The way you analyze statistics matters too. For instance most don't cross-reference crime statistics with economic background and see how drastically that affects any prior (likely racist) conclusions.
Have you cross-referenced that? What I saw was that the lowest economic quintiles of some groups committed less crime than the wealthiest quintiles of other groups. The "economic" in "socio-economic" is important, but so is the "socio", which may include being discriminated against in the past.
Unfortunately, statistics don't really bear out a lot of popular claims about the impact of poverty. For example, per-student funding does not make as big a difference in academic performance in schools as demographics and the local social environment: there are a lot of schools with bottom barrel funding that perform great and schools with exorbitant funding that perform miserably. And family income is not the strongest predictor of SAT scores.
