Following a convincing demonstration of the prediction power of the Gurson model for ductile fracture, it is now required to select realistic material parameters for practical applications of the model. In this paper, using studies on a smooth tensile specimen, a notched tensile specimen, a centre-cracked tensile panel and an analytical cell model, a sensitivity analysis of the material parameters is performed that includes the initial void volume fraction of the primary inclusions and the void volume fraction of the secondary inclusions when fitting the critical void volume fraction. Voids that nucleated from primary and secondary inclusions have been considered separately. It has been found that in either case the selection of material parameters for the finite element analyses is not unique, and the most significant parameter for the predictions is the nucleation burst strain. Some general conclusions concerning the selection of material parameters for the Gurson model have also been made.