One quick way to get people to not take you seriously is with a bad cost effectiveness estimate. There’s a much bigger risk of doing a sloppy/overconfident job than benefit of having a high number at the end of it (in EA circles). Also, there is a reputation of these estimates to both produce amazing numbers and also be very wrong, so while I support attempts, I’d also recommend lots of clarification, hedging, and consideration of ways the number could be poor. I think the default expectation is for the number to not be great; but even if the median isn’t good, it’s possible upon further investigation it could be better than expected, which could be quite worthwhile.
“to reach all chronic sufferers” → I’d recommend targeting 30%-60% of sufferers. The last several percent would be much more expensive.
I’m quite skeptical of the click → cure stats in particular. For-profit websites often have a 1% rate of people who go from click → purchase, and this could be a pretty significant amount of work to purchase.
Is this equation taking into account that the “cure” could last for many years? Would the result be in “QALYs per year”?
I’m sure you’ve answered this elsewhere, but why the American focus? Would it be possible in India or similar?
This estimation seems like something that Charity Entrepreneurship would have a lot more experience in. The program seems quite similar to some of their others.
Thanks!
Some quick thoughts:
One quick way to get people to not take you seriously is with a bad cost effectiveness estimate. There’s a much bigger risk of doing a sloppy/overconfident job than benefit of having a high number at the end of it (in EA circles). Also, there is a reputation of these estimates to both produce amazing numbers and also be very wrong, so while I support attempts, I’d also recommend lots of clarification, hedging, and consideration of ways the number could be poor. I think the default expectation is for the number to not be great; but even if the median isn’t good, it’s possible upon further investigation it could be better than expected, which could be quite worthwhile.
“to reach all chronic sufferers” → I’d recommend targeting 30%-60% of sufferers. The last several percent would be much more expensive.
I’m quite skeptical of the click → cure stats in particular. For-profit websites often have a 1% rate of people who go from click → purchase, and this could be a pretty significant amount of work to purchase.
Is this equation taking into account that the “cure” could last for many years? Would the result be in “QALYs per year”?
I’m sure you’ve answered this elsewhere, but why the American focus? Would it be possible in India or similar?
This estimation seems like something that Charity Entrepreneurship would have a lot more experience in. The program seems quite similar to some of their others.
I’d suggest reading up on the mini-fiasco of the leafletting research, if you haven’t yet. Just make sure not to make some of the mistakes made around that. Some context: https://animalcharityevaluators.org/blog/ace-highlight-updated-leafleting-intervention-report/ https://acesounderglass.com/2015/04/24/leaflets-are-ineffective-tell-your-friends/ https://medium.com/@harrisonnathan/the-problems-with-animal-charity-evaluators-in-brief-cd56b8cb5908
Consider using Guesstimate for clarity, but I’m biased :)
Kudos for the efforts, and good luck!