Yeah, statistics can be twisted to support a view that someone might want to prove, but they can provide useful insight if you approach with some education in statistics and an open mind.
Of course there will be some significance in how different riding styles and crashes are defined. It can get pretty complicated. Consider a recent incident I had. Cycling on a greenway MUP, I was going left around a sharp curve, with vegetation close to the trail, blocking the view around the corner. Coming around the curve the other way was one of those 6-wheeled, gasoline powered Gators that are commonly used in sports arenas. She was taking up the whole path, going around a right hand curve, enjoying the view of nature off to her left, all while talking on the cell phone! I was able to avoid a collision by going off the paved path and stopping. How would you define a collision like that?! She was not in a car, but it was a motor vehicle, it was not a public road, but it was paved with asphalt. Add into the statistical mix cell phone use, and the lines start to blur. But I would still like to hear some sort of legitimate statistics concerning road vs. off-road injuries and death.