This is an update from the Centre for Effective Altruism on our work in the fourth quarter of 2020. Most of this information was not included in our annual review, which we began to draft in October.
Our mission is to build a community of students and professionals acting on EA principles, by nurturing high-quality discussion spaces.
In Q4, we began to focus on achieving our annual goals.
Going into the quarter, we had a much clearer scope and goals than before. We began to focus more on those goals, and test out new programs that could further them, like online fellowships and student-focused events. We also reviewed our progress in 2020, made plans for 2021, and began a major fundraising campaign.
Huw Thomas joined the groups team. Now that CEA’s scope and goals are clear, I think that we’re in a position to make several excellent hires. This will be a major focus for next year.
Joan (our managing director) took over our groups and recruitment work, and Katie (formerly head of groups) became the interim head of community health (while we search for a permanent manager). This will mean each goal has a single owner, and will also allow Joan and Katie to play to their strengths.
Progress towards our goals
Goal 1: We aim to help onboard new students and young professionals, particularly from highly-ranked universities and underrepresented groups. We plan to do this by following a cohort of students through the academic year, and helping them to engage with quality content, make connections, and take action.
We’ve begun to search for examples of EAs who became highly engaged through our work in 2020 to pressure test our ability to measure this target. So far, we’ve found 2 highly-engaged EAs from the Student Summit, 3 from CBGs, and 6 from introductory fellowships.
One example is Neel who has just completed an AI Alignment visiting fellowship at FHI, and is about to start a research engineering internship at DeepMind. He graduated top of the year in Maths at Cambridge and is an IMO Gold Medallist. He has previously interned at Jane Street and is considering earning to give. In his response to the CBG group member survey in 2019, he estimated attending 20+ of the Cambridge group’s events in the last 12 months. He also mentioned the following:
“[As a result of] the EA Cambridge retreat last year […] and contacts I made, I ended up interacting a lot with the EA London community (met >20 new people and made >5 good new friends), went to a CFAR workshop, EAG, got 80K coaching and went to an AIRCS workshop. I think that being involved in helping to run/organise EA Cambridge events was pretty helpful for me to become more agenty and to feel more emotionally invested in things (eg, doing outreach/giving careers advice helped me clarify my own thoughts). As a result of this much stronger exposure to the EA Community, ideas like changing my career plan and working on things like AI Safety feel much more plausible and OK on an emotional level, and more likely to happen.”
Goal 2: We aim to increase the total value that highly-engaged EAs get from EA discussion spaces by 30%, by producing more high-quality content (which will encourage continued learning) and helping them make more connections.
We are still developing a dashboard to track the number of engagement hours and new connections.
Forum and event engagement time continued to grow this quarter (after a strong Q3).
Goal 3: We aim to mitigate key risks to EA’s and CEA’s success.
The Groups team advised group organizers on handling conflicts within the group, handling tense discussions, etc.
The Community Health team advised several other EA-align organizations about best practices in HR and diversity, equity, and inclusion.
We worked with Rethink Priorities to plan a survey measuring US students’ inclination/understanding of EA.
Other program updates
We continue to pilot introductory and advanced fellowships in groups (especially at highly-ranked universities), and we’re planning to promote fellowships more broadly online.
We had ~100-150 attendees at three groups training sessions, and >75 calls with organizers.
We made 5 Community Building Grants, including a new one at Harvard.
We ran two major online events and began to plan a more goal-focused program of events for next year.
80,000 Hours hosted a popular careers open thread on the Forum.
We began to import Pablo Stafforini’s effective altruism wiki (currently in development) onto the EA Forum.
We help local group organizers by advising them, providing resources they can use, and creating online spaces where they can share resources and support each other.
Progress this quarter
Huw Thomas will be joining CEA in a Groups Associate role.
Huw will develop resources for EA university groups, and aid them with strategy and planning.
Before joining CEA, Huw studied Maths and Philosophy at the University of Oxford, and spent two years funded by a CEA community building grant to work on various projects, including leading the EA Oxford student group and launching the Student Career Mentoring Program.
Our contractors ran 2 strategy workshops for a total of 8 university group leaders from Stanford, Harvard, and Yale. (The average likelihood of recommending the event was 8.8/10.)
Joan hosted a call for university groups with Community Building Grants in December to align on plans for 2021.
