On Thursday 13 Sep 2007 7:13 am, Charles Haynes wrote: > On 9/13/07, shiv sastry <[EMAIL PROTECTED]> wrote: > > He has less absolute amounts of money than many of his peers and the > > people in his new nation and is therefore more likely to be tempted to > > cheat or steal. This would call for racial profiling as a way of tackling > > crime, and it would be right. > > Class, not race and it would only be "right" if you believe > correlation is the same as causality.
I stand corrected. I should have said class and not race, though in the particular example that I used, race could be convenient starting point for profiling. Still, I see some very interesting fallout from the correlation "relative poverty==tendency to criminal behavior" because of what I believe is a widespread tendency to ignore less visible "other things" and making correlation=causality. For example, in an ancient social system such as we find in India, there are pools of castes who are far less likely to get high paying jobs than others. A large majority of people in the highest paying jobs and professions belong to higher castes such as Brahmins and Vysyas, and a large proportion of the people in the lowest paid jobs tend to be from the lowest castes or the former outcastes - now called Dalits. Using the correlation "relative poverty==tendency to criminal behavior" and ignoring invisible "other things" that could render this as a poor predictor of causality, we can come up with the concept of "criminal castes". For example, a Dalit who gets a college education by means of reservation and then gets a high paying job will still come from a social group who are paid relatively much less. Using the correlation in the previous paragraph we can reach the conclusion that this Dalit's relatives and extended family are more likely to show criminal behavior related to temptation and disparity. This would be in contrast to say a Brahmin or Kshatriya colleague of his, whose social group is well off and are less likely to show criminal behavior. shiv
