I am far from an expert on this subject, but I'm sure that I have read that the holes left in paper targets are usually a little bit smaller than the caliber size. If this were true, it would make groups measured using this method prone to being slightly smaller than they actually are. The amount of effect this would have on relative group size (that is percentage) should increase as the group size decreaes. That is, on a 2" group a .1" error would make much less difference than on a .2" group. I've heard of actually measureing the centers of the bullet holes, however, I don't know of any superior way to do this.