Stats also form a bit of a problem when looking at repeat offenders. The below is obviously in no way representative of the truth, I'm just using it to illustrate how numbers can be misleading.
Population:
900 white people
100 black people
100 crimes committed.
60 crimes committed by 60 different white people.
40 crimes committed by 8 different black people.
Ignoring unique criminals, you can say black people make up 10% of the population, but commit 40% of the crime, their crime : population ratio is 4:1.
The same metric for white people (60% of the crime, 90% of the population) gives a crime : population ratio of 2:3, or 0.67:1
Oh noes, black people are 6 times more likely to commit crimes than white people (4 / 0.67 = 6)
However, only 8% of the black population actually committed any crimes, whereas 6.7% of white people committed crimes.
So from a unique person perspective, black people are only 20% more likely to commit crimes than white people ( 8 / 6.7 = 1.2), a far cry from the 600% we saw in the previous calculation.