Applied Behavior Analysis

Applied Behavior Analysis (ABA) refers to labor-intensive, one-on-one educational interventions with young children with autism. Its hallmark is breaking communication and interaction into their behavioral components in order to teach them systematically, one small behavioral task at a time. Many programs use highly structured environments, saturated with rewards and occasionally punishments, to teach the linguistic and social skills that most children learn simply by listening, watching, and imitating. ABA is widely accepted today in special education as well as in clinical settings, in part because its procedures of careful documentation and measurement fulfill the mandate that interventions be “evidence-based.” It originated in the research of Charles Ferster, Marian DeMyer, Ivar Lovaas, and other learning theorists and behaviorists. During the 1960s, their work with autistic children was unconventional and controversial. It attracted both admiration and criticism. 

Reinforcement was the principle at the foundation of all behaviorist approaches and it informed all of its teaching techniques. It was applied to even the most severely affected non-verbal and self-injurious autistic children. The apparent indifference of these children to their social environments posed enormous challenges to any kind of human connection, let alone learning.

Researchers who championed behaviorism did not care why autistic children lived in worlds of their own or what caused their autistic symptoms. They approached learning empirically and placed their trust in the experimental method. The pioneers of ABA argued that efforts to understand and shape human development required measuring and manipulating what could be directly observed: behavior. Everything else was beyond the reach of scientific research, including the family’s emotional climate, the history and quality of mother-child attachment, psychological factors located in the mind, and the brain’s biology. Philosophically, ABA therefore offered an alternative to the dueling theories of psychogenesis and biogenesis.

In the field of animal behavior, training techniques featuring reinforcement had been well established for decades. Ivan Pavlov’s famous experiments in the 1890s showed that dogs could be trained to associate the sound of a bell with food and thereby “conditioned” to salivate whenever a bell rang. B.F. Skinner worked with a variety of animals in the 1930s and 1940s to show that more complex behaviors could be shaped by deliberately varying the pattern of environmental responses, or “reinforcement schedules.” During World War II, Skinner even showed that this technique, which he called operant conditioning, could turn pigeons into reliable missile guidance systems after he trained them to recognize and peck at pictures of targets.

Skinner became a lightning rod for controversy when he turned his attention from pigeons to people. His utopian novel, Walden Two (1948) explored what a society based on behaviorist principles might look like. In Skinner’s fictional society, social arrangements were engineered by a committee of planners. The result was that people were peaceful and happy. They worked four hours daily and devoted the rest of their time to creative and leisure activities. Moral and religious objections to the novel surfaced quickly. How could a good society flourish without free will? Was there any place for spirituality if nothing fundamental separated humans from other animals? Questions like these were responsible for the reluctance to embrace behavioral techniques in the real worlds of education and medicine. Wouldn’t treating autistic children like rats and pigeons turn them into automatons and deprive them of their essential human dignity? How could an approach that circumvented thinking and feeling possibly work? What would it mean if it did?

Behaviorism was often accused of being mechanistic, soulless, and antithetical to freedom, but it was still very tempting to apply it to humans, especially in cases where difficult and dangerous behaviors had resisted other methods of change. Children with mental retardation were among the first subjects of reinforcement techniques. In 1948, psychologist Paul Fuller described these children as “vegetative human organisms” who were “behaviorally speaking, considerably lower in the scale than the majority of infra-human organisms used in conditioning experiments—dogs, rats, cats.” Adults with mental illness were typical subjects too. One article often cited as a landmark in the birth of ABA described how nurses in a large Canadian state hospital became “behavioral engineers” as they managed adults with schizophrenia according to behavioral principles. Such experiments demonstrated that even such “low” human beings could learn.

In 1960, Charles B. Ferster and Marian K. DeMyer were the first to conduct behavioral experiments with autistic children, who were often also labeled as mentally retarded. Their achievement was to show that institutionalized autistic children did respond to environmental reinforcements, albeit very slowly. Ferster was a psychologist at Indiana University who had worked in Harvard’s pigeon lab with B.F. Skinner and then with chimps in the Yerkes Laboratory of Primate Biology. DeMyer was a psychiatrist at Indiana University and Director of Children’s Services at the Larue D. Carter Memorial Hospital, a state hospital in Indianapolis.

Ferster and DeMyer put children in a room with a coin dispenser. After training them to press a key on the dispenser to obtain coins—much as Skinner had trained pigeons to peck for food—the children used them in vending machines holding desirable objects: candy, trinkets, picture viewers, phonographs, telephone sets, and color wheels. Left alone to make their choices, candy turned out to be the most popular, so Ferster and DeMyer used candy as reinforcement during the rest of the experiment. By dispensing candy at varied times and frequencies, they deliberately changed the children’s behavior. Their findings, summarized in detailed graphs and tables, “indicate at least the existence of normal processes at a very basic level.” If the same laws that regulated all human learning applied to autistic children, then it was theoretically possible to train those children to do (and not do) all kinds of things.

Other behaviorists were inspired to expand on this work. Dicky, a child diagnosed with childhood schizophrenia, was another subject of operant conditioning in the early 1960s. After a variety of drugs and physical restraints failed to alter his challenging behavior, Dicky was admitted to Washington State Hospital at the age of three. Three researchers at the University of Washington’s Institute of Child Development designed a behavioral program to reduce Dicky’s tantrums and help with his sleeping and eating problems. Dicky had also had a number of eye operations which required him to wear glasses, something he refused to do. Utilizing behavioral approaches, the researchers succeeded in modifying all of Dicky’s problem behaviors over a 3-month period. A more manageable Dicky was able to return home as “a new source of joy to the members of his family.” Thirty years later, Todd Risley, one of the researchers, visited with Dicky. He was living independently in Portland, Oregon, working as a part-time custodian in the apartment complex where he lived.

Of all the behaviorists interested in autism, Ivar Lovaas became the best known for his long-term studies of severely affected children. Lovaas did his graduate work at the University of Washington, a hub of behavioral research, and spent most of his career at UCLA. Under the auspices of the Young Autism Project, Lovaas and his team designed intensive interventions, frequently lasting for months or years, to advance the linguistic and social development of children with autism. They reported dramatic success. His first research subject was 9-year-old Beth, an institutionalized child with echolalia. Lovaas worked with her for a year, using positive reinforcement to teach her fifty new words and experimenting with electric shock to eliminate Beth’s self-injurious behavior. Lovaas moved on to work with Chuck and Billy, two non-verbal autistic boys, who were able to imitate 30 new words after 26 days of work, 7 hours each day, also using a combination of positive and negative reinforcements. Lovaas was careful to note that learning to vocalize and imitate words was one step. Using words and understanding their meanings was quite another.

Behavioral approaches had the distinction of being independent of all theories about the origin of autism. Advocates did not attempt to free autistic children from their condition for two reasons. First, behaviorists were skeptical that a unified disease called autism had been scientifically established, in spite of compelling clinical descriptions like those offered by Leo Kanner. Second, their philosophy allowed for no reservoir of individuality locked inside the bodies of children with autism, waiting to be liberated by the right cure.

For behaviorists, working with affected children was more like a construction project and constructing human beings entailed taking one tiny, practical step after another. If human status and belonging existed anywhere, they existed in responses to the social environment. In Ivar Lovaas’ words, that entrusted teachers, parents, and others with “the task of building a person where little had existed before.” This way of thinking provided a stark alternative to both the psychogenesis that dominated thinking about autism until the mid-1960s and the neuroscience that came later. ABA was founded on a paradigm that advocates believed offered optimism where others had offered only pessimism and failure.