This study was concerned with psychotherapists' evaluations of the outcome of therapy. Staff and resident psychiatrists employed in 23-item questionnaire to rate the seccess of psychotherapy with 85 of their inpatients. These were patients for whom psychotherapy constituted a significant part of the treatment that they received in the hospital. The therapists' responses to the questionnaire items were intercorrelated. The results indicated that a major aspect of a psychotherapist's judgment of the success of treatment-as these judgments usually are employed in current psychotherapy research-is the therapist's affective reaction to the patient, Better-liked patients were viewed as having improved more. However, when the therapists' responses were subjected to factor analysis, independent Improvement and Affection factors emerged. It is suggested that when therapists' ratings of success are used in psychotherapy research they should be treated or refined statistically to minimize potential confounding with their affection for patients.