Schizotypy's shape

Abstract

Dimensional models of schizotypy and associated traits have taken on current relevance in the DSM-5 (http://www.dsm5.org) proposal for personality disorder (PD), which includes a personality trait initially conceptualized as a five-facet schizotypy domain and then simplified into a three-facet psychoticism domain that has yet to be evaluated extensively. In this study, I (1) reviewed the literature to develop hypotheses about the content and boundaries of the schizotypy domain, and (2) measured this content in a mixed sample of students and patents with 657 usable protocols at Time 1 (193 Notre Dame students, 301 University of Iowa students; 163 outpatients) and 263 usable protocols (74 Notre Dame students, 76 University of Iowa students, 113 outpatients) at Time 2, at least 1.5 weeks later. I then (3) evaluated confirmatory models including DSM-5 schizotypy and psychoticism facet models and other empirically grounded models and (4) used the best confirmatory model to provide item pools for classically constructing scales of schizotypy facets. This four-factor structure provided content pools used to create four corresponding scales: Unusual Perceptions, Unusual Beliefs, Dissociation Proneness, and Cognitive and Communicative Peculiarity. Additionally, (5) I used item response theory (IRT)-based analyses to evaluate items in these facet scales, both in terms of the level of schizotypy they best measure and the strength of their relations to the schizotypy construct. I also (6) examined the short-term test-retest reliability of the schizotypy scales, as well as that of the established measures used in this study; new and existing measures were comparably stable. Finally, (7) I evaluated schizotypy's convergent and discriminant validity in relation to three other types of traits: (a) those correlated with the domain (e.g., Obsessive Compulsive Disorder [OCD] and non-delusional mistrust), (b) other higher level traits (i.e., measures of the 3-factor and 5factor models of higher order personality/temperament), and (c) familially related traits (e.g., social anxiety). Overall, the schizotypy facet measures appeared to assess moderate