The first thing I want to clarify is that I don’t think GiveWell needs to settle questions like the one I posed above. Under our new funding model, our donors aren’t stuck with the charity we proclaim as “best”; they will choose between several charities that we will select specifically to represent a range of philosophical values and judgment calls. I think it is worthwhile to discuss these issues, so that we can get our own values out in the open and hear from others on them, but I don’t think that settling this question is necessary for what we’re doing.

That said, I’m happy to try to explain my values further. I disagree that “the best outcome is measured by the summed quality of each day of life.” I’m not sure exactly why I disagree with this – it comes down to intuitions – but here are a few observations that seem relevant:

• I don’t think of the quality of my own life as the summed quality of each day. It’s certainly not what I aim for. To use a somewhat classic example, I’m almost certain that I could increase this quantity for myself through a hypothetical “happiness box,” but I wouldn’t choose to. I care about things like what I accomplish in life (and not just because of the feeling I expect accomplishment to give me), which are not well captured by the sum of the quality of the days.
• Accepting “the best outcome is measured by the summed quality of each day of life” would seem to lead to many beliefs that violate my intuitions very strongly. For example, assuming that the expected value of a life is much greater than the “zero” of nonexistence, this would imply that any time a mother decides to have one fewer child than she could is a tragedy on par with death. That would imply, in turn, that we should morally encourage people to have as many children as the planet will bear. Using a definition of “quality of life” that is more in line with my own view of the good life doesn’t seem to lead to any similar problems.

Again, I speak only for myself, and I don’t feel that “resolution” of these questions is necessary for what we’re doing – though we do need to recognize that we’re dealing with such questions, and it seems useful to be open about where we stand on them. (And, of course, I welcome your continued thoughts; my mind is open.)

Do you prefer program A or program B?

Program A will have a 1/3 chance of saving 600 lives, and a 2/3 chance of saving no lives. Program B will save 200 lives for certain. Which do you prefer?

Most people prefer Program B, because it seems better to save 200 lives for certain.

Do you prefer Program A or Program B?

Progam A will have a 1/3 chance of no deaths, but a 2/3 chance of 600 deaths. Program B will result in 400 deaths for certain. When presented that way, most people prefer program A, because they can’t bear to condemn 400 people to certain death.

Of course, it is simply two different ways of presenting the same scenario, but it results in dramatic shifts of people’s preferences. In a situation like that, you’d want to lead people to see that it is just a result of how you frame it, and lead them to reconsider.

What GiveWell is trying to do is HARD, as you of course know. What is the best way to measure outcome, and why? If you don’t agree that the best outcome is measured by the summed quality of each day of life (as evidenced by your valuing the 20 year old more than the infant) then try to explicate why? What is your rationale?

I’m just trying to provoke thought. I think what you are trying to do at GiveWell is a great thing.

Ron: for what it’s worth, your reminder doesn’t change my mind - that’s exactly what I had in mind when I laid out the initial stylized “choice.” I agree that serious thought may lead people to change their views on questions like this, but I don’t think it’s safe to assume that all disagreements are a matter of who is thinking “most seriously” - different donors may have legitimately different values.

I suspect that your preference for #2 has something to do with the simple fact that good health is a lot better than poor health. After folks have made their choice, I’d ask that they consider it but with the following reminder: If you choose to fund #1, the 100 people you could have funded with choice #2 will lead lives in poverty and illness. If you choose to fund number #2, the 100 people you could have funded with choice #1 will die. My guess is that this will cause a lot of people to switch from choice #2 to choice #1, though rationally, they should have thought of that any so it shouldn’t have an effect.

Having worked in a developing country for several years, my guess is that your attitude is rare among people with development-related work experience from poor countries. I say this because the three alternatives, including your attitude, fail to take into account the complexities of poverty.

#2 also seems pretty worthwhile but I have a hard time to even consider #3 over #1…

@RubyK, losing your own baby isn’t part of the question. You need to think about the real options: a 20-yr-old stranger dies or an unknown baby dies. Which do you dislike more? If you’re still stuck on the point about emotional attachment, you need to at least keep it consistent: your own baby dies or your own 20-yr-old dies (but this is less relevant for thinking about where to donate, in my opinion).

And it’s great that there are certain procedures that are very leveragable and save lots of infants, but the question of the relative worth of saving an unknown infant vs improving the life of an unknown adult is still crucial.

Interventions that will best accomplish (1) are actually health system and capacity building (e.g. skilled attendants at childbirth, case management of pneumonia and neonatal sepsis) and not vertical bednet-type programmes so provide knock on effects greater than just the one “potential” life that you have saved.