Blog post

Interview with GiveWell: Should we be funding research?

7 min read
13 Jun 2013


Research Director Robert Wiblin interviewed GiveWell co-founder Elie Hassenfeld (pictured right) on a range of issues suggested by volunteers and members of Giving What We Can. This is the first instalment, focussed on the issue of meta-charity - or research into which forms of charity are most effective.

An unabridged transcript of the interview is available here and audio for download here.

RW: Do you think there’s any chance that giving to organisations that research which interventions and organisations are most cost-effective, such as GiveWell, but also other research outfits like the Jameel Poverty Action Lab, the Copenhagen Consensus Centre, Aid Grade and so on could have a greater impact with each dollar than the organisations that you actually recommend?

EH: It’s definitely possible, I think it’s an interesting question. We’ve evaluated Innovations for Poverty Innovation (IPA), one of the organisations you’ve listed in the past. Our view is that, the organisations that we rate most highly right now, Against Malaria Foundation (AMF), Give Directly and the Schistosomiasis Control Initiative (SCI), are all better bets than IPA. We’ve definitely thought a lot about funding research, we’ve talked about this in the context of this cause of meta-research.

We’ve talked about this part of GiveWell Labs that would be funding a group like the Cochrane Collaboration. We’ve also thought about funding randomised controlled trials (RCTs) themselves. There’s questions involved there; the research we found in our work is actually more the research done by the global health community. That would be groups that are performing the RCTs on malaria nets, water and nutrition trials. That’s work that we’ve found really helpful.

I do think it’s interesting that people ask us about funding research or meta charity, and they tend to focus on the groups that are more visible, the groups like IPA or Jameel Poverty Action Lab, and less the groups that I think have been most helpful to us, which is just the general global health research. The other group that’s been incredibly valuable to us is the work done by literature reviewers broadly and the Cochrane Collaboration specifically. That’s a group to which we directed what we call a quick grant about a year ago and we are interested in what they do and have been following that.

RW: So did you manage to find any organisations that were doing that basic health research, who you could potentially fund?

EH: My impression is that a lot of that research comes from broader academic departments. There are researchers at universities working in the World Health Organisation. This is the type of thing where the type of groups to which an individual donor can write a check and cause something to happen may be different than the opportunity for a larger funder to fund a specific project. And that distinction is one that is becoming particularly salient to us as we think about the move from our traditional focus on direct aid-providing organisations, to other areas where we almost want to be thinking more like a major philanthropist. And so I think that thinking how to fund RCTs or global health research will likely be done more in the vein of here is an RCT that we find particularly valuable, so we’re going to go and identify the funding to implement this RCT and then provide it all in one go, rather than giving more money to organisations to support RCTs more broadly. Obviously one can do that for IPA and then all of the questions you want to ask are the ones that we discussed in our review of IPA.

RW: Sure, so did you have much success finding out whether these kinds of meta-charities were having influence on [government or charity] spending. Was that something that was possible to identify?

EH: It’s tough. I mean, IPA is the one we’ve looked at the most and we have a whole page up on them, about our best assessment of the impact that their research they’ve done has had on the conversation. The studies that seem most impactful to us are studies on deworming, studies on free bednets, the studies they conducted on microfinance. That does have an influence. Without a doubt, other trials have had an influence. I don’t think this is the type of thing where it was easy to possible to perform a very direct and obvious case of the exact nature of the influence that they had.

RW: Have you found organisations receptive to getting that kind of tied grant to perform a particular RCT or a particular literature review. Is that something that donors like our members should be thinking about trying to pursue?

EH: I think it’s difficult to pursue on their own. So this is just my general impression and I can’t promise you that this is 100% correct, but my impression is that a great deal of individual RCTs are funded by donors who are reasonably restricted with their grants. So it’s possible for them to say ‘I want to fund a study on water in a particular area, with a particular technology’; so that would be very restrictive. I could also imagine a funder providing a great deal of money for many RCTs somewhat related to - let’s say - water. So I think that is incredibly common as a model for false plunders: providing money for research.

I think it’s not a thing that makes a lot of sense for individual donors to be trying to do on their own. Depending on the study, studies can cost half a million dollars. So if you’re a donor with five thousand dollars, you’re not able to make a study happen with just that money and if you really want to give to this, you need to look to someone who can aggregate up the funds. Now this is something we’ve thought about doing. I think, down the line, GiveWell will do this. Essentially provide the mechanism by which individuals who want to provide to a larger pool of money that goes to things that those individuals couldn’t otherwise afford, we’ll provide the vehicle for them to do that - not something that exists right now.

All that said, I just want to take a step back. I think there’s a lot of reasons that research could be among the best things. At this stage in time, you know, I haven’t found any research that is right now obviously competitive with AMF. I think it’s important for anyone who’s thinking about this to not just think about the conceptual case that research could be good, but to try as hard as they can to nail down the specific organisation or the specific research that they want to do, and think about how that would trade off against giving to the top performing organisations.

There are two additional thing to be aware of when it comes to research. One is when we look at studies, often what we find is something goes wrong that diminishes the amount of information that the study can convey. So recently, we’ve been looking at some clean water trials. And this is a page that we’re trying to publish shortly. So these are studies of water purification technology. Basically, giving some people a filter, don’t give some other people the filter, then measure what happens to the water quality, measure what happens to health outcomes. In some of these studies, what happened is that the control group water filter actually causes the water to purify to some extent and this diminishes the power of the study to communicate results. One challenge of conducting studies is that you can conduct them, but there may be some problem in their implementation that reduces the impact.

Another issue to be aware of is that a single RCT is not particularly informative. And the programs that have the strongest evidence have multiple RCTs behind them. For example, malaria has tens of RCTs showing impact of nets on malaria, and so to get to the point where one has this body of knowledge that enables them to have a strong view about the impact of a program requires more than just funding a single trial.

All of that said, this remains an area that we, GiveWell, are very interested in considering, and I think it’s something we’ll be doing in the next few months. If we find the areas where research might have the biggest impact, that’s something we’re definitely considering trying to provide funding for.