Purpose -In the optimisation of processing parameters for additive manufactured parts using, e.g. selective laser melting (SLM) or electron beam melting, the measurement of the part densities is essential and of high interest. However, there is no common standard. Different institutes and system providers are using their own principles and guidelines. This study investigates the accuracies of the three measurement principles: Archimedes method, microscopic analysis of cross sections and X-ray scanning. Design/methodology/approach -A total of 15 test samples on five density levels (densities between 90 and 99.5 per cent) were produced using the SLM process. The samples are analysed regarding the accuracy of the measurement principles and their reproducibility taking into account influencing parameters like the buoyancy of a sample in air (Archimedes method) or different magnifications of a cross section. Findings -The Archimedes method shows a very high accuracy (^0.08 per cent for high densities) and repeatability (^, 0.1 per cent) on all density levels. In contrast to the Archimedes method, taking a micrograph of a specific cross section allows to influence the resulting density and the coefficient of variation reaches values . 4 per cent. However, for low porosities, mean densities are comparable to the results of the Archimedes method even though calculated densities are typically somewhat too high. The advantage of the image guided analysis (2D and 3D) is getting more information about the distribution, size and form of pores in the part. Originality/value -The findings do not only refer to metallic parts but generally to all parts having a specific porosity. The study is a contribution to the American Society for Testing and Materials initiative F42 "Additive Manufacturing Technology" and especially to the subcommittee "test methods".