Inferential reasoning during the psychodiagnostic assessment : attribution, hypothesis-testing strategies, and final inferences as a function of theoretical orientation, level of experience, and temporal order

Psychological problem representation, a complex task, is underpinned by clinicians' inferential processes, processes which are not immune to bias and logical error. Problem representation influenced by faulty reasoning can have deleterious effects on treatment planning and consequently on treatment outcome. Using a think-aloud procedure, 32 clinical and counselling psychologists examined a case file of the initial interview. The following three variables thought to contribute to clinicians' conceptualization of clients' problems were investigated: (1) theoretical orientation, (2) level of experience, and (3) temporal order of client information. The reasoning processes of psychodynamically-oriented and behaviourally-oriented clinicians were studied. The existence of expertise in psychotherapy was examined. Comparison of hypothesis generation among novices and experts was made. The impact of the temporal order of client information on the inferential processes of clinicians was also investigated. Subjects processed one of two versions of the case file. Levels of theoretical orientation, experience, and temporal order were compared in order to determine (1) whether clinicians posit significantly more dispositional hypotheses than contextual hypotheses, (2) whether initial hypotheses of a dispositional or contextual nature are confirmed or disconfirmed, and (3) whether initial hypotheses of a dispositional or contextual nature are retained or rejected. Based on absolute numbers, psychodynamicists posited more contextual inferences than behaviourists. Novices confirmed more dispositional inferences than experts. Clinicians rarely disconfirmed or rejected their initial hypotheses. Behaviourists retained more dispositional inferences than psychodynamicists. The order of the case material significantly affected the types of hypotheses generated and the hypothesis-testing strategies utilized. No significant differences, however, were found in the proportion of dispositi