Does a latent class underlie schizotypal personality disorder? Implications for schizophrenia.

Despite growing enthusiasm for dimensional models of personality pathology, the taxonic versus dimensional status of schizotypal personality disorder (PD) remains a point of contention in modern psychiatry. The current study aimed to determine empirically the latent structure of schizotypal PD. We examined the latent structure of schizotypal PD in the Psychiatric Morbidity Survey in Great Britain and the second wave of the U.S.-based National Epidemiologic Survey on Alcohol and Related Conditions (NESARC) survey. We analyzed composite indicators created from participant responses using the mean above minus mean below a cut (MAMBAC), Maximum Covariance (MAXCOV), and latent mode factor analysis (L-Mode) taxometric procedures. We also analyzed item-level responses using two latent variable mixture models--latent class analysis and latent class factor analysis. Taxometric and latent variable mixture analyses supported a dimensional, rather than taxonic, structure in both epidemiological samples. The dimensional model better predicted psychosis, intellectual functioning, disability, and treatment seeking than the categorical model based on DSM-IV diagnosis. People meeting criteria for schizotypal PD appear to exist on a spectrum of severity with the rest of the population. The possible dimensionality of schizotypal PD adds to growing support for a dimensional structure of PDs including other Cluster A disorders.