A guide to effective altruism fellowships

In­tro­duc­tion to the guide

This doc­u­ment is based on three semesters’ worth of ex­pe­rience run­ning effec­tive al­tru­ism fel­low­ships at Yale Univer­sity, though most of what I write here re­lates to the Fall 2018 semester, which was novel in sev­eral im­por­tant ways. The doc­u­ment is in­tended to help groups run suc­cess­ful fel­low­ships or similar pro­jects. The doc­u­ment can ei­ther be read in its en­tirety or used as a en­cy­clo­pe­dia. The ap­pen­dices are com­pre­hen­sive and are es­pe­cially use­ful for group or­ga­niz­ers. For any ques­tions, don’t hes­i­tate to reach out to joshua.mon­rad@yale.edu. If you wish to reach Yale Effec­tive Altru­ism (YEA) in or­der to col­lab­o­rate, get in­volved or in­quire about any­thing, con­tact yaleeas@gmail.com.


The Yale Effec­tive Altru­ism Fel­low­ship is a semester-long pro­gram, in which 15-16 peo­ple are se­lected to learn a great deal about effec­tive al­tru­ism over a few months. The fel­lows sub­mit a writ­ten ap­pli­ca­tion and are in­ter­viewed prior to ac­cep­tance. The fel­lows meet once a week for din­ner, where they dis­cuss a range of EA top­ics. Be­fore each meet­ing, the fel­lows are as­signed be­tween 30 and 45 min­utes worth of read­ings, pod­casts or videos. Din­ner dis­cus­sions are mod­er­ated by a group leader and usu­ally do not in­volve pro­fes­sors. Beyond the din­ner dis­cus­sions, the YEA Fel­low­ship in­volves an im­pact challenge, wherein the fel­lows are tasked with de­cid­ing where to donate a sig­nifi­cant amount of money. The fel­low­ship also in­cludes a cou­ple of work­shops, so­cial events, and op­tional events with speak­ers and pro­fes­sors. Fel­lows were also in­vited to take part in or­ga­niz­ing our GWWC pledge drive.

Some keys to a suc­cess­ful fel­low­ship seem to be:

  • Only ad­mit­ting the right peo­ple, who are com­mit­ted and gen­uinely interested

  • Set­ting high ex­pec­ta­tions for fellows

  • Mak­ing the fel­low­ship seem and feel like a se­ri­ous and well-or­ga­nized program

  • Only in­volv­ing pro­fes­sors who are ex­cep­tion­ally good speak­ers and/​or speak on top­ics that are di­rectly re­lated and al­igned with effec­tive al­tru­ism ideas

  • Do­ing enough out­reach that you can get a large amount of applicants

  • Plan­ning well in ad­vance of the fellowship

If you plan to do a mas­sive Giv­ing Game, it might be good to:

  • Fo­cus on com­par­ing or­ga­ni­za­tions within one cause area (e.g. global health)

  • Let fel­lows choose from a pre­de­ter­mined list of can­di­dates, in­stead of start­ing from scratch

  • Have one or more ses­sions of col­lab­o­ra­tive, in-depth research

  • Fo­cus on teach­ing fel­lows to eval­u­ate, rather than pro­duce, re­search on effectiveness

In the 2018 Fall Fel­low­ship, we ac­cepted 16 out 65 ap­pli­cants. We had an as­ton­ish­ingly high level of re­ten­tion and at­ten­dance through­out the 12 weeks of weekly din­ner dis­cus­sions and work­shops. Among them, our fel­lows spent, in to­tal, more than fifty hours re­search­ing and de­liber­at­ing as part of the im­pact challenge and ul­ti­mately de­cided to al­lo­cate the $3,000 dona­tion (pro­vided by CEA) to the Against Malaria Foun­da­tion. One of the fel­lows took the GWWC pledge dur­ing the fel­low­ship, with sev­eral oth­ers in­di­cat­ing that they strongly con­sid­ered it. Fel­lows’ views on top­ics in effec­tive al­tru­ism up­dated in a pos­i­tive di­rec­tion dur­ing the fel­low­ship. (See ap­pen­dices for re­sults of post-fel­low­ship eval­u­a­tion sur­vey.)


Although I (Joshua) wrote this doc­u­ment and for­mally was in charge of our fel­low­ship for the past cou­ple of semesters, it would be both wrong and dishon­est if I failed to rec­og­nize the hard work of other YEA board mem­bers; in par­tic­u­lar, Frankie An­der­sen-Wood, Jes­sica McCurdy, and Mo­jmír Stehlík, each of whom have been es­sen­tial in mak­ing the fel­low­ship what it is. I should also ex­press our deep­est grat­i­tude to CEA for their mon­e­tary and strate­gic sup­port.

Goals of the fel­low­ship

The fel­low­ship was mainly aimed at peo­ple who did not have a lot of prior fa­mil­iar­ity with effec­tive al­tru­ism. For those, more ex­pe­rienced, peo­ple, we had other events and ac­tivi­ties such as a weekly dis­cus­sion group.

We had sev­eral goals with the fel­low­ship. Look­ing back, I re­call these be­ing the main ones:

  • Bring­ing mem­bers into our “fun­nel.” The fel­low­ship ap­pears to be a great way of at­tract­ing po­ten­tial core mem­bers and we hoped that at least a cou­ple of fel­lows would go on to join our board and that even more would re­main mem­bers of the gen­eral EA com­mu­nity on cam­pus.

  • Direct im­pact of af­fect­ing fel­lows’ views. 12 weeks isn’t a lot when it comes to deeply af­fect­ing some­one’s views, but we hoped that we could at least teach the ba­sic tenets and con­cepts of effec­tive al­tru­ism to al­most all fel­lows and that this could af­fect dona­tion and ca­reer de­ci­sions for a good por­tion of them. More­over, we hoped that this might leave a huge im­pres­sion on 1-2 fel­lows, who could at some point dras­ti­cally change ca­reer paths as a re­sult of be­ing in­tro­duced to EA. We as­sumed that a large part of the ex­pected value of the fel­low­ship came from this pos­si­bil­ity.

  • Strength­en­ing YEA as a promi­nent group on cam­pus. The fel­low­ship acts as a “high-pro­file” ac­tivity, which can be ad­ver­tised widely and which board mem­bers can high­light when asked “So, what does your group do?” Fur­ther­more, the fel­low­ship can be­come well-known on cam­pus through word-of-mouth and at­tract at­ten­tion to the group.

Re­cruit­ment of fel­lows

Reach­ing the right num­ber and the right type of peo­ple will be es­sen­tial to a good fel­low­ship.


