Scores by kind of claim
Each kind of claim is scored on its own, out of 100: the share that held up, with the range it likely falls in. Fewer resolved claims give a wider range.
-
Factual claims
Statements about the present or the past, checked against the record: did they hold up?
76 63 88 -
Predictions
Statements about events outside the person's control, checked once they fell due: did they come true?
Insufficient record: 1 prediction resolved, 5 needed for a score.
-
Promises
Commitments to the person's or organisation's own future action, with something to deliver: was it delivered?
70 52 87
All types combined: 74 (likely 64 to 84, medium confidence), from 22 claims: 15 factual claims, 7 promises. It mixes kinds of claim in different proportions for different people, so compare people kind by kind.
GiveWell's claims
“We will make these grants if we believe they are more cost-effective than the opportunities to which we would otherwise direct funds.”
GiveWell, unsigned blog post 28 May 2020 GiveWell's plans for 2020 (The GiveWell Blog)
GiveWell will make additional COVID-19 response grants if it believes they are more cost-effective than the opportunities to which it would otherwise direct funds.
Mostly true True in substance, wrong in a detail.
The second judge would not resolve this claim; the verdict shown is the first judge's.
What would count as true or false
This is a conditional commitment. True: any further COVID-19 grants GiveWell made after 28 May 2020 were justified in its grant writeups by a judgement that they were more cost-effective than its alternative uses of funds, and there is no record of grants being made despite GiveWell judging them less cost-effective. Making no further grants is consistent with the promise if GiveWell explains it judged the opportunities not cost-effective enough. Mixed or mostly false: grants were made with a stated rationale other than comparative cost-effectiveness. False: GiveWell's own records show it made COVID-19 grants it described as below its cost-effectiveness bar.
Evidence
- Grantmaking in Response to COVID-19 GiveWell · March 2021 Own record
“We looked for grants that we thought would be as or more cost-effective than where we would have otherwise directed marginal funding.”
- Grantmaking in Response to COVID-19 GiveWell · March 2021 Own record
“These grants were part of GiveWell's work to identify COVID-19-related giving opportunities that might be comparable to our top charities in cost-effectiveness.”
- Innovations for Poverty Action — Randomized Controlled Trial on the Effect of Face Masks on COVID-19 GiveWell · February 2021 Own record
“The grant resulted from GiveWell’s work to identify giving opportunities related to COVID-19 that could be more cost-effective than support for our top charities.”
- Optimizing Your Pandemic Charity Michigan Public (Planet Money newsletter excerpt, NPR) · 5 May 2020
“And, Snowden says, that $2,000-per-life figure is just "a really high bar to clear." So Snowden says they're still recommending people donate to the same top charities they did before the coronavirus.”
Reasoning
GiveWell's grant writeups for COVID-19 grants made after 28 May 2020, including the July 2020 IPA mask RCT grant (E3) and the IDinsight grants (E2), justify them in comparative cost-effectiveness terms against its top charities. Its 2021 summary says it sought grants as or more cost-effective than its marginal alternatives (E1). No evidence shows a grant described as below its bar, but all post-claim evidence comes from GiveWell itself, and E3 notes the cost-effectiveness estimate was less rigorous than usual. E4 is independent but predates the claim and only shows the bar was high.
Cite this claim
GiveWell, “GiveWell will make additional COVID-19 response grants if it believes they are more cost-effective than the opportunities to which it would otherwise direct funds.” (said 28 May 2020). Verdict: mostly true. Reliability API, claim c_317d5b778c3a, method 0.1.0+7831dd9d. https://epistacklabs.com/claims/c_317d5b778c3a
Claims without a verdict
Is a verdict here wrong, or do you speak for GiveWell? Reply on the claim's own page, with the form at its foot (how replies work). Verdicts are made automatically; how verdicts and scores are made.