Three models of clinical judgment.

Abstract

3 simulation models were applied to clinical judgments of "psychotic" vs. "neurotic" for 29 judges across 7 samples of MMPI profiles. A crossvalidated multiple correlation was used as a goodness-of-fit measure for the 3 judgment models. On this basis, each of the 29 clinical judges was characterized as being either "linear" or "configural" in his utilization of MMPI test data, where configurality was operationally defined by the fit of the data to a model employing configural, or patterned, variables. It was shown that configurality was a consistent judgmental characteristic distinguishing 16 of the judges, but this characteristic was unrelated to both accuracy and amount of clinical training. Also, judges, as a group, tended to change their manner of judgment with changes in samples of MMPI profiles.

Consciousness, context, and know-how

“…Despite 50 years of analysis, even the most favorable assessments of clinical judgment using non-linear models can explain only 5% of the variance not accounted for in linear models (Faust, 1989;Ganzach, 2001;Meehl, 1954;Wiggins & Hoffman, 1968). Thus, clinical expert reports regarding their own know-how seem to falsely attribute highly non-linear judgment methods to their performance of clinical tasks.…”

Individual Differences in Diagnostic Judgments of Psychosis and Neurosis from the MMPI

“…Except for the variable &dquo;amount of clinical training&dquo; (which was unrelated to any of the factors), Horn and Stewart had no variables by which to identify their subject factors. This is particularly unfortunate in light of the considerable body of research on these same 29 clinicians (Goldberg, 1965(Goldberg, , 1968(Goldberg, , 1969(Goldberg, , 1970Wiggins and Hoffman, 1968).…”

The Roles of Scientific and Clinical Epistemologies in Forensic Mental Health Assessments

“…During interviews, clinicians are confronted with a great deal of complex and continuously changing qualitative observations that reveal an individual's internal world. Clinicians must continuously adjust the meaning and weight given to different parts of this data at different times in order to understand the unique person before them (Sines 1959;Wiggins and Hoffman 1968). Ruscio (2003) argues that the vast number of possible combinations created by so many separate pieces of data makes their clinical integration impossible and therefore inherently untrustworthy.…”