s analysis of verbal behavior that was published in 1957. Skinner believed that language is a result of operant conditioning, similar to other learned behavior. Children learn to speak through selection by consequences (operant conditioning). But verbal behavior, according to Skinner, requires a separate analysis. With verbal behavior, the child does not operate directly on his environment. The behavior of others in a verbal community is an additional event that must be considered (Skinner, 1957 cited on B.F. Skinner Foundation Web Site).

Verbal behavior (VB) differs from other ABA models in it that it is focused primarily on Skinner’s analysis of language as a learned skill. Skinner proposed that language is behavior that is primarily caused by environmental variables such as reinforcement, motivation, extinction and punishment. This view of language differs substantially from others that assume language is primarily caused by cognitive or biological variables (Sundberg & Partington, 1998).

With the VB model, language (i.e. verbal behavior) is analyzed into formal and functional properties. Formal properties include grammar, syntax, articulation, pitch, etc. Functional properties of language are the stimulus for the behavior, the motivation for it and the consequences that control the response. Using these concepts, language dysfunction can be targeted very specifically and programs are designed to teach verbal behavior in its functional context. VB is an intensive treatment model that attempts to develop language as rapidly as possible using highly structured direct teaching and intensive natural environment teaching.

In the current edition of their verbal behavior training manual, Teaching Language to Children . . . Sundberg and Partington stress the need for both intensive, one to one discrete trial teaching (DTT) combined with natural environment training (NET), which focuses on the child's immediate interests as a guide for language instruction. The advantages and disadvantages of each method are discussed at length in their manual (pages 255-271). Sundberg and Partington conclude that the balance between DTT and NET may change frequently during the language acquisition process, but training should always include both approaches (Sundberg and Partington, 1998: 270).

In general, DTT is recommended as the primary instructional method for:

• academic activities
• specific skill development
• structured learning of academic skills in later elementary classrooms

NET is considered more effective for teaching:

• early requesting
• compliance
• instructional control
• pairing
• observational learning

A combination of both is most effective when teaching:

• making requests
• labeling
• receptive language
• motor imitation
• verbal imitation
• early conversational skills

Sundberg and Parting ton have developed a comprehensive language assessment to identify the deficits that are present in the child and to guide the development of an appropriate language intervention program for the child. The Assessment of Basic Language and Learning Skills (ABLLS) is also a valuable tool for determining IEP (Individualized Educational Placement) objectives. It is designed to be used by parents and teachers. With it, key developmental skills are evaluated under several general headings.

(Sundberg and Partington, 1998)

The ABLLS separates specific skills into various phases of development. Parents or teachers can assess the present level of the child at each stage. Goals and curriculum are developed to emphasize these unaccomplished steps. They are then taught systematically using basic behavioral techniques in both DTT and NET settings.

Sundberg and Partington propose that the majority of negative behavior issues, tantrums, social withdrawal, self-stimulation, etc. are directly related to dysfunctional language. They suggest that eliminating an undesired behavior that is linked to a defective verbal repertoire without considering the language link is an example of the mistake of treating the symptom and not the cause. Language training is an essential part of behavior reduction programs (Sundberg & Partington, 1998).

Discrete trials in verbal behavior are structured differently than in typical DTT models.

Sessions are less scripted and there is less differentiation between teaching programs than in other DTT models. The overall program is highly structured but the treatment sessions are not. Typically, data is not taken during teaching sessions, so the teacher can spend more time engaging the child. Sessions should be creative and flexible while targeting curriculum programs. A high level of positive reinforcement is maintained throughout (errorless) teaching. High priority is placed on making learning fun.

Other differences include the pace of instruction, concept integration (i.e. not adhering to sets of trials), and the use of receptive language training as a direct bridge to expressive language.

While speech is the primary goal, those children who do not learn to speak must have an alternative communication system. Sundberg & Partington provide a lengthy discussion of augmentative communication systems and conclude that American Sign Language is the preferred method with PECS a close second

VB training is a very precise and sophisticated model of intervention. A high level of training and ongoing support and supervision through qualified case managers is required for people working with children in VB programs. The team should be trained in ABA and VB techniques. They should also be trained in augmentative communication systems when appropriate.

Parents are active members of the treatment team, facilitating development and use of skills in home and community settings. Parents need to know what the child is currently learning. Additionally, parents should be trained in the teaching skills they wish to know.

The set of manuals by Sundberg and Partington, Teaching Language to Children with Autism Or Other Developmental Disabilities and The Assessment of Basic Language and Learning Skills present materials and training to conduct a comprehensive language assessment of the child, used to direct an individualized curriculum for the child. A language curriculum and teaching techniques are provided along with data systems and tracking forms. Guidelines to assist with program selection (home or school-based) are also included, as are options for augmentative communication systems.

