This is crossposted from Metaculus (I wrote the original question). If you'd like to forecast it, go here:
- Feel free to ask questions (here, there, on my twitter @nathanpmyoung)
- There are forecasting group chats and the metaculus discord if you'd like to forecast more, finding a group to do it with would be my top recommendation
Full text of question
GiveWell has recommended grants to over 10 charities over the years. They are currently investigating 14 charity areas including breastfeeding promotion.
The following sections are quoted from GiveWell’s explanation of the topic:
“The World Health Organization (WHO) and UNICEF recommend early initiation of breastfeeding, exclusive breastfeeding to 6 months, and partial breastfeeding to age 24 months to improve infant and maternal health, but the majority of infants are not fed according to these guidelines. Mothers may not breastfeed as long or as intensively as they would like due to a lack of skills and support, so various maternal counselling and support interventions promote breastfeeding behaviour change.”
Effectiveness for the following reasons: “We believe there is reasonably strong evidence that breastfeeding support programs can lead to increases in rates of exclusive breastfeeding up to 6 months (compared to some or predominant breastfeeding) and breastfeeding duration, as recommended by WHO/UNICEF. Their impacts on exclusive breastfeeding may be larger in low-income countries. There is additional evidence that increasing breastfeeding reduces diarrhoea morbidity, which likely leads to reductions in childhood mortality from diarrhoea. It may also cause additional benefits that we have not yet vetted.”
Will GiveWell recommend 1 or more grants to support breastfeeding promotion or denote it a "top charity" before 2027?
This question resolves positively if GiveWell publishes a recommendation that grants be given to support breastfeeding promotion or on or before December 31, 2026. Or if a breastfeeding promotion charity is designated a top charity before December 31, 2026.
This may happen in the following ways:
- GiveWell writes “yes” in the column “Have we recommended one or more grants to support this program?” in the “breastfeeding promotion” row of the GiveWell program reviews spreadsheet or the spreadsheet which supersedes this one (see below).
- Givewell publishes a blog in which which they announce they are recommending a grant in the area of breastfeeding promotion
- Givewell adds an incubation grant to a breastfeeding promotion charity on this page https://www.givewell.org/research/incubation-grants
- A GiveWell blog post announcing that a charity in the area of breastfeeding promotion is a “Top Charity”
- GiveWell publishes a section on breastfeeding promotion on its Top Charities page
Fine print
If GiveWell publishes a recommendation that is no longer in force in 2027, this question resolves positively. Eg GiveWell no longer recommends grants to GiveDirectly, but it once did.
If GiveWell revokes a recommendation due to making errors in its calculations, this question will resolve ambiguously.
The category “standout charities” used to exist. It no longer does, so can be ignored.
If the GiveWell program reviews spreadsheet is no longer linked from this page, a superseding spreadsheet may be used. A superseding spreadsheet should have very similar column headings or the same title (“GiveWell programme reviews) or be linked under “prioritized list of programmes” from this page or failing that, be the consensus choice of admins commenting on the question.
For clarity, if the above criteria aren’t met, but a reasonable spreadsheet exists which contains a grant for breastfeeding promotion exists, admins can post their suggestions of the spreadsheet in the comments of this question. Admins can change their votes, but at the time of resolution, all admins should have commented in favour of the same spreadsheet. If they have not, and the question does not resolve positively by any of the other methods, the question resolves ambiguously.
Some thoughts
- Could I have written this question better? If so how?
- I am interested in how forecasts can support researchers. If I do this a lot, it will take me about an hour per question. If that provides useful information to researchers it could be very cheap
- eg If 3 of the cause areas have much higher % forecast, then GiveWell might choose to look at them first. If this is the right call, it could be a useful prioritisation
- I think this question is harder to forecast than "will breastfeeding support be a top charity?". Someone at Givewell I spoke to said "any grants" would be more useful to them, so I went with it. But a more complex question means that forecasters might try and forecast the GiveWell's actions than the impact of the cause area which would be less useful.
- If this gets a lot of takeup I intend to write a question for all of Givewell's potential cause areas (there are about 15)
- Are there EA questions you think could be usefully forecasted?
Hi, everyone! I think the question here is focused on the higher order questions of what GiveWell will recommend (and one level higher) how to ask forecasting questions about it. But, to not answer that, I will say that at r.i.c.e., our fieldwork in India is, in the sense of a majority of our time, effort and staff, mostly spent on lactation consulting and Kangaroo Mother Care for low birthweight babies right now. (One question for the author might be whether a KMC program, which is fundamentally about breastfeeding but also has other mechanisms (keeping the baby warm, keeping the caregiver informed about the baby's status) would could as breastfeeding promotion for this forecast.)
We're working in Uttar Pradesh, which is a context where women's modesty is an important cultural issue. But one challenge we are facing is demand: it's hard for us to get mothers and families more broadly to want to go along with our program. I don't think this should have surprised us so much, but in fact the bindingness of this constraint did surprise us. I would be very excited to hear about practical research, ethnography, or just trial and error anyone is doing about this.
From a prioritization point of view, this sort of approach competes against more resource-intensive neonatal care, such as investing in incubators (which may or may not be plugged in and turned on) which work against establishing breastfeeding by separating the moms and babies. So, another indirect way of doing "breastfeeding promotion" may be to lean against that sort of medicalized policy.
Like so many things for early-life health in poor populations (open defecation, clean cooking fuel use,...), this is an area where the "second stage" biological mechanism seems very well established to me, and what we need is good evidence and strategies for a better "first stage" effect of programs on behavior change.
Thanks Dean, love to hear from an expert.