Psychological Medicine

Original Articles

Contrasting prototypes and dimensions in the classification of personality pathology: evidence that dimensions, but not prototypes, are robust

N. R. Eatona1 c1, R. F. Kruegera1, S. C. Southa2, L. J. Simmsa3 and L. A. Clarka4

a1 University of Minnesota, Minneapolis, MN, USA

a2 Purdue University, West Lafayette, IN, USA

a3 University at Buffalo, The State University of New York, Buffalo, NY, USA

a4 University of Notre Dame, Notre Dame, IN, USA


Background DSM-5 may mark the shift from a categorical classification of personality pathology to a dimensional system. Although dimensional and categorical conceptualizations of personality pathology are often viewed as competing, it is possible to develop categories (prototypes) from combinations of dimensions. Robust prototypes could bridge dimensions and categories within a single classification system.

Method To explore prototype structure and robustness, we used finite mixture modeling to identify empirically derived personality pathology prototypes within a large sample (n=8690) of individuals from four settings (clinical, college, community, and military), assessed using a dimensional measure of normal and abnormal personality traits, the Schedule for Nonadaptive and Adaptive Personality (SNAP). We then examined patterns of convergent and discriminant external validity for prototypes. Finally, we investigated the robustness of the dimensional structure of personality pathology.

Results The resulting prototypes were meaningful (externally valid) but non-robust (sample dependent). By contrast, factor analysis revealed that the dimensional structures underlying specific traits were highly robust across samples.

Conclusions We interpret these results as further evidence of the fundamentally dimensional nature of an empirically based classification of personality pathology.

(Received April 20 2010)

(Revised July 30 2010)

(Accepted August 05 2010)

(Online publication September 22 2010)