An examination of the factor structure of the SCI-PANSS

Schizophrenia is a debilitating disorder with far reaching effects on individual functioning. The characteristic symptoms include positive symptoms which are described as psychosis and bizarre behavior. Then there are the negative symptoms which are more numerous in presentation. Negative symptoms include problems in communication, affect, cognition, and behavioral control. While the positive symptoms are more readily defined, the negative symptoms are less so. All of these symptoms present in complex patterns, which has led to difficulty in the accurate measurement of symptom severity. The SCI-PANSS has sought to measure the severity of schizophrenic symptoms. There has been considerable debate centered on the number and composition of dimensions that make up the phenomenon of schizophrenia. The goals of this project were (1) to examine the factor structure of the PANSS rating scale in a sample of individuals diagnosed with schizophrenia, and, (2) simultaneously compare on a single sample the obtained factor structure with several previously derived multi-factorial models of the SCI-PANSS.; This study describes a latent structure of SCI-PANSS symptoms using an exploratory factor analysis (EFA). This model was compared to previously derived models through a confirmatory factor analysis (CFA). Previous models have been described using a variety of patient samples. This left the competing models vulnerable to variation caused by these varied samples. This study improves upon the present literature by removing this confound through the comparison of multiple models on one sample.; An EFA was conducted on a large sample of subjects diagnosed with schizophrenia. This analysis produced a five factor solution. These dimensions include: positive symptoms, negative symptoms, cognitive symptoms, an excitement dimension, and an anxiety/depression component.; This derived factor solution was compared to a number of other competing models that have been derived through other factor analytic studies. These included a null model, a three and four factor solution, several five factor models, and a seven factor model. While none of these models met adequate fit, the results suggest that a pentagonal model reflects the latent structure of schizophrenia. Implications of these findings are discussed with recommendations to improve model specification.