3D printing, using material extrusion (a.k.a. FDM), is increasingly used for final parts, and not only prototypes. This increases research into printed material strength. For such research to be reproducible, many parameters must be controlled. Assessed here are several such parameters currently underappreciated in the literature: variation between printers, printing toolpath, void content, and bed placement. Through systematic testing (n = 780), the first three parameters are proven to be statistically significant. With these controls, the effect of layer height, print speed, nozzle temperature, and print orientation on strength as previously reported is accurately reproduced, and contradictory prior results regarding print orientation are explained. A single material was used: polylactic acid, benchmarked by injection molding and testing of specimens with comparable geometry to the printed specimens.