If you survey 100 people who were mugged once last year about how many times they have been mugged this year and 50% said the same (lets say 1) and 50% said increased to 2, then you would have a 50% increase (from 100 to 150 total) this year. So yeas, these stats can be perfectly valid.