STATEMENTS MADE BY A THERAPY CLIENT WERE TYPED ON SMALL CARDS AND PRESENTED AS AN ARRAY TO 18 SUBJECTS. EACH SUBJECT CATEGORIZED THESE STATEMENTS TO REPRESENT CONCEPTS HE USED IN CONCEPTUALIZING THE CLIENT. ON COMPLETING THIS TASK, EACH SUBJECT DESCRIBED THE CLIENT USING A Q-SORT. DESCRIPTIVE ACCURACY WAS DEFINED CONSENSUALLY VIA FACTOR ANALYSIS OF Q-SORT INTERCORRELATIONS. THE RESEARCH TASK FAILED TO DIFFERENTIATE BETWEEN FOUR HIGHLY ACCURATE SUBJECTS AND FOUR HIGHLY INACCURATE SUBJECTS IN TERMS OF (1) THE DIVERSITY OF THEIR CONCEPTUALIZATIONS, (2) THE PERMEABILITY OF THEIR CONCEPTS, (3) THE NUMBER OF CLIENT EVENTS INCORPORATED IN THEIR CONCEPTS, AND (4) THE ARRANGEMENT OF ITEMS IN CONCEPTS. IN ADDITION, THE THEORY ENDORSED AND THE LANGUAGE EMPLOYED DID NOT RELATE TO DESCRIPTIVE ACCURACY. IMPLICATIONS ARE THAT THE RESEARCH METHOD EMPLOYED REVEALS THE IDIOSYNCRATIC ASPECTS OF CLINICAL THINKING BUT LITTLE ABOUT THE LAWFUL PROCESSES WHICH ENABLE CLINICIANS TO CONVERGE AS THEY DID ON THE CRITERION. (AUTHOR)