Research in client-oriented psychotherapy on the effect of "accurate empathy," "non-possessive warmth," and "genuineness" in therapists relates these variables ("therapeutic conditions") to the outcome of treatment. This paper is a critical review of a number of such studies. Two types of experimental designs were used. In the first, patients treated by therapists who were rated high in the "therapeutic conditions" were compared with untreated control groups. In the second, a correlation design was used relating high and low levels of the "conditions" to outcome. Truax and Mitchell had earlier reviewed many of these same studies and had concluded that the therapists who were rated high in the three "therapeutic conditions" were more effective regardless of their training or theoretic orientation. The present author disagrees with this conclusion, finding