Should I give to Our World In Data?

by Cullen_OKeefe1 min read10th Sep 201910 comments


Donation choiceOur World in DataCommunity

I’m considering making a donation to Our World in Data, which is currently fundraising. My thinking is that:
1. I haven’t seen any EAs (except Steven Pinker) donate or grant here so it might be an overlooked/undervalued opportunity. Besides Pinker and the Bill and Melinda Gates Foundation, I didn’t recognize any of the donors on their donor list.
2. EAs frequently use this as a source that is high-quality in both substance and presentation and focuses on important issues. They have an explicit focus on doing work that neglected. Expanding their work could therefore be useful as source for both substantive EA work and movement-building. I also think there might be something of a network effect of having a bunch of macroeconomic data in one place.

Reasons against:
1. Same as 1 above: there might be a good reason for no grants to-date. If the EA donor "market" is efficient, a good opportunity like this would have been taken, especially since the site is well-known.
2. Unclear what the effect size of a donation is. I asked Max Roser on Twitter and hope he’ll answer.
3. The data collected and published might not be particularly important on the margin. If you look at the latest posts, the importance of the topics varies pretty widely. But overall I think the new work is still impressively important on average!

Assume the donation size in consideration is between $1,000-$10,000.

10 comments, sorted by Highlighting new comments since Today at 4:18 PM
New Comment

Note that their sponsors now include Longview Philanthropy, Effective Altruism Infrastructure Fund, and some EA-adjacent folk such as Paul Graham, Vitalik Buterin and Patrick Collison. The EA Infrastructure Fund report from November 2019 offers the following justification for supporting Our World In Data:

  1. In a world where misinformation is becoming an increasingly bigger problem, we see some considerable upside in funding organisations trying to make it more likely that decision-makers act based on truthful information. Often, this data exists but is difficult to access and interpret. Our World in Data (OWID) aims to address this by publishing clear, relevant, open-source analysis on a range of important issues. It targets a large audience and ranks high in relevant Google searches, entirely through organic growth.
  2. OWID has some evidence that there is a trend towards even senior decision-makers ‘asking Google’ where they would have previously had an aide do a research project. If this is true, then OWID’s goal of getting compelling, ready-to-use data and charts high in search rankings seems potentially valuable for aiding evidence-based decision making.
  3. Much of OWID’s work is directly relevant to effective causes; for example, see its reports on technological progress, genome sequencing, peacekeeping, meat production, mental health, and smoking. Given how much we've seen OWID data used in research focused on solving high-impact problems, we saw this as an additional reason to support its work.
  4. OWID’s audience is still growing very quickly. The site had over 1 million unique users last month, double its average monthly visitors for last year. OWID is among the top Google results for a number of keyword searches, such as CO2 emissions and extreme poverty. 60% of users find OWID through organic search. OWID is referenced extensively in media and education; see its data on audience and coverage here.
  5. OWID also aims to counter society’s fatalism, prompt appreciation for achieved progress, and inspire us to work toward further advancement. Although success here is hard to measure, the group seems to be making some progress. OWID was listed as source for 25 out of 75 graphs in Steven Pinker’s Enlightenment Now, and Bill Gates tweeted the cleverly framed ‘World as 100 People’ infographic to much applause. More about OWID’s motivation is here.
  6. The OWID team is surprisingly small. We had assumed the group had a large research and support staff, but it has only 6.5 FTE equivalent. The team went through YCombinator last year and they seem highly capable and experienced.
  7. OWID is fundraising £500,000 for 2020, which will allow it to expand its team, particularly to hire more developers (it currently has only two). OWID seems to have accomplished a lot with a very small team, and we expect that growing the team further will have high additionality. Projects we are excited for in 2020 include a deep report on the history of war, the development of the OWID grapher (an open-source tool that anyone can use to make their own visualizations), and the plans to transform the OWID database into a fully searchable and manipulatable library.

I was personally very surprised to learn that, back when the report was written, OWID had only ~6 full-time employees. A quick-and-dirty comparison of their current Our Team page with the archived version from December 2019 suggests that staff has since grown by a factor of 18 / 8 = 2.25. I still get the impression that the value per employee is pretty high, and I'd imagine they are primarily constrained by funding, so giving to them looks like a promising opportunity to me. I would be interested to hear other people's opinions, especially from those who don't think donations to OWID are unusually impactful.

I haven’t seen any EAs (except Steven Pinker) donate or grant here so it might be an overlooked/undervalued opportunity. Besides Pinker and the Bill and Melinda Gates Foundation, I didn’t recognize any of the donors on their donor list.

It seems strange to use the fraction of donors who are EAs as a negative signal, holding their total fundraising constant. Surely the information about implicit EA cost-effectiveness estimates is dominates the crowding out effect? In equilibrium this consideration would suggest that EA giving should be equally spread among all groups!

And if this was valid, you might not want EA opinions on the subject. Would you not then equally take persuasive EA arguments for giving there to be a reason to think it is not overlooked/undervalued?

Yes, on balance I suspect you’re right. But I also think individual donors should be looking for black swan opportunities.

Then you'd be looking for something with no donors. (Or nearly none.) The percent of them who were EAs wouldn't be relevant.

I’m not sure that’s entirely true. The ones who don’t care about effectiveness make the opportunity less neglected, but it still might be neglected relative to its effectiveness if EAs haven’t noticed the potential there.

Ah, like, if EAs have already discovered it, then you expect it to quickly reach saturation, efficient market style?

Two positions at Our World in Data are listed on the 80,000 Hours job board.

Did you ever hear back from Max Roser? I'd be interested to hear his thoughts, as the donation page on the website is bare-bones, with no information on runway, projects they might pursue or not pursue based on certain funding targets, etc.

Nope. I might follow-up by email.