The articles in this special section on "Paradigms for Psychotherapy Outcome Research" are filled with useful ideas, but the major risk in this discussion is that they talk past each other, rather than to each other. This risk exists, in large part, if we mistakenly assume that one label, "treatment outcome research," refers to a single thing for which we need a single paradigm rather than to a family of activities that may have different goals and utilities. It is still incumbent upon us, however, to try to identify the most useful members of this family rather than shrugging and insisting that everything is equal under the sun, even though we are still at a stage in the development of the field where no single paradigm is likely to be (or should be) ascendant.
In providing an overview of the special section, our first step is to articulate as clearly as possible the paradigms that are being suggested and the historical context in which they have arisen. Since the late 1970s, a major influence on psychosocial treatment outcome research has been the regulatory model of the Food and Drug Administration (FDA). Within this model, the goal for any treatment is to establish evidence of its safety and statistically significant indications of its efficacy in at least two separate clinical trials. The focus is on internal validity, with relatively little consideration of clinical significance (i.e., there are no requirements for documenting a minimum effect size or proportion of positive responders). Relatively little attention is paid to the total weight of the evidence (i.e., it does not matter much how many negative studies there are so long as two positive ones can be produced). Meta-analytic accumulation of discrete efficacy trials becomes the extant strategy for treatment research and further treatment development, although this trend has provoked considerable criticism.