I don't think this is remotely true. Compiling the statistics does not commit you to a particular use. Obviously, people can abuse them, but here's an example from just this month of a case where you're glad to have the racial breakdown, because it clues you in to a piece of good news and a possible root cause:

http://www.motherjones.com/kevin-drum/2013/08/lead-crime-rac...

Impoverished communities are a problem here.

For some reason whether you have a community of muslims or africans, it happens to be impoverished. On the other hand, if you have a community of asians or western europeans, it doesn't.

That's exactly the problem parent is trying to convey. It's not that africans are inherently criminal, it's just that they have dangerous group dynamics.

> On the other hand, if you have a community of asians or western europeans, it doesn't.

I might be mis-understanding the context here: I've been to a few impoverished communities composed of 98% descendants of Western Europeans. Think: white trash.

And they commit only a fraction of the amount of crime of their black counter-parts: https://news.ycombinator.com/item?id=6279991