Investigating the Best Structure to Conceptualize and Assess Multidimensional Schizotypy

Schizotypy offers a useful and unifying construct for understanding schizophrenia-spectrum psychopathology. Current conceptualizations and findings support a three-factor model of schizotypy consisting of positive, negative, and disorganized dimensions. However, recent studies have suggested four- and five-factor structures. The present study compared and examined the extent to which each factor model predicted interview and questionnaire outcome measures. Confirmatory factor analysis (CFA) and exploratory structural equation models (ESEM) were employed to compare the three-factor model with five competing models using the items from the Multidimensional Schizotypy Scale (MSS) in 10,814 adults. We subsequently compared the supported models in terms of the prediction of interview (<i>n</i> = 339) and questionnaire assessed symptoms (<i>n</i> of 1,342 to 1,430) using hierarchical linear regressions. The three-, four-, and five-factor models provided the best fit. However, the four-and five-factor models generally did not improve the prediction of outcome measures, although exploratory analyses suggest some potential value for the negative schizotypy subfactor model. The present findings, along with previous validation studies, support positive, negative, and disorganized factors. Ultimately, we recommend that alternative factor and subfactor models of schizotypy should be developed from conceptual, not atheoretical-exploratory approaches.