How to prop­erly ad­ver­tise the fel­low­ship. One of the most im­por­tant things I ever learned about run­ning the fel­low­ship is this: You get what you ad­ver­tise for. Be­fore the Spring 2018 semester, I would of­ten ad­ver­tise the pro­gram by say­ing things like “It’s not a huge time com­mit­ment, you just have to come for some din­ners, and you’ll be eat­ing any­way right?” and “It’s definitely NOT like an ex­tra class!” My rea­son­ing was that I wanted to bring in more ap­pli­cants. The prob­lem is that if you say that the fel­low­ship doesn’t re­quire a lot of com­mit­ment, you’ll get ex­actly what you asked for: peo­ple who aren’t very com­mit­ted.

In Fall 2018, I in­stead said things like “We’re look­ing for peo­ple who will be re­ally en­gaged” and I would make it clear that the fel­low­ship would re­quire 4-6 hours a week. This made it so that peo­ple who weren’t that ex­cited would self-se­lect away be­fore the ap­pli­ca­tion pro­cess, mak­ing it much more likely for us to ad­mit very en­gaged fel­lows. (At the same time, you don’t want to go too far. Col­lege stu­dents are busy and they need to know that the fel­low­ship can fit into their weekly sched­ules.)

Chan­nels for ad­ver­tise­ment. We had a two-pronged strat­egy for re­cruit­ment: broad & tar­geted.

      • Broad re­cruit­ment: We aimed to make sure most first-years (at least) knew about the Fel­low­ship, through a broad re­cruit­ment strat­egy. We hoped that well-thought-out mes­sag­ing (ac­cu­rate—not just fo­cus on dona­tions) could elimi­nate the risk of spread­ing low-fidelity con­cep­tions of EA. The rea­son­ing was that if we achieved this goal, hope­fully only peo­ple who were truly in­ter­ested in EA ideas would ap­ply, and the se­lec­tion pro­cess would be much eas­ier. I be­lieve that was the case.

    • Posters on cam­pus (See Ap­pendix 1 for Fall 2018 poster de­signed by Frankie An­der­sen-Wood)

    • Website

    • Face­book posts to peo­ple who ‘like’ our page, posts in rele­vant groups, & tar­geted ads

    • Fresher’s fair/​ex­tracur­ricu­lar bazaar, handed out fly­ers in Ap­pendix 1

    • E-mails to spe­cific newslet­ters (es­pe­cially to rele­vant de­part­ments and grad­u­ate schools such as eco­nomics, com­puter sci­ence, philos­o­phy, school of man­age­ment, etc.)

    • Word of mouth

      • Tar­geted re­cruit­ment: refer­rals. For fall 2018, we also made a “Refer-a-friend” form that we sent to former fel­lows, where they could in­put an email ad­dress and we’d con­tact their friend about ap­ply­ing. We aimed that all former fel­lows and board mem­bers are aware of the refer­ral form and we would re­ceive at least ten high qual­ity refer­rals.

How large should a co­hort of fel­lows be?

In Fall 2018, we opted for a co­hort size of 16 fel­lows. This seems to be a re­ally good size for dis­cus­sions, al­low­ing for talk­ing time to ev­ery­one (if well mod­er­ated). In the past, we’ve tried with slightly larger co­horts (18-20), rea­son­ing that some peo­ple in­evitably would drop out and bring us to the ideal size of 15-16. How­ever, if 20 peo­ple show up for the first dis­cus­sion, it will be very crowded and leave peo­ple with a poor first im­pres­sion of dis­cus­sion qual­ity. More­over, if you suc­ceed in se­lect­ing only very com­mit­ted peo­ple (see Ad­ver­tise­ment and Selec­tion) and run­ning a re­ally good fel­low­ship (see Run­ning the fel­low­ship), you can achieve high at­ten­dance and re­main at the ini­tial co­hort size.

One or two co­horts?

If you have many ap­pli­ca­tions, it is cru­cial to only ac­cept the peo­ple who seem like they’re are very com­mit­ted and a good fit for the pro­gram.

In ear­lier semesters, when we started re­ceiv­ing 40+ ap­pli­ca­tions, we were re­luc­tant to turn any­one away who had an in­ter­est in EA. In­stead, we chose to ac­cept al­most all of them into two co­horts of 20 peo­ple. The down­side to this was that, in each group, only half of them were re­ally com­mit­ted and/​or a good fit for the fel­low­ship, lead­ing to de­clin­ing at­ten­dance and qual­ity of dis­cus­sions.

In Fall 2018, we re­sisted the temp­ta­tion to take ev­ery­one and de­cided in­stead to go for just one co­hort. (Later on, we cre­ated a new co­hort speci­fi­cally for grad­u­ate stu­dents, which was a dis­tinct and less ex­ten­sive pro­ject, based on weekly dis­cus­sions of an as­signed read­ing.) This definitely seemed like the right move, as it al­lowed us to get a co­hort en­tirely com­prised of en­gaged fel­lows. Re­jected ap­pli­cants were still strongly en­couraged to get in­volved with our com­mu­nity, though did so to a dis­cour­ag­ingly low de­gree.

This isn’t to say that an EA group couldn’t run mul­ti­ple fel­low­ship co­horts at once, only that you shouldn’t ac­cept ev­ery sin­gle ap­pli­cant. It seems plau­si­ble that 30-40% of ap­pli­cants (at Yale at least; rate may be higher or lower el­se­where) will be a good fit. So, if you wind up with 100 ap­pli­ca­tions, you might have enough good can­di­dates to fill two co­horts. Just make sure that you also have the ca­pac­ity to run two pro­grams, then.

Selec­tion process

We re­ceived 65 ap­pli­ca­tions for 16 spots. Once the dead­line had passed, we (four peo­ple from the lead­er­ship of YEA) met. Each of us were as­signed 25% of the ap­pli­ca­tions, which we then read and graded ac­cord­ing to se­lec­tion crite­ria (see be­low). We blinded the names of the ap­pli­cants to miti­gate bi­ases and con­flicts of in­ter­est. After ev­ery­one had graded their por­tion, we ranked the ap­pli­cants from high­est to low­est score. Then, we went through from the bot­tom up and de­cided on the 20 least com­pet­i­tive ap­pli­cants, whom we weren’t go­ing to in­ter­view. We talked each ap­pli­cant over, just to make sure that the se­lec­tion wasn’t just a func­tion of one grader hav­ing differ­ent stan­dards than the oth­ers.

Once that was done, we in­vited about 40-45 peo­ple for 30-minute in­ter­views. I would strongly en­courage you to do in­ter­views, even if you don’t have that many ap­pli­cants. For one, it gives you a much bet­ter pic­ture of the ap­pli­cant pool and how much they would gain from the fel­low­ship. Se­cond, it sends a strong sig­nal to prospec­tive fel­lows that the fel­low­ship is a se­ri­ous thing with high stan­dards.

For the in­ter­view, we asked them to read the In­tro­duc­tion to EA from CEA and asked them to “come pre­pared to share your re­ac­tions, dis­agree­ments or un­cer­tain­ties.”