ABA approaches are consistent in their use of operant conditioning principles and techniques (reinforcement, shaping, prompting, chaining, behavior extinction, etc.,).

In all cases, target skills are broken down into smaller units of information and practiced repeatedly. All place a high priority on data collection and qualified interpretation of the data to drive the program. Keeping the child "on task" is emphasized.

Differences between the models appear most obvious in the predominant teaching techniques used - intensive one-on-one discrete trial teaching (UCLA, Autism Partnership), more natural environment training (PECS) or a combination of both environments as instructional vehicles (VB, Eden). Significant differences also exist between the overall intensity of the methods used, the level of team training and support required or recommended by each model, the hierarchy of program stages, behavior reduction programs and philosophies, and the role of the family.

Perhaps ABA models can be separated into two broad categories: ABA treatment models, and ABA educational models, with differences in approach that each view inherently incorporates. Those examples that focus on intensive, short term (2-3 years) intervention can be placed in the former category, with the objective of bringing the child into a more typical pattern of development. Others are more keenly focused on long-term methods of adaptation around the disability to educate the child and foster supported independence.

Contemplating the different approaches to ABA brings intriguing questions to mind. Is the key in keeping the child engaged in his surroundings, "on task" for the majority of the time? Is the intensity of the focus and attention required related to outcome? Do the various approaches affect brain development? If so, what are the components required for optimal long-term gains? It will be interesting to see what ongoing and future scientific research tell us about the underlying relationship between behavioral and neurodevelopmental issues and effective treatment for children and others on the autism spectrum.

Ashcraft, M. H., (1998). Fundamentals of Cognition, New York: Addison Wesley Longman, Inc.

Bondy, A. & Frost, L. (1994) The Picture Exchange Communication System. Focus on Autistic Behaviior, 9:1 – 19

Berk, L., (1996). Infants, Children and Adolescents, Boston: Allyn & Bacon

Frost, L. & Bondy, A., (1994). Picture Exchange Communication System Training Manual. Cherry Hill, NJ: Pyramid Educational Consultants, Inc.

Frost, L. & Bondy, A., (1999). Paper for Autism 99 Conference

Green, G.,(1996). Evaluating Claims About Treatment for Autism. In Maurice et al. (Ed.) Behavioral Intervention for Young Children With Autism (pp. 29-43), Austin, TX : Pro-Ed Inc.

Holmes, David L., (1997). Autism Through the Life Span :The Eden Model, Bethesda, MD: Woodbine House

Johnson, C., (1994). Interview With Ivar Lovaas, The Advocate (Autism Society of America), Nov-Dec 1994

Leaf, R., and McEachin, J., (1999). A Work In Progress: Behavior Management Strategies and a Curriculum for Intensive Behavioral Treatment of Autism, New York: DRL Books

Lovaas, O.I., (1981). Teaching Developmentally Disabled Children: The Me Book. Baltimore : University Park

Lovaas, O. I., (1996). The UCLA Young Autism Model of Service Delivery. In Maurice et al. (Ed.) Behavioral Intervention for Young Children With Autism (pp. 241-248), Austin, TX : Pro-Ed Inc.

Lovaas, O.I., (1987), Behavioral Treatment and Normal Educational and Intellectual Functioning in Young Autistic Children. Journal of Consulting and Clinical Psychology, 55, 3-9

Lovaas, O.I., & Smith, T., & McEachin, J.J. (1989) Clarifying Comments on the Young Autism Study. Journal of Consulting and Clinical Psychology, 57, 165-167

McEachin, J.J., Smith, T., & Lovaas, O.I. (1993) Long Term Outcome for Children With Autism Who Received Early Intensive Behavioral Treatment. American Journal on Mental Retardation, 4: 359-372

McClannahan, L. E., and Krantz, P. J., (1999). Topics In Autism: Activity Schedules for Children With Autism: Teaching Independent Behavior, Bethesda, MD : Woodbine House

Satcher, D. (1999). Mental Health: A Report From the Surgeon General. [http://www.surgeongeneral.gov/library/mentalhealth/chapter3/sec6.html#autism]

Skinner, B.F. (date unknown) A Brief Survey of Operant Behavior. B.F. Skinner Foundation Web Site [http://www.bfskinner.org/Operant.asp]

Smith, T., (1999). Outcome of Early Intervention for Children With Autism. Clinical Psychology and Practice, 6, 33-49.

Sundberg, M, & Partington, J, (1998). Teaching Language to Children with Autism or Other Developmental Disabilities, Danville, CA: Behavior Analysts

Sundberg, M, & Partington, J, (1998). The Assessment of Basic Language and Learning Skills, Danville, CA: Behavior Analysts