Factor versus cluster models of schizotypal traits. I: a comparison of unselected and highly schizotypal samples.

Factor analytic studies have long supported the division of schizophrenic symptoms into three relatively orthogonal factors: positive symptoms, negative symptoms, and disorders of relatedness/disorganization. Similarly, factor analyses of schizotypy often yield three factors: positive symptoms, negative symptoms, and social anxiety or disorganization. Recent cluster analyses, however, suggest that not all patients can be simply categorized according to these factors. Cluster analyses of schizotypal symptoms tend to result in clusters of individuals who are low in all factors, high in more than one factor, or high predominantly in one factor. The present study sought to compare factor and cluster models of schizotypal symptoms, as measured by the SPQ, PAS, and MIS, in unselected individuals and highly schizotypal individuals. Consistent with prior research, factor analysis of a large unselected undergraduate sample yielded three factors: "positive", "negative", and "disorganized." Factor analysis of schizotypal undergraduates produced the same three factors, plus a fourth designated "paranoid thinking." In contrast, cluster analysis of the unselected sample yielded four clusters ("low schizotypy", "average schizotypy", and "high schizotypy", plus "positive/disorganized"). Cluster analysis of the schizotypal subsample produced four clusters: "low schizotypy", "positive", "negative" and "high schizotypy."