Dur­ing the in­ter­view, we dis­cussed the ar­ti­cle and asked ques­tions which we thought might help us rate them ac­cord­ing to our se­lec­tion crite­ria.

Note: Young peo­ple take in­ter­views very se­ri­ously and may be a lot more ner­vous than you think they are. It’s in­cred­ibly im­por­tant to be kind, pa­tient, and en­courag­ing. In­timi­dat­ing them will be bad for them, for your group, and for the rep­u­ta­tion of EA as a whole. Also, be mind­ful that peo­ple have differ­ent styles of think­ing and talk­ing, and it’s im­por­tant not to be dis­mis­sive of those styles which are differ­ent than yours.

Selec­tion criteria

When eval­u­at­ing ap­pli­cants for the Fall 2018 co­hort, we rated them on the ba­sis of a set of crite­ria that had also been out­lined on our ap­pli­ca­tion:

    • Altru­ism: Pas­sion­ate about helping others

    • Effec­tive­ness: Am­bi­tious in their al­tru­ism, with a drive to do as much good as they can. Po­ten­tial to be al­igned with the cen­tral tenets of EA.

    • Po­ten­tial: Ex­cited to ded­i­cate their ca­reer to do­ing good or to donate a sig­nifi­cant por­tion of their in­come to charity

    • Open-mind­ed­ness: Open-minded and flex­ible, ea­ger to up­date their be­liefs in re­sponse to per­sua­sive evidence

    • En­thu­si­asm: Willing and able to com­mit ~3-4 hours per week

    • Fit: How good a fit are they with the fel­low­ship for­mat? Will they be good in dis­cus­sions? Will they do good work for the Im­pact Challenge?

The ques­tions we used to judge these crite­ria can be found in Ap­pendix 11 (the Fel­low­ship Ap­pli­ca­tion.)

We rated the ap­pli­cants from 1-5 on each of these crite­ria. We dis­cussed some of the ap­pli­cants to­gether and shared how we’d each rate them, to en­sure that we had similar stan­dards.

Ul­ti­mately, the se­lec­tion pro­cess was very time in­ten­sive and ex­tremely difficult. (It should be noted that our abil­ity to in­ter­view 40+ was very much a product of the fact that we had some­one work­ing full-time, on a CEA grant, on run­ning our group dur­ing this semester.) The end re­sult was cer­tainly ac­cept­able and of all the 16 ad­mit­ted fel­lows, there was only one per­son whom I maybe would have preferred to sub­sti­tute for an­other ap­pli­cant. Still, it is clear that there’s some room for im­prove­ment and I’d be happy to hear if some­one knew of a more sys­tem­atic ap­proach to ad­mis­sions.

Other im­por­tant con­sid­er­a­tions: Diver­sity and ho­mo­gene­ity in lev­els of ex­pe­rience


Beyond the above crite­ria, I be­lieve it is im­por­tant to con­sider two kinds of di­ver­sity when cre­at­ing a fel­low­ship co­hort.

First, di­ver­sity of race, gen­der, and so­cioe­co­nomic back­ground should be se­ri­ously con­sid­ered. The EA move­ment gen­er­ally has a huge prob­lem with di­ver­sity. Among about 1,000 re­spon­dents in the 2017 EA Sur­vey, al­most 90% iden­ti­fied as white; less than 1% iden­ti­fied as black; and 70% iden­ti­fied as male. The full causes, con­se­quences, and reme­dies for this lack of di­ver­sity are se­ri­ous but be­yond the scope of this piece. In­stead, I will briefly note that the com­mu­nity could stand to benefit a great deal from more di­ver­sity and that this should be kept in mind when putting to­gether a co­hort of fel­lows.

One (small) way of ad­dress­ing is, for in­stance, to have non-male or non-white mem­bers of your group ad­ver­tise the fel­low­ship pub­li­cly. In our group, we are so for­tu­nate to have sev­eral women in our lead­er­ship (as of Spring 2019, both Co-Pres­i­dents are fe­male), and they did (and do) a great job at mar­ket­ing the fel­low­ship in a wel­com­ing and in­clu­sive way.

We also made sure to in­clude an en­courage­ment to ap­ply, re­gard­less of one’s iden­tity, on the ap­pli­ca­tion page.

Another im­pli­ca­tion of this is­sue is to be care­ful about judg­ing ap­pli­cants too much by their prior fa­mil­iar­ity with EA jar­gon or ter­minol­ogy. Some peo­ple may have great po­ten­tial to be­come core mem­bers of the com­mu­nity and do great things, even though they have not yet been in­tro­duced to the “ra­tio­nal­ist ver­nac­u­lar” or haven’t hap­pened to hear about Peter Singer when do­ing de­bate in high school. Aim to eval­u­ate can­di­dates for their po­ten­tial and their pas­sion, not their cur­rent level of fa­mil­iar­ity with EA.

Se­cond, it is a good idea to get a di­verse range of aca­demic back­grounds rep­re­sented in the co­hort, if pos­si­ble. The dis­cus­sions benefit a great deal from con­tri­bu­tions from a va­ri­ety of dis­ci­plines.

Ho­mo­gene­ity in lev­els of experience

On the other hand, there is one area where ho­mo­gene­ity is de­sir­able: lev­els of fa­mil­iar­ity and in­volve­ment with the EA move­ment. While some het­ero­gene­ity is OK, it will be a dis­ad­van­tage for ev­ery­one con­sid­ered if there are some fel­lows who are vet­er­ans in the EA com­mu­nity while oth­ers are to­tal new­com­ers. Since our goal for the fel­low­ship was to serve as an early stage of the “EA pipeline,” we erred on the side of get­ting new­com­ers. That said, we do also con­sider the fel­low­ship a vi­able op­tion for peo­ple who are some­what fa­mil­iar with the move­ment, but who haven’t yet had the chance to re­ally study it, such as newly joined mem­bers of our board. We had a few ap­pli­cants who were ob­vi­ously overqual­ified, whom we en­couraged to join our board in­stead of do­ing the fel­low­ship.

Put­ting to­gether a pro­gram and curriculum

Con­tent of the Fellowship

The fall 2018 fel­low­ship con­tained the fol­low­ing el­e­ments:


  • Weekly din­ner dis­cus­sions from 6pm-7.15pm (10 meet­ings)

  • A so­cial ini­ti­a­tion event on the first Fri­day evening of the fellowship

  • A 3-hour in­tro­duc­tory work­shop on EA con­cepts dur­ing the first week­end of the fellowship

  • A 3-hour Re­search-a-thon


  • Din­ners with pro­fes­sors and guest speakers

  • YEA so­cial events

  • Other YEA pub­lic events

  • A work­shop on ca­reer strategies


In my view, the fel­low­ship should give an in­tro­duc­tion to sev­eral im­por­tant top­ics in EA, but shouldn’t nec­es­sar­ily at­tempt to dive deep in to any one of them. In 2018 Fall, the dis­cus­sion top­ics were as fol­lows:

Dis­cus­sion #1: Meth­ods and Ap­proach of Effec­tive Altruism

Dis­cus­sion #2: Effec­tive­ness in Global Health and Devel­op­ment

Dis­cus­sion #3: Im­pact Challenge I

Dis­cus­sion #4: Im­pact Challenge Dis­cus­sion II. Con­clu­sion of Im­pact Challenge.

