One attacks on both are rare.
Two you would need at least two groups, an OC group, a CC group, and a control group. You would need to insure all groups are similarity situated, in terms of income, neighborhood, age etc.
Of course for a statistical study you also need to determine how many people actually carry a gun frequently, and your study needs to define "frequently"
There is so many factors in this, and of course attacks on both are exceedingly rare, in fact attacks on people who don't carry guns are rare for that matter.
Basically it's such a specialized statistic that no one tabulates it, and it would be incredibly complicated to make your own statistic, at least if the statistic has to be credible. If you are arguing a CC only guy, it suffices to say, attacks on open carriers are so rare no one bothers to keep statistics