No SOONER were computers able to handle the data than did psychologists and psychiatrists interested in establishing a more objective and reliable classification of psychiatric patients begin to apply empirical classification methods to psychiatric phenomena. Methods of factor analysis, cluster analysis, Q-sorts and a variety of other techniques have been used. The objective has been to identify the distinct patterns that occur repeatedly in nature in highly similar form. From a simple classification point of view, such distinct homogeneous types should provide the basis for most reliable segregation of patients. Some investigators have reasoned further that syndromes, which are recognized by the coexistence of particular unique patterns, are often found to be associated with distinct disease processes. Others have been more pragmatic, being concerned only with objective, reliable, and useful descriptive classification. WITTENBORN et al.,’ were perhaps first to report extensive multivariate analyses of the phenomena of psychopathology in 1953. BECKY followed the next year with an application of Q-sort methodology that resulted in description of six distinct types of schizophrenia, and GRINKER et al.,3 followed shortly with the description of four sub-types of depression (which bear considerable similarity to those identified in our own work). LORR et al.,4 undertook extensive cluster analyses of symptom and behavior rating profiles for large samples of patients from the general inpatient psychiatric population and identified six major phenomenological types. It is not possible, in this brief introduction, to analyze in detail the similarities and differences in classification concepts that might be derived from these and other empirical investigations. We would note, however, that given certain tolerance for minor discrepancies (due probably to differences in observational techniques, data bases, and methods of analysis), a substantial thread of consistency emerges. Our own work, which was the direct basis for the investigation reported here, is in many particulars not inconsistent with the empirical results obtained by others using different data bases. The primary question left unresolved by all of the empirical classification research concerns whether the typologies have clinical meaning. One aspect of this question involves the extent to which the empirically derived types “ring true” in clinical experience. It is our impression that previous investigators have stopped short of considering this important question.
Read full abstract