Programs based on applied behavior analysis (ABA) have become increasingly popular as interventions for childhood autism. A number of leading behavior analysts even have claimed that many children with autism can be "cured" through such programs and therefore strongly recommend ABA over alternative interventions. The extant research literature, however, does not support these claims. ABA programs for childhood autism are indeed promising, but exaggerated claims may undermine confidence in the approach, and are misleading to families of autistic children and to other stakeholders.
Empiricism has always been one of the key defining features of applied behavior analysis (ABA) (Cooper, Heron & Heward, 1987). Both theoretical and technological developments are closely tied to objective data (Zuriff, 1985). In keeping with the high value placed on pragmatism, the link between basic behavioral principles and the technological developments that stem from these principles is emphasized. Technological developments such as interventions for psychological disorders and behavioral problems are not relegated to an afterthought, but are central to the entire enterprise.
The recent movement toward empirically supported treatments (ESTs) in the field of clinical psychology therefore represents nothing new to applied behavior analysts. Although debates are currently raging among clinical psychologists and other mental health professionals regarding the relevance of outcome research to clinical practice, the idea that interventions should be informed by data is axiomatic to the behavior analyst. Applied behavior analysts may join other empirically-minded professionals in quibbling over details (e.g., the specific criteria that should be used to justify claims of empirical support for an intervention; cf. Herbert, 2000; Lohr, DeMaio, & McGlynn, in press), but the idea that "the data matter" is not in dispute.
ABA AND CHILDHOOD AUTISM
The conventional wisdom within the field of ABA is that one of the most significant if not the single most significant - empirically supported success story is ABA-based early intervention for developmental disabilities, and childhood autism in particular. In fact, the association of ABA with intervention programs for childhood autism has become so strong that Hayes (2001) recently warned that ABA "is gradually becoming a subfield of developmental disabilities" (p. 61). Indeed, the Internet home page of the Cambridge Center for Behavioral Studies (www.behavior.org)prominently features ABA programs for autism. The premier ABA journal, the Journal of Applied Behavior Analysis, frequently publishes articles on the application of ABA for developmental disabilities. In a feature article on Autism, Newsweek magazine described ABA as "the standard approach" to intervention with autism (Cowley, July 31, 2000, p. 52). The ABC News program Nightline recently featured ABA for autism (March 9, 2001). The New York State Department of Health (1999) recently issued clinical practice guidelines for autism in which ABA is recommended as a critical element of any intervention program for childhood autism.
The astounding success of ABA early intervention programs for children with autism is proclaimed not only in the professional literature, but also in books and Internet sites targeted to parents and other stakeholders. These publications frequently announce that ABA programs can result in dramatic developmental gains for many autistic individuals. In fact, it is claimed that many of these children can eventually function in normal education settings, and are indistinguishable from their normally developing peers. In other words, many can be "cured" of their disorder. Consider, for example, the following passages gleaned from the literature on ABA for autism:
Several studies have now shown that one treatment
approach--early, intensive instruction using the methods of Applied
Behavior Analysis--can result in dramatic improvements for children
with autism: successful integration in regular schools for many,
completely normal functioning for some (Green, 1996b, p. …