Exploring the structure of psychopathological symptoms: a re-analysis of AMDP data by robust nonmetric multidimensional scaling.

This paper investigates the structure of psychopathological symptoms. Based on AMDP symptom profiles, a symptom space was calculated by robust nonmetric multidimensional scaling (NMDS) and the symptom structures of a sample dating from 1980 and a sample from 2002/2003 were compared. The method of NMDS presented in this study allows results from other studies to be confirmed and complemented. The symptom factors identified in the past by factor-analytic studies were replicated as clusters in two-dimensional symptom maps. Additionally, some theoretically assumed clusters of symptoms were detected that were not found in previous factor analysis approaches. From the results, which are depicted in a continuous space, new insights can be gained, especially with regard to questions of categorical and dimensional classifications. The comparison of the structural aspects of the symptomatology across more than two decades resulted in only small divergences and allows conclusions to be drawn about the stability of these structures and consequently of the symptom clusters and dimensions.