Lenzenweger M F, Dworkin R H, Wethington E
J Abnorm Psychol. 1989 Feb;98(1):62-70. doi: 10.1037//0021-843x.98.1.62.
The present investigation empirically evaluated three competing models of the relations between positive and negative symptoms in schizophrenia, namely the severity-liability model (Gottesman, McGuffin, & Farmer, 1987), Andreasen's unidimensional bipolar model (Andreasen & Olsen, 1982), and Crow's independent dual-process model (Crow, 1980a, 1980b). Using positive and negative symptom ratings based on 220 schizophrenic subjects, the results of a LISREL VI (Joreskog & Sorbom, 1984) confirmatory factor analysis revealed that Crow's model of positive and negative symptoms provided the best fit to the observed data among the three models. The severity-liability model provided a modest fit to observed data, and Andreasen's model fit the data poorly. Results are interpreted as supporting the validity of the positive and negative symptom distinction in schizophrenia and as providing substantive empirical support for Crow's independent dual-process model. The methodological advantages of confirmatory factor analysis in the specification and evaluation of theoretical models in experimental and developmental psychopathology are discussed.
