This study explored specific and differential effects of metacognitive beliefs on proneness to both hallucinations and delusions in a general population sample, including a control for the alternate symptom. The study then examined whether similar findings were reproduced in a sample of people with psychotic disorders.Linear and hierarchical regressions were used to determine the role of metacognitive beliefs in the proneness to symptoms, whilst ANCOVAs analysed group differences.Participants were recruited to a non-clinical sample (N = 133) and a psychosis sample (N = 100). Both groups completed the Launay-Slade Hallucinations Scale-Revised (Laroi et al., ; Eur. Psychiatry, 19, 15), the Peters Delusions Inventory (Peters et al., ; Schizophr. Bull., 25, 553), and the Metacognitions Questionnaire-30 (Wells & Cartwright-Hatton, ; Behav. Res. Ther., 42, 385).Metacognitions were predictive of both hallucination- and delusion-proneness in the non-clinical sample. Controlled analyses in the non-clinical sample revealed specific effects: low cognitive confidence (CC) predicted hallucination-proneness, whilst negative beliefs about the uncontrollability and danger of thoughts (NBUD) predicted delusion-proneness. Mean ratings on NBUD, low CC, and need to control thoughts were elevated in the psychosis sample; however, after controlling for comorbid symptoms, no metacognitive belief predicted symptom-specific vulnerability in the clinical sample.The pattern of findings provided little support for Morrison's theoretical model of symptom-proneness. Metacognitive beliefs may be related to sub-acute vulnerability to psychosis symptoms; however, the specificity of the relationship between individual metacognitive beliefs and positive psychosis symptoms appeared no longer significant in psychosis patients. The possibility that these metacognitive beliefs are evoked by psychotic experiences, rather than primarily functioning as a driver of them, warrants greater attention.Metacognitive beliefs appear at least equally associated with delusion-proneness as hallucination-proneness. Negative metacognitive beliefs appear more central to delusion-proneness than hallucination-proneness in the general population. When controlling for alternate symptom, no individual metacognitive belief appears reliably able to predict symptom-proneness in psychosis patients.Consistent with existing literature on metacognitions in psychosis, this study adopted a cross-sectional design, meaning we were unable to determine the causal direction of the observed associations between metacognitive beliefs and symptom-proneness. Although a strength of this study design was its control for alternate psychotic symptoms, we did not control for non-psychotic symptoms, particularly, anxiety and depression. The symptom measures used were developed primarily for assessment of psychosis-proneness within the general population; thus, their use by people with established psychosis may have been less sensitive to clinical manifestations of these phenomena.