Dis­cus­sion #5: The Philo­soph­i­cal, Em­piri­cal, and His­tor­i­cal Case for a Long Term Focus

Dis­cus­sion #6: Global Catas­trophic Risks from AI, Biose­cu­rity, and More

Dis­cus­sion #7: The Mo­ral Case for An­i­mals & An­i­mal Welfare Interventions

Dis­cus­sion #8: Ca­reer Choices and Effec­tive Altruism

Dis­cus­sion #9: Effec­tive Altru­ism Com­mu­nity Build­ing and Liv­ing As An Aspiring Effec­tive Altruist

Dis­cus­sion #10: Wrap­ping Up the Fel­low­ship and Dis­cussing Re­main­ing Issues

Too lit­tle fo­cus on the long term? You might no­tice that only 2 out of 10 top­ics are di­rectly re­lated to ex­is­ten­tial risks and “long-ter­mism.” This might seem at odds with the “main­stream” view within EA. (Whether this re­ally is the main­stream view, and whether it should be, are perfectly in­ter­est­ing ques­tions that are be­yond the scope of this piece.)

Here’s the rea­son­ing be­hind this de­ci­sion:

A lot of peo­ple in the EA com­mu­nity (in­clud­ing my­self and other stu­dent group lead­ers as well as promi­nent peo­ple such as Will MacAskill and Toby Ord) be­gan their in­volve­ment with the move­ment with a fo­cus on global poverty and health. Grad­u­ally, many tran­si­tion into car­ing more and more about the long term tra­jec­tory of hu­man­ity. I don’t think the or­der of this tran­si­tion (global de­vel­op­ment then long-ter­mism) is co­in­ci­den­tal and I do think it’s ex­tremely im­por­tant to keep in mind when try­ing to bring new peo­ple into the com­mu­nity.

Let’s face it. Long-ter­mism is not very in­tu­itively com­pel­ling to most peo­ple when they first hear of it. Not only do you have to think in very con­se­quen­tial­ist terms, you also have to be ex­tremely com­mit­ted to act­ing and pri­ori­tiz­ing on the ba­sis of fairly ab­stract philo­soph­i­cal ar­gu­ments. In my view, that’s just not very ap­peal­ing—some­times even off-putting—if you’ve never even thought in terms of cost-effec­tive­ness or to­tal-view con­se­quen­tial­ism be­fore.

Ac­cord­ingly, I be­lieve that start­ing off with ab­stract ar­gu­ments for fu­ture gen­er­a­tions or apoc­a­lyp­ti­cal risks from AI has a very high like­li­hood of turn­ing peo­ple off from EA—peo­ple, who could come to sup­port those very same causes down the line, if only they stuck around long enough to be­gin un­der­stand­ing and agree­ing with the core prin­ci­ples that lead to those views.

On the other hand, by start­ing with global health and de­vel­op­ment—which of­ten is com­pel­ling even to new­com­ers—peo­ple get used to the method­ol­ogy and philo­soph­i­cal as­sump­tions of effec­tive al­tru­ism. Once you start think­ing “Hm, maybe I should care about peo­ple, even though they are far away in space and maybe I should have a max­i­miz­ing ap­proach” then you’re one step closer to think­ing the same thing about peo­ple far away in time as well.

In sum: I’m not say­ing that it shouldn’t be a pri­or­ity to get peo­ple to care about long-ter­mism and x-risks. But I be­lieve pretty firmly that, for most peo­ple, the path to that point goes through more “con­ven­tional” cause ar­eas such as global health and de­vel­op­ment.

Ob­vi­ously, this is up for dis­cus­sion and ul­ti­mately a judge­ment call. But I should add that my “model of EA-new­comer psy­chol­ogy” is based on not only my own ex­pe­rience and the anec­do­tal ex­pe­rience of EA thought lead­ers, but also on the ex­pe­rience of talk­ing to tens, if not hun­dreds, of peo­ple about effec­tive al­tru­ism and see­ing what they re­spond well to.


I think it is rea­son­able to as­sign be­tween 30-90 min­utes of read­ings for each week, es­pe­cially if you have a co­hort of very en­gaged peo­ple. Read­ings should be in­for­ma­tive and use­ful as talk­ing points dur­ing dis­cus­sion. They shouldn’t re­quire too much former fa­mil­iar­ity with a sub­ject.

See Ap­pendix 2 for the as­sign­ments for the Fall 2018 co­hort.

In­volve­ment of pro­fes­sors

For the Fall 2017 and Spring 2018 fel­low­ships, I was con­vinced that it was es­sen­tial to the fel­low­ship that it in­cluded weekly din­ners with pro­fes­sors. I had two main rea­sons for be­liev­ing this, but I no longer be­lieve in ei­ther of them.

First, I thought that well-known names, such as Paul Bloom and Lau­rie San­tos, would at­tract more ap­pli­cants. While this may be true, you can still ob­tain that effect with­out hav­ing to have pro­fes­sors for ev­ery din­ner. More­over—as dis­cussed above—it’s not just a mat­ter of get­ting more ap­pli­cants. You want more ap­pli­cants of the right kind—i.e. peo­ple who are in­ter­ested in the ideas of EA, not just look­ing to meet fa­mous pro­fes­sors.

Se­cond, I thought that pro­fes­sors would make for more in­ter­est­ing con­ver­sa­tions. This can be true, if it’s an amaz­ing pro­fes­sor or if you have a poor group of fel­lows. How­ever, not all pro­fes­sors are amaz­ing and hope­fully you have a great group of fel­lows.

But the biggest is­sue with pro­fes­sors is prob­a­bly that it’s rare that their work is di­rectly re­lated to EA and some­times their per­sonal prin­ci­ples are at odds with core EA ideas. You might think that a pro­fes­sor in global health would be perfect for an EA dis­cus­sion. But that won’t be the case if they fo­cus on niche re­search in a niche (not-so-EA) area or per­haps haven’t done much work on cost-effec­tive­ness, cause pri­ori­ti­za­tion etc. This means that while con­ver­sa­tions with the av­er­age pro­fes­sor might be su­per in­ter­est­ing, they will only be tan­gen­tial to an ed­u­ca­tion in EA. Most of the time, the fel­lows will learn much more about EA ideas if they have a free con­ver­sa­tion mod­er­ated by some­one who’s very fa­mil­iar with the move­ment.

