The Giessen Physical Complaints Inventory for children and adolescents (GBB-KJ) is currently the only German test for the standardised, multidimensional measurement of physical complaints in the self and external assessment of children and adolescents. The present study analysed the scalability of the five complaint dimensions in a mixed Rasch model with up to four clusters and four response models (partial credit, dispersion, equidistance and rating scale). The self-report data of N=1027 11- to 18-year-olds used in the study stem from the nationally representative Hamburg Health Survey (Barkmann, 2004). Of the five original 7 item scales, only cold problems could be Rasch scaled in a 3 cluster rating scale model without the exclusion of nonconform responding cases. In the 5 item rescaling approach, more than one valid model could be identified for each of the five dimensions. In principle, physical complaints proved to be adequate for probabilistic measuring models. Future studies must decide which of the solutions offers the more significant and consistent results in scientific research and clinical practice.