Calculating friendly fire deaths is a tricky proposition at best. If you remove the number of non-combat deaths; (training accidents, illness etc.) the number killed in Viet Nam can drop as low as 48,000. That alone skews the friendly fire percentages up to the 25% range. If you add Reservists, National Guardsmen and Active Duty Personnel killed stateside during Viet Nam to the equation; the percentage goes down.

I've read one study after another over the years and I can't say anyone has pinned it down exactly. You just about have to decide which parameters you want to use and pick a number from there.