Variation management in additive manufacturing (AM) is progressively more important as technologies are implemented in industrial manufacturing systems; hence massive research efforts are focused on the modeling and optimization of process parameters and the effect on final part quality. These efforts are, however, hampered by the very problem they are seeking to solve, as conclusions are weakened by poor validity, reliability, and repeatability. This paper details an elaborate experiment design and the subsequent execution with the aim of making the research data available without loss of validity. Test artifacts were designed and allocated to fixed positions and orientations in a grid pattern within the build chamber to facilitate rigid analysis between different builds and positions in the build chamber. A total of 507 specimens were produced over three builds by laser sintering PA12 before inspection with a coordinate measuring machine. This research demonstrates the inherent variations of laser-based powder bed fusion of polymers (LB-PBF/P) that must be considered in experiment designs to account for noise factors. In particular, the results indicate that the position in the xy-plane has a major influence on the geometric accuracy, while the position in the z-direction appears to be less influential.