@Geri
These stats have nothing to do with each other.
22 times higher may be true if the number for non-Afghans is super low. Say there's 1 rapist in 10,000 in the general population, and 22 in 10,000 among Afghans. This does not mean that a large percentage of Afghans are rapists.
3 Afghans out of 186 offenses can be lots (if there are 5 Afghans in Dorset) or nothing (if there are 50,000).