Exactly. If you assume the population in general is "people caught up in violent crime", "people unlucky enough to be caught for petty crime" and "everyone els

Exactly. If you assume the population in general is "people caught up in violent crime", "people unlucky enough to be caught for petty crime" and "everyone else" then take a sampling of the first two groups (i.e. prison) and then re-check your crime stats of course it's going to look bad compared to the overall population stats. That doesn't mean the risk of becoming a victim of violence for the "people unlucky enough to be caught for petty crime" group goes up anymore than it would if they suddenly found themselves in equal contact with the "people caught up in violent crime" group outside of prison.