IS BEHAVIORISM BECOMING A PSEUDO-SCIENCE?: POWER VERSUS SCIENTIFIC RATIONALITY IN THE ECLIPSE OF TOKEN ECONOMIES BY BIOLOGICAL PSYCHIATRY IN THE TREATMENT OF SCHIZOPHRENIA

Behavior and Social Issues, Fall 2006 by Wakefield, Jerome C

THE PROBLEM OF GENERALIZATION

Wong traces the rise, first, of individualized operant programs, and then token economies for entire wards. The research left no question that patient behavior can be brought significantly under control using behavioral contingencies. True, token effectiveness is by no means universal; some percentage does not respond across studies (Kazdin, 1983). Wong et al. (1987) found that independent recreational activity reduced stereotypic vocalizations, but "that contingent tokens given to one subject for on-task behavior did not contribute significantly to treatment impact" (p. 81), or, as the abstract states, "results were the same with or without contingent tokens" (p. 77), a result mentioned but oddly not pondered in the discussion.

Although Wong's characterization of the success of these approaches is on the optimistic side, the fact is that the data were overall quite positive. So, reasonably, Wong asks: "Why did behavior analysis and related behavioral approaches to treating this disorder fail to develop and thrive? Why did biomedical interventions, particularly psychotropic medication, become the prevailing treatment for severe mental disorders? Is this prevalence based on sound scientific research and technology or something else?"

There is a simple and quite scientific reason why token economies were set aside in favor of drugs; lack of generalization. Drugs could be prescribed for the patient and had roughly the same effect whether inside the institution or out, and outside the institution in principle they can be administered with minimal additional supervision. Behavioral methods depended on continued control over patients of the sort that occurs on a ward, and those environments became rare and some control procedures came to be considered unethical except under exceptional circumstances. In the community, reinforcers cannot be easily controlled, and what control can be achieved is expensive to maintain indefinitely.

That generalization is the issue about token economies as applied to most patients has gradually been realized (Glynn et al., 2002; Kazdin, 1982; Kazdin & Bootzin, 1972; Stokes & Baer, 1977). In her much-cited review, Glynn (1990) stated the obvious but often ignored truth: "One of the major challenges for behavior therapy in general (Stokes & Baer, 1977) and token economies in particular (Kazdin, 1982, 1985) is the maintenance and generalization of effects" (p. 387); indeed, "Maintenance and generalization of treatment gains are, of course, critical tests of the utility of treatment interventions" (p. 401). The problem is that "the power of the token economy rests in large part on control of external environmental contingencies," and as the patient moves from the controlled clinic environment to the community and thus can obtain desired rewards through alternative means and avoid punishments thus voluntarily exiting from the token economy, "the treatment modality becomes less useful" (p. 387). Glynn further noted that "the crucial question is, Are token economies primarily prosthetic, resulting in behavioral change only when it is supported by token reinforcement; or are they therapeutic, resulting in long-term behavioral change across settings?" (p. 401), and observed that "The answer to this question has not been fully determined" (p. 401). Nothing has happened since 1990 to change that sober assessment. Glynn does refer to various factors suggested in the literature that might attempt to address generalization, but these techniques have not been demonstrated to work to maintain gains in open environments after token reinforcement programs for schizophrenics end.