If you do want din­ners with pro­fes­sors...

All of this said, pro­fes­sors can add tremen­dous value to your group. For this rea­son, we de­cided to have a se­ries of “Din­ners with Pro­fes­sors” par­allel to the fel­low­ship, which was open to board mem­bers and fel­lows. If you run these, here are some helpful tips.

Pick pro­fes­sors care­fully and never bring a pro­fes­sor with­out meet­ing them in per­son first

Some pro­fes­sors are sur­pris­ingly awful at giv­ing talks and lead­ing dis­cus­sions. Others seem like they do in­ter­est­ing work, but then it turns out that their opinions are an­ti­thet­i­cal to core effec­tive al­tru­ism ideas. Few semesters ago, we man­aged to bring a pro­fes­sor who was an awfully dull speaker and was con­vinced that it was ridicu­lous to even be the slight­est wor­ried about AI al­ign­ment. To be sure, it can be could to have pro­fes­sors push back on main­stream EA think­ing, but it’s not very use­ful if they don’t even agree with the very ba­sic prin­ci­ples that make EA what it is.

As­sign a read­ing or video

Dis­cus­sions are usu­ally bet­ter if peo­ple are on the same page in terms of the topic of dis­cus­sion. Talk to the pro­fes­sor about choos­ing a good as­sign­ment that isn’t much longer than 20 min­utes worth of read­ing/​watch­ing/​listen­ing.

Get par­ti­ci­pants to sign-up affirmatively

Stu­dents are busy and will flake from op­tional events if they feel like it. For this rea­son, it’s good to get peo­ple to com­mit to show­ing up. Usu­ally, I first send out an RSVP form (us­ing google forms or similar) to de­ter­mine which 15-16 peo­ple get a spot for the din­ner. Then, I send an email to those peo­ple, ask­ing them to re­spond with an email con­firm­ing that they are able to come and that they will do the read­ing.

Beyond that, all you’ve got to do are the ba­sic lo­gis­tics:

  • Agree on a date that works for the pro­fes­sor. It can also be a good idea to fit it with the fel­low­ship cal­en­dar. (E.g. in Fall 2018 we had a con­ver­sa­tion with Shelly Ka­gan about ethics of fu­ture gen­er­a­tions dur­ing the same week that this was the topic for the fel­low­ship).

  • Book a din­ing room well in advance

  • Pick up the key (if ap­pli­ca­ble) for the din­ing room at the day of the dinner

So­cial events

One of the main pur­poses of the fel­low­ship is to in­te­grate new mem­bers into the core EA com­mu­nity at Yale. To this end, so­cial events are pretty es­sen­tial.

Ini­ti­a­tion. As part of the fel­low­ship, we had an “ini­ti­a­tion” event on the first week­end of the pro­gram. This was framed as be­ing “manda­tory” and was pretty well at­tended. The ini­ti­a­tion was a chill mixer tak­ing place in a dorm com­mon room. Bev­er­ages and snacks were served and we played a few games, in­clud­ing name games and ice-breaker games like “two truths and a lie.”

Other so­cial events. Through­out the semester, we in­vited the fel­lows to all YEA so­cial events but it was pretty rare that any fel­lows at­tended. In their eval­u­a­tion sur­vey re­sponses, fel­lows seemed to sug­gest that ac­tu­ally wanted to be more of a part of the com­mu­nity, but just didn’t have the time. There’s some room for im­prove­ment on this front!


As part of the fel­low­ship, we had an in­tro­duc­tory work­shop on im­por­tant EA con­cepts. The pur­pose was the in­tro­duce use­ful con­cepts and ter­minol­ogy to guide fu­ture dis­cus­sions. The work­shop was largely based on ver­sion made by Huw Thomas of Oxford EA, who de­serves much credit for his work.

The work­shop was fairly in­ter­ac­tive, lasted 3 hours, and took place dur­ing the first week of the fel­low­ship.

See Ap­pendix 10a for work­shop con­tent and Ap­pendix 10b for work­shop slides.

Run­ning the fel­low­ship

Once a co­hort of fel­lows has been se­lected, the work of ac­tu­ally run­ning the fel­low­ship be­gins.


You need a re­cur­ring book­ing. Prefer­ably some­where cen­tral and some­where that al­lows for a good con­ver­sa­tion held by 15-18 peo­ple. At Yale, The Say­brook-Bran­ford room and Bran­ford Men­dell Room are both ex­cel­lent op­tions. I pre­fer a room that is sep­a­rate from the din­ing hall (un­like e.g. the Silli­man din­ing an­nex), be­cause this choice dis­cour­ages get­ting up in the mid­dle of the con­ver­sa­tion to get more food, which is dis­rup­tive.

Send­ing weekly emails

Every week, I usu­ally send out two emails. See Ap­pendix #4 with all the emails sent out dur­ing the 2018 Fall Fel­low­ship. If the fel­low­ship din­ner is on Wed­nes­day, the first email should ideally be sent out dur­ing the week­end. The sec­ond email is a re­minder email sent on the day of the din­ner. An email usu­ally in­cludes:

In­tro­duc­tion to next week’s topic. How is it rele­vant to EA? What will we dis­cuss and why?

As­sign­ment. In­clude links to ev­ery as­sign­ment (dou­ble check that links work be­fore send­ing) and a note about whether the read­ing is op­tional or manda­tory.

Other in­for­ma­tion. Men­tion other group events of the week, sign-ups for spe­cial events, etc.

You should definitely use Mailchimp, as it makes for some nice-look­ing emails.

Run­ning dis­cus­sions

Make sure ev­ery­one has a chance to speak. Some peo­ple are more talka­tive; louder; bet­ter at “seiz­ing the micro­phone”; and quicker at for­mu­lat­ing their ideas than oth­ers. This means that you’ll some­times have dis­cus­sions where a few peo­ple are do­ing most of the talk­ing. This is bad for two rea­sons. One, it can be less ex­cit­ing and in­ter­ac­tive for the peo­ple who don’t talk. Two, you get an un­rep­re­sen­ta­tive sam­ple of peo­ple’s opinions and peo­ple might think that no one be­lieves in po­si­tion X, just be­cause no one hap­pens to be ar­gu­ing in fa­vor for it.

If this hap­pens, it might be use­ful to some­times say “How about some of the peo­ple who haven’t re­ally had the chance to talk yet—is there any­thing you’d like to add?” Another trick is to pay close at­ten­tion to peo­ple who aren’t say­ing much. Some­times, their face will re­veal that they ac­tu­ally have some­thing on their mind, in which case you can gen­tly say “Laura, you look like you’re [nod­ding/​dis­agree­ing/​what­ever]. What’s on your mind?” Some­times, they’ll chime in and some­times they’ll po­litely say no.

