When I was last on the job market, I spent a bunch of my free time trying to come up with well-justified cost-effectiveness estimates for a wide array of different interventions across several cause areas. I think something like this technically meets your three criteria, but I suspect it isn’t quite consistent with the spirit of what you’re looking for (i.e. projects that take longer than a week to do and will actually probably have some positive impact). Even though I don’t think my CEAs did anything at all to improve the world themselves, I’d still recommend this to early-career EAs, if only because I thought it was a huge help when applying for jobs at GiveWell and Open Phil (which, for the sake of full disclosure, I was not ultimately offered, though I made it fairly far in the process). Even for people who have no interest in working somewhere like GiveWell or Open Phil, I think doing this trains a lot of important skills: conducting literature reviews, thinking about counterfactuals and measuring counterfactual impact, thinking in terms of DALYs or QALYs, some math… etc., and it just isn’t that much of a commitment, either. I probably spent a few hours a day, most days, for up to week on each estimate, but you could spend more or less as you saw fit. It has an appealing kind of flexibility. Finally, it’s highly scalable — there’s no shortage of things to estimate the cost-effectiveness of, so there’s no reason why tons of people couldn’t all reap the human capital and intellectual benefits of doing this. In the aggregate, I think that itself could have a pretty positive impact, and if someone were to find strong evidence that some previously overlooked intervention was actually competitive with, say, the AMF, that would be a pretty great thing for the EA community to know!
When I was last on the job market, I spent a bunch of my free time trying to come up with well-justified cost-effectiveness estimates for a wide array of different interventions across several cause areas. I think something like this technically meets your three criteria, but I suspect it isn’t quite consistent with the spirit of what you’re looking for (i.e. projects that take longer than a week to do and will actually probably have some positive impact). Even though I don’t think my CEAs did anything at all to improve the world themselves, I’d still recommend this to early-career EAs, if only because I thought it was a huge help when applying for jobs at GiveWell and Open Phil (which, for the sake of full disclosure, I was not ultimately offered, though I made it fairly far in the process). Even for people who have no interest in working somewhere like GiveWell or Open Phil, I think doing this trains a lot of important skills: conducting literature reviews, thinking about counterfactuals and measuring counterfactual impact, thinking in terms of DALYs or QALYs, some math… etc., and it just isn’t that much of a commitment, either. I probably spent a few hours a day, most days, for up to week on each estimate, but you could spend more or less as you saw fit. It has an appealing kind of flexibility. Finally, it’s highly scalable — there’s no shortage of things to estimate the cost-effectiveness of, so there’s no reason why tons of people couldn’t all reap the human capital and intellectual benefits of doing this. In the aggregate, I think that itself could have a pretty positive impact, and if someone were to find strong evidence that some previously overlooked intervention was actually competitive with, say, the AMF, that would be a pretty great thing for the EA community to know!
Thank you so much!
I agree and am adding this to our list of types of projects to suggest to students :)