A recent paper of Shinohara et al. [1] systematicly reviewed the effectiveness of behavioral therapies (BT) for treating depression in comparison to other psychological therapies (e.g. psychodynamic and ‘third wave’ cognitive and behavioral therapies). From the 25 studies which attended the inclusion criteria, taking into account the purposes of the review, poor conclusions were drawn. According to the authors, most of the studies were provided with small samples and were at unclear/high risk of bias. There were no evidences for differences in respect to response to the treatment or dropouts rates. Besides the small samples and the risk of bias, the quality of the evidence was weak because in most of the articles there was, for example, a lack of information about treatment fidelity (i.e. “How the investigators can assure that the adopted procedures were those intended?”) or about beneficial (improvements in quality of life, social functioning etc.) or adverse (worsening of symptoms, attempted or completed suicides, etc.) ‘secondary’ outcomes.

Many enthusiasts of BT would find these results as quite disappointing. Behavioral therapy has profound roots in mid 20th century operant and respondent psychology. It is based on the Behaviorism, which has an agenda of denial to ‘mentalism’ and of building knowledge over single-subject designs. The main characteristic of these designs is that the subject is used as his/her own control [2]. Behaviorism, due to its particular methods and interests, gave birth to some specialized journals. Thus, many psychologists committed with this approach selectively read articles from these journals. Most of the citations are indeed to articles from the same journals. This cycle of reading and citing articles from the same journals inadvertedly makes the behaviorist community a scientific ghetto (a situation apparently seen as unavoidable by Skinner [3].

The importance attributed to single-subject designs in BT literature may explain the scarce available literature for the above cited review. Thus, while the superiority of BT over other methods remains to be demonstrated, the way out of this situation will not be found with single-subject designs. Suggestions for the validation of BT can in most of the cases be summarized by the phrase “Test the therapy in the same way as a pill should be tested”. Some steps in this direction are as follows: (1) number is (statistical) power: Comparisons among groups in which BT is compared against a control condition, other psychotherapies and pharmacotherapy can be heuristic. Increasing the number of participants in each group improves the statistical power of the conclusions. (2) validated tools: for the aim of comparisons, both within the study and between studies, the use of validated tools (scales, questionnaires etc.) is mandatory. They help in the evaluation not only of the efficacy, but also of secondary outcomes. (3) treatment fidelity: To guarantee that the therapist will follow the planed procedure, the use of a manual is desirable. Video or audio recording of the sessions is important for monitoring the procedures. To these steps, many more can likely be added. These and other suggestions in the same direction would put BT in a ground where it could be compared to other therapies.