Eric Herboso mentioned recently that GiveWell doesn’t value non-human animal lives much.
What a controversial subject! I know people who value animal lives at 0% of human lives; I know people who would say that all animals are equal. Whether you value an animal life at 0.002 of a human life, as in Eric’s example, 0, or 1 makes a huge difference to which kinds of programs you think one ought support.
This really highlights what I’ve always said about how there are two separate issues that those in the non-profit business all-too-often conflate: what is the most effective way at obtaining an outcome, and how to value outcomes.
When advocating for particular causes, people often shove under the rug the fact they’re saying anything about the second one. It might be clear which outcomes are under discussion (e.g. DALYs/QALYs), but there should still be more discussion of why these are the right ones. If we can’t even agree on the value placed on non-human animal lives, there is no way we can agree on what we should do, and this is just one issue among many on which we might disagree. (For example, do we value education at all, or just health?)
I’m not saying that people’s values are sacrosanct; there could theoretically be value in arguing with someone over what they should value. But let’s transparently separate the two issues. AidGrade, for its part, has been simply presenting outcomes. People with a wide range of beliefs can thus use the results. I would be all for weighting these results in some more complicated function, but arguing for a particular function is an entirely different enterprise.