Background. A metacognitive approach to the conceptualization of paranoia as a strategy for managing interpersonal threat has gained some support in studies of non-clinical populations. This study reports a clinical validation of the Beliefs about Paranoia Scale (BaPS), a self-report measure to assess metacognitive beliefs about paranoia. We aimed to replicate the factor structure of a brief version of the measure and test the specific hypotheses that positive beliefs about paranoia would predict levels of suspiciousness, and that negative beliefs about paranoia would predict problematic persecutory delusions. Method. A total of 122 patients meeting criteria for a diagnosis of a schizophrenia spectrum disorder completed the questionnaire assessing beliefs about paranoia. In addition, 61 of the participants were administered the Structured Clinical Interview for DSM-IV, and 60 were administered the Positive and Negative Syndromes of Schizophrenia Scale. One hundred and seventy-eight non-patients were also recruited (an undergraduate sample). Results. Principal components factor analysis showed that the three-factor solution was replicated (comprising negative beliefs about paranoia, paranoia as a survival strategy, and normalizing beliefs). This measure showed good internal consistency (alphas ranged from .85 to .91). Correlational analyses revealed that positive beliefs about paranoia were positively associated with levels of suspiciousness, and independent t tests showed that negative beliefs about paranoia were significantly higher in patients with a diagnosis of schizophrenia meeting criteria for persecutory delusions in comparison to those without. Analyses of covariance showed that patients scored higher than non-patients on both positive and negative beliefs, but logistic regression did not demonstrate that co-occurrence of these beliefs predicted patient status. Conclusions. Three of our four hypotheses were confirmed, suggesting that a metacognitive approach to the conceptualization of paranoia as a strategy for managing interpersonal threat may have some utility for understanding clinical paranoia. Such a model is described and the clinical implications of the findings are also discussed.