Lost in the numbers
May 14, 2009 10:42 AM Subscribe
In light of the proper statistics, can discrimination be a form of rational reasoning?
A large part of my work deals with predicting events objectively using large amounts of data. I've been wondering lately how this applies to my stereotypes and beliefs, as well as the workings of society.
If someone reads that homicides, for example, are largely perpetrated by African-American youth, then when faced with a young black man walking down the street, is it rational to be more cautious than when faced with a young white man in identical garb? Given no other knowledge of these men a priori than their skin color, it's a pretty straightforward application of the Bayes' Theorem to note that the probability of a crime occurring between the two scenarios is clearly not the same, yet we are commonly advised that this type of thinking is racist and wrong.
Race and age are protected characteristics, but it strikes me as wasteful and suboptimal for regulations to put, say, elderly white females, through the same dance at airport security than everyone else (than even myself, a 24 year old minority male). On the other hand, I've had Muslim friends tell me that they make sure to shave and don a nice blazer before going on a flight, as clearly this type of profiling already happens informally. Instead of skirting around the subject, why can't we figure out how to use the data we have to more effectively distribute resources when dealing with heterogeneous populations in a respectful way?
How do you think about this in your own daily life? If you were going to invest your own money in prosper loans, or allocate dosages of scarce medicines, would you refrain from using age/ethnicity as a factor? On the flip-side, when does this type of statistical reasoning cross over to hurtful behavior and racism?
Thanks!
A large part of my work deals with predicting events objectively using large amounts of data. I've been wondering lately how this applies to my stereotypes and beliefs, as well as the workings of society.
If someone reads that homicides, for example, are largely perpetrated by African-American youth, then when faced with a young black man walking down the street, is it rational to be more cautious than when faced with a young white man in identical garb? Given no other knowledge of these men a priori than their skin color, it's a pretty straightforward application of the Bayes' Theorem to note that the probability of a crime occurring between the two scenarios is clearly not the same, yet we are commonly advised that this type of thinking is racist and wrong.
Race and age are protected characteristics, but it strikes me as wasteful and suboptimal for regulations to put, say, elderly white females, through the same dance at airport security than everyone else (than even myself, a 24 year old minority male). On the other hand, I've had Muslim friends tell me that they make sure to shave and don a nice blazer before going on a flight, as clearly this type of profiling already happens informally. Instead of skirting around the subject, why can't we figure out how to use the data we have to more effectively distribute resources when dealing with heterogeneous populations in a respectful way?
How do you think about this in your own daily life? If you were going to invest your own money in prosper loans, or allocate dosages of scarce medicines, would you refrain from using age/ethnicity as a factor? On the flip-side, when does this type of statistical reasoning cross over to hurtful behavior and racism?
Thanks!
This post was deleted for the following reason: this really really seems like a chatfilter post with a modicum of data tossed in. -- jessamyn
« Older Emergency whitening | Legal Advice Filter (Yes, I know YANML) Newer »
This thread is closed to new comments.