The selection of spatter stains from an impact pattern is the foundation by which the estimation of the area of origin is derived. In this paper the authors establish bloodstain selection criteria based on a statistical model. The model is constructed with data collected from several impact patterns created within a laboratory environment. The response variable considered is the distance between the known source and the estimated trajectory emanating from a stain; the explanatory variables are impact angle, glancing angle, zone location of the stain on the wall, and distance from the source to the wall. The model obtained indicates that zone location is relevant for improving accuracy, and that stains with any alpha angles can be used to calculate an area of origin estimation. Results from a validation study for the stain selection rules derived from the model are presented and analyzed. (Published Abstract)