Table un­re­solve­able or tan­gen­tial dis­cus­sions.Every once in a while, the con­ver­sa­tion will move to­wards re­ally big pic­ture ques­tions like “cap­i­tal­ism ver­sus other eco­nomic sys­tems” or “rad­i­cal ver­sus mod­er­ate poli­ti­cal plat­forms.” When this hap­pens, I like to in­ter­ject and say some­thing like “This is an ex­tremely in­ter­est­ing and im­por­tant is­sue. How­ever, I think we should table it for now and stick to the topic at hand. The con­ver­sa­tion should definitely be con­tinued out­side the fel­low­ship, though.” The main benefits of this ap­proach are:

  • Keep­ing fo­cus on the topic of the night

  • Mak­ing progress on nar­rowly defined ques­tions in­stead of aim­lessly dis­cussing ex­tremely difficult ques­tions that could not pos­si­bly be re­solved in the span of one hour

  • Avoid­ing heated ide­olog­i­cal debates

Mak­ing sure ev­ery­one is on the same page. Every once in a while, some­one will re­fer to an ad­vanced con­cept, ei­ther from their course­work or from the read­ing, and go on to dis­cuss that. In some of those cases, it might be a good idea to in­ter­ject and say some­thing like “Would you mind defin­ing that term?” or “Can you re­mind us what her ar­gu­ment is?”, which al­lows other peo­ple to par­take even if they don’t know the term or con­cept.

Be re­spect­ful of peo­ple’s opinions and in­ter­ests.Many fel­lows will un­doubt­edly—as they should—care deeply about is­sues other than those com­monly con­sid­ered in EA. Of course, you want to pre­sent sus­tained ar­gu­ments for pri­or­ity EA-ar­eas. How­ever, be­ing dis­mis­sive to­wards other causes and or­ga­ni­za­tions will do you no benefit and can cause a great deal of harm. Not only will it make for a less en­joy­able ex­pe­rience for the fel­lows but you also risk that fel­lows be­come defen­sive and even less likely to con­sider EA ar­gu­ments. Even more trou­bling is the risk that you af­fect the rep­u­ta­tion of EA nega­tively by mak­ing the com­mu­nity and philos­o­phy ap­pear hos­tile to and in­com­pat­i­ble with a con­cern for other, im­por­tant is­sues.

Ed­u­cate, don’t ar­gue. Oc­ca­sion­ally, some­one will pre­sent a claim that is some­what at odds with main­stream think­ing within EA. Often, I be­lieve that the right course of ac­tion is to re­sist the temp­ta­tion to re­fute their ar­gu­ment right away. You’re not there to prove how much you know about EA or to con­vince ev­ery­one that EA is 100% right on 100% of is­sues (as it most likely isn’t). In­stead, you are there to offer in­sights and ar­gu­ments to peo­ple for their con­sid­er­a­tion. Some peo­ple won’t change their mind on some is­sues, and it can be a waste of ev­ery­one’s time to try and con­vince them that they’re wrong. One of the best pieces of ad­vice I ever got was to think of EA ad­vo­cacy as be­ing about “an offer of in­for­ma­tion, not about per­sua­sion.”

Ap­proaches to keep­ing at­ten­dance high

Ar­guably the most clear mea­sure of suc­cess for the 2018 Fall fel­low­ship was the very high at­ten­dance rate span­ning 12 weeks of weekly din­ner dis­cus­sions and work­shops. Most of the ab­sences were “ex­cused” in the sense that the fel­low had con­tacted me prior to the din­ner to let me know. So, what are the keys to en­sur­ing high at­ten­dance?

Re­cruit the right peo­ple. I think a lot of the things that can be done to achieve high at­ten­dance hap­pens be­fore the first meet­ing even takes place. At­ten­dance will start at a high baseline if you ac­cept fel­lows who seem com­mit­ted; who aren’t scared off by the ex­pec­ta­tion of spend­ing 3-5 hours on the fel­low­ship per week and who are gen­uinely in­ter­ested in EA ideas.

Make it clear that the ex­pec­ta­tions are high. Again, “you get what you tol­er­ate.” Re­sist the temp­ta­tion to say things like “The read­ings are op­tional” or “It’s to­tally okay if you ar­rive a bit late.” In­stead, em­pha­size that this is a fel­low­ship with high stan­dards.

We tell fel­lows that they must at­tend 80% of meet­ings in or­der to be “rec­og­nized as grad­u­at­ing fel­lows” by the end of the semester. (You don’t nec­es­sar­ily have to en­force this, but it’s a good idea to set the bar high.) At the be­gin­ning of each manda­tory event, take at­ten­dance, not­ing if peo­ple are “Pre­sent, Ex­cused ab­sence (if they’ve no­tified you in ad­vance), Ab­sent with­out ex­cuse, or Late”. I told them that I’d take at­ten­dance, and they must have seen that I pul­led out my com­puter to do it, but I didn’t in­form them that I also noted if they were late. This was mostly to keep track so that I could nudge some­one of they were con­sis­tently late.

Make the fel­low­ship feel like a big deal. Peo­ple are more likely to in­vest a lot of time and en­ergy if they feel like they’re part of some­thing se­ri­ous. For this rea­son, it seems to have a good effect to do things that in­crease the feel­ing of be­ing in a real pro­gram. Some things we did:

Web­page. Create a page on your stu­dent group web­site (e.g. the Yale EA site) where you post pho­tos and in­for­ma­tion about the fel­lows. In­stead of just hav­ing the fel­lows send in ran­dom pho­tos, spend fif­teen min­utes be­fore din­ner dis­cus­sion tak­ing pho­tos with a good qual­ity cam­era.

Hand­book. Print out a hand­book with a cal­en­dar, a state­ment of the fel­low­ship’s pur­pose, a list of norms and ex­pec­ta­tions for the fel­lows etc. See Ap­pendix #5 for the Fall 2018 hand­book.

Use­ful read­ings for dis­cus­sion moderators

I be­lieve any­one who mod­er­ates EA dis­cus­sions should read these guidelines on de­pict­ing poverty, this piece on em­pathic com­mu­ni­ca­tion of effec­tive al­tru­ism and this piece on why & how to make progress with di­ver­sity and in­clu­sion in EA.

Im­pact Challenge

As part of the fel­low­ship, we launched our first-ever “Im­pact Challenge” (IC). This was es­sen­tially a mas­sive Giv­ing Game, span­ning three weeks. Dur­ing the IC, which was gen­er­ously funded by the Cen­tre for Effec­tive Altru­ism, fel­lows were asked to choose one of four or­ga­ni­za­tions to re­ceive $3,000 in real cash. Ul­ti­mately, the fel­lows voted to donate the $3,000 to the Against Malaria Foun­da­tion.

