THE INTERRATER RELIABILITY OF THE POSITIVE AND NEGATIVE SYNDROME SCALE (PANSS)

Methodological weaknesses of previous reliability studies on the Posit ive and Negative Syndrome Scale (PANSS) include too small samples of r aters and inadequate use or report of measures. In this study, videota ped interviews with 13 schizophrenic patients were rated by an average of 5.2 raters per interview according to a Norwegian version of PANSS . The study shows relatively low ICCs of most of the negative and some general psychopathology symptoms, whereas ICCs of positive symptoms w ere in general high. Interrater reliability is very good on all subsca les and on symptom groups based on principal component analysis. This indicates that it is much safer to use scores based on symptom groups and subscales than on single items.