Are universities a cult? Do charitable interventions like de-worming work? How much should we trust the conclusion of well-respected charity evaluators like GiveWell?
Uri is the publisher of The Browser and The Listener, the world's favourite curation newsletters, and the author of Thinking Statistically and The Business of Big Data. Uri can be found at uribram.com or firstname.lastname@example.org.
As we mention in the audio, this episode includes a critique of Givewell. Givewell were kind enough to listen to our recording and send us a reply. Here's their reply:
We're excited to see this level of detailed engagement with our research. As Uri and Spencer note, one of the key reasons we share the full analysis behind our recommendations is precisely this: inviting fresh perspectives and debate on the conclusions we reach.
We operate in an expected value framework when recommending top charities. We recommend deworming programs because of the possibility that deworming may have a large impact on long-term economic well-being. At less than $1 per treatment, we think it's a pretty good bet. We've discussed our views publicly over the years, such as in our blog post titled "Deworming might have huge impact, but might have close to zero impact."
The case for deworming's long-term benefits does rely on a relatively small number of studies. And the mechanisms by which it has long-term impact are unclear. But when we account for these uncertainties in our impact estimates, it still remains promising.
We've also supported research to better understand the impacts of deworming. We funded part of a study that measured the economic welfare of children who received deworming treatments 20 years later. This work was recently published, and at a high level, seems to support the story of deworming's long-term effects.
Thanks again for discussing this topic—it's an important and thorny one!
Givewell also mentioned some corrections to some of the claims made in the episode. They said:
[We] noticed some comments outside of the deworming conversation that didn't reflect our views and flagged a few of the more important ones below.
- In addition to the groups you listed, our current list of top charities includes Malaria Consortium's seasonal malaria chemoprevention program and Helen Keller International's vitamin A supplementation program. The full list is here: https://www.givewell.org/charities/top-charities.
- The two outcomes we recommend our current list of top charities for are averting deaths (not improving nutrition) and increasing incomes/consumption. We are open to considering additional outcomes in the future.
- Uri said the following in regards to cash transfers: "I might be wrong but I think GiveWell doesn't count—if you took the money and spent it on a one-off way that didn't increase your long-term wealth or income—then GiveWell wouldn't count that." This is not accurate. We model short-term as well as longer-term benefits to cash transfers. This is reflected in our cost-effectiveness model and discussed in this blog post.