Announcing Rethink Priorities

This es­say was jointly writ­ten by Peter Hur­ford and Mar­cus A. Davis.

Re­think Char­ity is ex­cited to an­nounce our new pro­ject, Re­think Pri­ori­ties, which is ded­i­cated to do­ing foun­da­tional re­search on ne­glected causes in a highly em­piri­cal and trans­par­ent man­ner. This work will be­gin this year, be­gin­ning with a fo­cus on les­sons that can be taken from an­a­lyz­ing vac­ci­na­tions as an in­ter­ven­tion in the de­vel­op­ing world and an­i­mal welfare cor­po­rate cam­paigns.

This work will be led by Mar­cus A. Davis and Peter Hur­ford, with Mar­cus work­ing full-time and Peter work­ing part-time.

Mar­cus is a co-founder of Char­ity En­trepreneur­ship and Char­ity Science Health, where he sys­tem­at­i­cally an­a­lyzed global poverty in­ter­ven­tions, led cost-effec­tive­ness analy­ses, and over­saw all tech­ni­cal as­pects of the pro­ject. Be­fore join­ing the Char­ity Science team, he ran Effec­tive Altru­ism Chicago and worked with LEAN co­or­di­nat­ing out­reach to lo­cal EA groups around the globe.

Peter is a data sci­en­tist work­ing from Chicago. He co-founded Re­think Char­ity, and is on the board of Char­ity Science and An­i­mal Char­ity Eval­u­a­tors. He has re­viewed and pro­duced re­search on cause pri­ori­ti­za­tion and effec­tive al­tru­ism since 2013.

In do­ing this re­search, we fo­cus on the fol­low­ing prin­ci­ples:

  • Gen­er­at­ing new in­sight and knowl­edge in ad­di­tion to syn­the­siz­ing ex­ist­ing re­search. We’re fo­cused more on pro­duc­ing cut­ting-edge con­tent rather than sum­ma­riz­ing ex­ist­ing con­tent.

  • Pub­lish­ing shorter and more di­gestible in­for­ma­tion more fre­quently, rather than pub­lish­ing sprawl­ing re­search less fre­quently. By tak­ing the same amount of in­for­ma­tion and break­ing it down into “min­i­mal pub­lish­ing units,” we make it eas­ier for our­selves and oth­ers to un­der­stand and build upon, and get quicker feed­back loops.

  • Em­pha­siz­ing quick feed­back loops try­ing to go from ini­tial re­search to an ini­tial pub­li­ca­tion in about a month.

  • Seek­ing tractabil­ity in re­search by look­ing for ques­tions that we may be able to make mean­ingful progress on. We’ll hack away on the edges, so to speak.

  • Keep­ing all re­search in touch with the big­ger pic­ture, build­ing upon in­di­vi­d­ual pub­li­ca­tions to try to an­swer key ques­tions.

  • Find­ing new op­por­tu­ni­ties over cre­at­ing more ranked lists of ex­ist­ing op­por­tu­ni­ties. We pre­dict im­prov­ing on ex­ist­ing un­cer­tain rank­ing meth­ods is less effec­tive than find­ing di­rec­tions that have been pre­vi­ously ne­glected in EA. We’re un­likely to con­vince peo­ple of the num­ber one cause, but we might find some­thing oth­ers have not thought about.

  • Track­ing the im­pact of our re­search to help us make key de­ci­sions. We have plans to do fol­low-up re­views on all our re­search, in­clud­ing sur­veys to de­ter­mine the in­fluence of the re­search.

  • Shut­ting the pro­ject down if it doesn’t gen­er­ate im­pact to avoid wast­ing our time and effort. We are pre-com­mit­ting to an ini­tial six month test phase and pivot­ing as nec­es­sary. We will com­pletely shut down if there’s no dis­cern­able im­pact within the first year. If our work isn’t helping oth­ers make bet­ter de­ci­sions, we should try some­thing else.

Right now, our re­search agenda is fo­cused on:

  • try­ing to learn how to ap­ply cost-effec­tive­ness frame­works to un­cer­tain do­mains. (As an ex­am­ple, see our first post on vac­cine re­search and de­vel­op­ment.)

  • pri­ori­ti­za­tion and re­search work within in­ter­ven­tions aimed at non­hu­man an­i­mals (as re­search progress here looks uniquely tractable com­pared to other cause ar­eas)

  • as­sist­ing LEAN and SHIC in gath­er­ing ev­i­dence about EA move­ment build­ing (as re­search here looks tractable and ne­glected)

What we’ve done so far:

Our re­search agenda and ap­proach are still in the very early stages and may change sig­nifi­cantly as we grow and learn.