The Effective Altruist's Prisoners' Dilemma
A discussion recently erupted with several friends over a tweet about an extra credit question posed by a professor at the University of Maryland:
One of my friends commented that the rational thing to do is to select 6% - unless you happen to be that marginal student whose choice brings everyone down, you can only expect to gain by selecting 6%.
My immediate reaction was that, well no, that's the rational thing to do provided you are egoistic and only care about your own exam score. If you're a rational altruist, though, the rational thing to do may be to select 2%, since in the unlikely event that you are the marginal student, you threaten to lose points for everyone. Depending on the size of the class and the way you value each additional point on the exam, this could easily outweigh the slight chance of getting an extra 4% for yourself.
As is often the case, things are more complicated. The reason is this: what if there is a curve? If there's a curve, then additional points on the exam only serve to set you aside from anyone else, and if you cause everybody to lose their bonus points, you just leave the relative distribution unchanged. From an altruistic perspective, if the value of everybody's exam score is equal, then it's unclear which way to answer this question.
It's more likely that everybody's exam score is not equal, though. If I'm a truly effective, altruistic person and (almost) the rest of the class is not, then I should select 6%, since in the scheme of things, it's better for those who will do good with their credentials to outcompete those who won't.
The irony is that a radically altruistic position leads to the same choice as a radically egoistic one. It seems to me that this is likely the correct assessment. I could see worries that effective altruism could lead to a cutthroat world, but these are easily allayed by the fact that the calculation changes if I know that 10% of the class is likely to be effective altruists.
In fact, this could be somewhat comforting for those who worry about effective altruism requiring some holier-than-thou self-sacrifice. Certainly, a dose of self-sacrifice is called for. But if you want to be not simply altruistic but also effective, the best default way of behaving in many situations may be the rationally egoistic one.
WHAT KIND OF PROFESSOR DOES THIS pic.twitter.com/ACtQ0FCwRm
— name (@shaunhin) July 1, 2015
One of my friends commented that the rational thing to do is to select 6% - unless you happen to be that marginal student whose choice brings everyone down, you can only expect to gain by selecting 6%.
My immediate reaction was that, well no, that's the rational thing to do provided you are egoistic and only care about your own exam score. If you're a rational altruist, though, the rational thing to do may be to select 2%, since in the unlikely event that you are the marginal student, you threaten to lose points for everyone. Depending on the size of the class and the way you value each additional point on the exam, this could easily outweigh the slight chance of getting an extra 4% for yourself.
As is often the case, things are more complicated. The reason is this: what if there is a curve? If there's a curve, then additional points on the exam only serve to set you aside from anyone else, and if you cause everybody to lose their bonus points, you just leave the relative distribution unchanged. From an altruistic perspective, if the value of everybody's exam score is equal, then it's unclear which way to answer this question.
It's more likely that everybody's exam score is not equal, though. If I'm a truly effective, altruistic person and (almost) the rest of the class is not, then I should select 6%, since in the scheme of things, it's better for those who will do good with their credentials to outcompete those who won't.
The irony is that a radically altruistic position leads to the same choice as a radically egoistic one. It seems to me that this is likely the correct assessment. I could see worries that effective altruism could lead to a cutthroat world, but these are easily allayed by the fact that the calculation changes if I know that 10% of the class is likely to be effective altruists.
In fact, this could be somewhat comforting for those who worry about effective altruism requiring some holier-than-thou self-sacrifice. Certainly, a dose of self-sacrifice is called for. But if you want to be not simply altruistic but also effective, the best default way of behaving in many situations may be the rationally egoistic one.
Comments
Post a Comment