[vhawk01]

[ QUOTE ]
[ QUOTE ]
I'm sorry, I interpreted your quote in roughly the following way, "I don't really care about the particulars of genetic classification--racial divisions based on skin color are useful because people with those skin colors behave in certain ways." If this isn't a fair paraphrase, please explain the subtleties I missed.

[/ QUOTE ]

You're not too far off. I'm more than content to use all the information at hand even if it fails the PC racist test. Do you have any doubt about the correlation of American blacks and violent crime? Causation is not the issue, personal safety is. Do you doubt white flight? Can all those whites be making the same wrong assumption, that avoiding urban American blacks makes for a safer life? It's statistically sound, PC poison. I'll stick with the obvious, you are welcome to the social constructs.

[/ QUOTE ]

The bolded part is obvious to the point of being absurd. White flight in no way supports your argument.

The problem I see with racial profiling is analogous to issues with medical screening tests. If the risk for a woman having breast cancer is 5%, and we have a diagnostic test that is 80% effective, a positive result means she is STILL unlikely to have cancer. So, the question is: Does your screening test (black or not, for example) have a higher sensitivity for crime than 80% or does your population (all black people) have a higher incidence of criminality than 5%? I think both of those are obvious no's. So, while it may be entirely valid that a positive result for 'being black' modifies the risk of criminality, it is still overwhelmingly unlikely they are a criminal.

For this reason, most medical societies would never recommend screening tests like the above (most of the screening tests we use either have better sensitivity or the prior risk is higher) or at the very least using them cautiously. The reason I would be opposed to racial profiling is not because of its statistical inaccuracy, necessarily (after all, it has some outcome as long as its done correctly) but the ACTIONS that are then taken from that. We just don't modify the risk enough to really justify ANY sort of action that could have negative consequences (like alienating an entire group of people, trampling on liberty, etc.)

Maybe I am in the minority here, but if this test conferred something like a 98% sensitivity, I would be in favor of it. The amount of discomfort would clearly be justified IMO by the amazing efficacy of the test. We could dramatically cut crime. But this is nowhere near the case.