[When does a study of different therapies allow comparisons of their relative efficacy?]

Psychotherapy research is a form of evaluation research. Among many others, two types of evaluation can be distinguished: "comparative" and "non-comparative" evaluation studies. This differentiation, though its consequences are often neglected, is fundamental, since a study designed to answer a question concerning the non-comparative or "isolated" efficacy of a therapy cannot simultaneously serve to answer the question concerning the relative efficacy of two or more therapies aiming at the same goals or objectives--and vice versa. In this article, a small part of psychotherapy research, the 22 studies which have already been used by Grawe et al. (1994) in their comparisons of behaviour therapies and short-term psychodynamic therapies, is reanalyzed with respect to the two types of evaluation and to some consequences of this differentiation.

On the whole, it is not possible to draw conclusions about the comparative efficacy of behaviour therapies and short-term psychodynamic therapies with regard to their specific goals, even if only some consequences of the distinction between two types of evaluation are taken into account. Among other reasons, this is due to the fact that the studies have not consequently been planned and executed as comparative evaluations. Only amelioration of the 22 studies can be regarded--with certain restrictions--as comparative outcome studies with respect to amelioration of certain symptoms. A further analysis of these studies shows that there is no evidence of a "highly significant" superiority of behaviour therapies over short-term psychodynamic therapies.