This pro­ject was similar to the Oxford Pri­ori­ti­sa­tion Pro­ject, though there were sev­eral sig­nifi­cant differ­ences. (The OPP’s di­rec­tor, Tom Sit­tler, wrote a nice re­view similar to this one.)


The IC con­sisted of:

  • One din­ner meet­ing spent go­ing over use­ful con­cepts in health and development

  • One din­ner meet­ing spent dis­cussing ini­tial im­pres­sions of the can­di­date organizations

  • A 3-hour “Re­search-a-thon” spent do­ing re­search of the or­ga­ni­za­tions in small groups

  • An op­tional lunch with Mush­fiq Mo­barak, pro­fes­sor in de­vel­op­ment economics

  • One din­ner meet­ing spent dis­cussing and vot­ing on the organizations


Be­fore the semester, we for­mu­lated the fol­low­ing goals for the Im­pact Challenge:

  • Fel­lows learn the core con­cepts of Effec­tive Altru­ism and char­ity pri­ori­ti­sa­tion

  • Fel­lows be­come more ded­i­cated to mak­ing Effec­tive Altru­ism a big part of their lives, by tran­si­tion­ing our dis­cus­sions from the­ory to prac­tice

  • Fel­lows feel en­gaged and we see a high re­ten­tion rate

  • The qual­ity of the pool of ap­pli­cants in­creases, as the fel­low­ship ap­pears even more se­ri­ous and professional

Can­di­date or­ga­ni­za­tions

Prior to be­gin­ning the IC, I se­lected five or­ga­ni­za­tions from which the fel­lows had to choose:

  • Against Malaria Foundation

  • GiveDirectly

  • Ev­i­dence Action

  • De­worm the World Initiative

  • StrongMinds

I did this for a few rea­sons:

Keep­ing the fo­cus on one cause area. On the face of it, one might think that fel­lows should com­pare or­ga­ni­za­tions work­ing in differ­ent cause ar­eas, such as global health or AI safety. After all, one of the most im­por­tant parts of effec­tive al­tru­ism is to com­pare differ­ent such ar­eas. How­ever, com­par­ing or­ga­ni­za­tions in differ­ent ar­eas poses sev­eral challenges.

For one, it is very difficult to make ap­ples-to-ap­ples com­par­i­sons of the pros and cons of or­ga­ni­za­tions in differ­ent fields; es­pe­cially con­sid­er­ing the lack of quan­ti­ta­tive meth­ods in our IC.

Se­cond, the choice be­tween cause ar­eas does of­ten not come down to anal­y­sis but rather to moral and philo­soph­i­cal judg­ments. Th­ese judg­ments are ex­tremely im­por­tant and con­sid­er­ing them is a cru­cial part of be­ing an as­piring effec­tive al­tru­ist. How­ever, I do not be­lieve that the IC is the ideal fo­rum for dis­cussing these philo­soph­i­cal is­sues. I sus­pect that peo­ple are much less likely to up­date their moral in­tu­itions through a few hours dis­cus­sion, than they are to up­date their em­piri­cal be­liefs about what’s effec­tive. For this rea­son, I think that within-area com­par­i­son would al­low fel­lows to gen­uinely learn a lot through open-minded re­search, whereas I fear that be­tween-area com­par­i­son might risk lead­ing the fel­lows to ar­gue for whichever cause area they felt most in­tu­itively com­pel­led by from the be­gin­ning.

Fi­nally, I have an idea that a lot of peo­ple in the EA com­mu­nity be­gan with an in­ter­est in global health and de­vel­op­ment and only later gained an in­ter­est for e.g. ex­is­ten­tial risks (as I dis­cuss el­se­where.) On this view, it makes sense to avoid pit­ting cause ar­eas against each other so early in peo­ple’s EA tra­jec­tory. We gave this a lot of thought prior to launch­ing the IC and it also seemed to have been the best choice af­ter the fact. If any­one wishes to dis­cuss this topic fur­ther, feel free to reach out.

En­sur­ing a min­i­mum im­pact. Since a con­sid­er­able amount of money will be donated, I reckon it is de­sir­able to elimi­nate the risk that the funds are donated to a poor choice. This also helps with ap­ply­ing for funds: you can guaran­tee the donor that the im­pact of their de­ci­sion will be equal to the value of donat­ing to a highly effec­tive or­ga­ni­za­tion plus what­ever the value of the IC is.

Fo­cus­ing the im­pact challenge. It is tempt­ing to think that the fel­lows should gen­er­ate the list of can­di­dates from scratch. How­ever, this is an ex­tremely difficult task and would likely be too much of a mouth­ful, un­less the IC had been scaled up to span sev­eral weeks and to re­quire a whole other level of com­mit­ment. Limit­ing the op­tions al­lows for more in-depth re­search of each can­di­date and for higher qual­ity dis­cus­sions.


I sort of im­pro­vised a vot­ing sys­tem, where each fel­low would rank the or­ga­ni­za­tions from 1 through 4. If an or­ga­ni­za­tion got ranked first, it got 1 point, if it was ranked sec­ond, it got 2 points, etc. In the end, the or­ga­ni­za­tion with the fewest points won. This sys­tem seemed to get the job done alright, but there’s likely a bet­ter solu­tion. I en­courage or­ga­niz­ers to listen to the 80,000 Hours pod­cast in­ter­view with Aaron Ham­lin on vot­ing sys­tems.

Some fel­lows ex­pressed in­ter­est in be­ing able to spread the dona­tion across all four or­ga­ni­za­tions or to at least have the al­lo­ca­tion be pro­por­tional to the num­ber of votes. While this is also an op­tion, I felt that the win­ner-takes-all model was bet­ter at cre­at­ing in­cen­tives for rigor­ous re­search and se­ri­ous de­liber­a­tion.


After the first din­ner, the fel­lows met on a Satur­day for a ses­sion of col­lec­tive re­search. Due to a few peo­ple not be­ing able to make it to the re­search-a-thon, I de­cided to drop one or­ga­ni­za­tion, so that the num­ber of groups would fit bet­ter with the num­ber of or­ga­ni­za­tions. I chose to drop StrongMinds, since it was the one that I had the low­est prior con­fi­dence in and also was the one that had met the most skep­ti­cism in our ini­tial dis­cus­sion.

The at­tend­ing fel­lows were split into groups of three and as­signed two out of the four can­di­date or­ga­ni­za­tions. (If some­one fails to at­tend, you’ll have to im­pro­vise a bit with the group ar­range­ment.) The or­ga­ni­za­tions were as­signed such that each was cov­ered by two groups. Fur­ther­more, each group was as­signed one or­ga­ni­za­tion for which they had the re­spon­si­bil­ity of writ­ing a re­search sum­mary.

