Theoretical models of anxiety have been developed in adult populations. The applicability of these models in child samples has been assessed using downward extensions of the questionnaires developed to assess the proposed theoretical mechanisms. This poses a challenge, as children are still in the process of developing the skills that are being assessed. Psychometrically sound assessment tools are therefore needed for this developing population, in order to ensure the early detection of mechanisms leading to anxiety disorders in children. This study examined if metacognitions, which play a key role in generalized anxiety disorder (GAD) in adults, can also be reliably assessed in childhood. The study investigated the psychometric properties of the 30-item Metacognitions Questionnaire for Children (MCQ-C30; Gerlach, Adam, Marschke, & Melfsen, 2008) in a national sample of 974 children and adolescents (538 girls) ages 9-17 years. Confirmatory factor analysis supported the 5-factor subscale structure and a 2nd-order total scale factor, which corresponds with previous versions of the scale. MCQ-C30 expectedly correlated significantly with anxiety symptoms and worry. Structural equation modeling revealed that both obsessive-compulsive disorder and generalized anxiety disorder symptoms regressed significantly onto the MCQ-C30. We fitted separate models for children and adolescents, and no noticeable differences are suggested between the models. Female gender was, expectedly, associated with increased levels of general metacognitions. This gender effect was mediated by level of anxiety. Overall, the MCQ-C30 exhibited acceptable psychometric properties in our community sample of children ages 9-17 years. Future studies should investigate the psychometric properties of the instrument in clinical samples and samples of younger children.