Each group spent the time as fol­lows:

  • 5 min­utes spent prepar­ing to re­search their first or­ga­ni­za­tion and as­sign­ing re­search ques­tions among each other;

  • 45 min­utes spent re­search­ing their first organization

  • 10 min­utes sum­ma­riz­ing, dis­cussing, and writ­ing down their ini­tial findings

  • 60 min­utes rins­ing and re­peat­ing for their sec­ond organization

  • 20 min­utes com­par­ing and dis­cussing the two organizations

  • 40 min­utes writ­ing up a sum­mary of find­ings for their sec­ond or­ga­ni­za­tion. See Ap­pendix 6 for ex­am­ple of re­search sum­maries.

Although this struc­ture worked fairly well, it was pretty much im­pro­vised on the day of the ses­sion, and it is pos­si­ble that an­other struc­ture could work just as well or bet­ter.

Eval­u­at­ing the or­ga­ni­za­tions

You could imag­ine an Im­pact Challenge where the goal was to have ex­tremely high stan­dards for rigor­ous de­ci­sion-mak­ing, in ac­cor­dance with EA meth­ods and val­ues. Such a pro­ject would likely em­ploy quan­ti­ta­tive mod­els, such as the ones used by GiveWell or in the Oxford Pri­ori­ti­sa­tion Pro­ject. How­ever, this would re­quire con­sid­er­ably more time and it would re­quire meth­ods that are un­fa­mil­iar to many fel­lows. One of Tom Sit­tler’s key les­sons from the Oxford Pro­ject seems to be that too high am­bi­tions for the eval­u­a­tion pro­cess can get in the way of a suc­cess­ful pro­ject and I agree with this as­sess­ment.

More­over, it could be ar­gued that the benefit of hav­ing such rigor­ous meth­ods might ac­tu­ally not be that great. To be sure, we want fel­lows to rec­og­nize that char­ity eval­u­a­tions re­quire rigor­ous meth­ods. How­ever, we are not so much aiming to train peo­ple in eval­u­a­tion as we are hop­ing to teach the im­por­tance of do­ing such eval­u­a­tion in a good way. To this end, ex­pos­ing fel­lows to the rigor­ous liter­a­ture of GiveWell seems perfectly ad­e­quate.

In­stead of quan­ti­ta­tive meth­ods, we re­lied on much less rigor­ous, more qual­i­ta­tive as­sess­ments based on liter­a­ture re­view and dis­cus­sion. This ap­proach seemed jus­tified in part by the fact that we were choos­ing from GiveWell-recom­mended or­ga­ni­za­tions, which gave us con­fi­dence that each al­ter­na­tive was largely in the same bal­l­park in terms of im­pact.

We tried to in­clude a some­what quan­ti­ta­tive (though non-rigor­ous) com­po­nent by ex­plic­itly ask­ing “What does this or­ga­ni­za­tion achieve with an ad­di­tional $3,000?” but the dis­cus­sion was mostly con­cen­trated on more gen­eral pros and cons of each or­ga­ni­za­tion.

Eval­u­a­tion of Fellowship

Be­fore and af­ter the fel­low­ship, we had fel­lows com­plete sur­veys. This had the dual aim of mea­sur­ing our im­pact and learn­ing ways to im­prove the fel­low­ship.

See Ap­pendix 7a-7c for sur­vey de­sign. See Ap­pendix 8 for sur­vey re­sults. (This ap­pendix is fairly com­pre­hen­sive and I en­courage read­ers to take a look.)

Over­all, based on my own sub­jec­tive ex­pe­rience and a qual­i­ta­tive as­sess­ment of the sur­vey re­sults, I’d say that the fel­low­ship was a suc­cess; that it was an en­joy­able ex­pe­rience for ev­ery­one in­volved; and that it was net-pos­i­tive in terms im­pact—even if we con­sider the op­por­tu­nity cost of the 100+ hours I spent or­ga­niz­ing and run­ning it.

In terms of the goals set out ear­lier in this doc­u­ment:

Bring­ing mem­bers into our fun­nel. This goal was not achieved as well as I had liked. As I write this, it’s a bit too early to tell whether alumni from the fel­low­ship will be in­volved with our group go­ing for­ward. My guess is that at least 30% of fel­lows will go on to at­tend our events and that at least 1-2 of them will join our board. Still, in­ter­est in join­ing the board hasn’t been ap­peared to be quite as strong as I had hoped, so there is room for im­prove­ment.

Direct im­pact of af­fect­ing fel­lows’ views. This goal was achieved. Judg­ing by the sur­vey re­sults as well as the gen­eral change in tone and view­points through­out the term, I be­lieve this goal was achieved. We didn’t have ap­pear to af­fect many rad­i­cal changes of per­sonal philos­o­phy or ca­reer plans—though there were at least a cou­ple of peo­ple who be­gan re­con­sid­er­ing their paths as a re­sults of the fel­low­ship—but I would also think that this is gen­er­ally difficult to achieve in just one semester.

Strength­en­ing YEA as a promi­nent group on cam­pus. This goal was achieved. The fel­low­ship con­tinues to be our flag­ship ac­tivity on cam­pus. It helps us reach more peo­ple and con­tributes to the feel­ing among board mem­bers that we are do­ing sig­nifi­cant work with clear out­comes.

Beyond the fel­low­ship

Fel­lows who have par­ti­ci­pated to a satis­fac­tory de­gree (in the hand­book, this was 80% at­ten­dance, though in prac­tice we tol­er­ated slightly lower rates than that) were given a printed cer­tifi­cate. See Ap­pendix 9 for ex­am­ple cer­tifi­cate (de­signed by Frankie An­der­sen-Wood). Fel­lows were also strongly en­couraged to join our group’s board and to stay in­volved with our ac­tivi­ties on cam­pus.


Th­ese ap­pen­dices are not in­tended to be read from A-Z. Rather, they might be use­ful to an­swer par­tic­u­lar ques­tions like “What read­ings should I as­sign this week?” or “What are some talk­ing points I can draw on in case tonight’s dis­cus­sion comes to a halt?”

Ap­pendix 1: Poster for re­cruit­ment (On re­quest, we can also send you the canva tem­plate to edit)

Ap­pendix 2: Read­ing list

Ap­pendix 3: Dis­cus­sion talk­ing points

Ap­pendix 4: Emails sent to fel­lows

Ap­pendix 5: Hand­book

Ap­pendix 6: Re­search sum­maries ex­am­ples

Ap­pendix 7a: Pre-fel­low­ship sur­vey de­sign

Ap­pendix 7b: Post-fel­low­ship sur­vey de­sign

Ap­pendix 7c: Notes on sur­vey de­sign

Ap­pendix 8: Sur­vey re­sults

Ap­pendix 9: Ex­am­ple cer­tifi­cate (On re­quest, we can also send you the canva tem­plate to edit)

Ap­pendix 10a: In­tro work­shop con­tent

Ap­pendix 10b: In­tro work­shop slides

Ap­pendix 11: Fel­low­ship Application