300+ students (the majority from highly-ranked universities) participated in an introductory fellowship in Q4. Fellowships are 6-8 week courses for learning more about topics related to effective altruism and discussing them in a weekly facilitated group.
We reviewed and drafted learning objectives for the In Depth Fellowship. This led to some major changes in our approach; we plan to launch an updated version in January. We surveyed students who had taken the introductory fellowship, and 130 of them expressed interest in the In Depth Fellowship. So did 425 people who took the EA Survey, and 68 Student Summit attendees. (Note: these may double- or triple-count some people).
Community Building Grants (CBGs) allow group organizers to professionally engage in local community building activities.
We made the following grants:
EA Brown − 1 FTE, 6 months
EA Harvard − 0.5 FTE, 5 months
EA Czechia − 2 FTE, 12 months
EA Stanford − 1.4 FTE, 12 months
EA Geneva − 2 FTE, 12 months
With our help, EA Norway hosted a training call for CBG recipients on user-centered design in December.
Broad resources & support
We drafted a website for making group-organizer-led online fellowships available to the public.
We’re working with Emma Abele (EA Brown) to pilot this in Q1.
We’re planning to offer introductory and advanced fellowships, and may expand offerings in the future based on data we’ve gathered about what students are looking for.
We hope that this will make high-quality groups-style content accessible to anyone who wants to join.
Trainings at the Student Summit:
~125 group organizers attended a workshop from Joan on goals for uni groups
~100 organizers attended a session on how to teach group members about EA principles
~150 organizers attended a workshop on using career resources in groups
These resources have now been published on the EA Hub to help other organizers.
We have also updated our Giving Season activity resources in collaboration with Giving What We Can.
We have passed information on 600+ attendees from the Student Summit to group organizers, so that the organizers can follow up with anyone they’re not already in regular contact with.
The Groups Survey has been completed (327 people participated).
The groups list is up to date online and in our CRM.
$5,060 in basic group funding has been disbursed.
In Q4, the Groups team had 75 calls with group organizers, plus several more casual, unlogged calls.
We now have information about areas of demand for further engagement from the Student Summit, the Intro Fellowship post survey, and the EA Survey. We’ve shared this information with CBG recipients. Sample below:
Huw will add a lot of capacity and relevant experience to the team.
We’re excited about the pilot online courses platform: we think that local group organizers could refer their members to this while still holding lower-effort events. This would allow them to provide quality content without spending much time: “not enough time” is the most common bottleneck for organizers (according to the 2020 EA Groups Leaders survey).
For the first time, we have data on what people interested in EA want from groups, which we can share with group leaders.
Mentorship for group organizers: we think that group organizers might benefit more from calls with people who can give them direction and feedback on their plans for the group, rather than calls with core EAs who can advise them on EA ideas and career plans (but might have fewer ideas about group plans). This is partly because we’ve found that mentees often aren’t sure what kind of support they want.
We had an objective to align on a set of 2021 goals and metrics for the CBG programme based on CEA’s objectives; this was deprioritized due to the management change between Joan and Katie.
The Community Health team aims to reduce key risks to the EA community’s future. Their work includes fostering a healthy culture, improving diversity, mitigating harm done by risky actors, reducing the harm of negative PR, and identifying risks to early field-building.
This quarter, the team handled:
7 inquiries or cases around media stories about EA
6 concerns around interpersonal problems such as sexual harassment
4 cases where we advised on situations in early field building (geographies or fields where EA is just getting established)
24 other situations where we advised groups, organizations, and individuals on situations like conflicts between staff or group members, online conflicts, best practices in HR, personal or mental health problems, and improving diversity, equity, and inclusion within organizations.
Progress this quarter
We have worked with the Operations team to set up systems to vet and monitor donations.
We have worked with Rethink Priorities to set up a repeatable survey of college students to monitor EA awareness and inclination in the US.
Diversity, equity, and inclusion
We read the available chapters of Will MacAskill’s in-progress book on long termism, and provided support to Will’s team on finding other reviewers who could advise on diversity, equity, and inclusion in the book.
We followed up individually with 13 promising Black / Latino leads from the Student Summit.
Culture and epistemics
We provided input on the epistemics section of the In Depth Fellowship
Advised and interviewed group organizers on handling group discussions about the relationship between EA and social justice; we are currently preparing a resource we hope will be helpful for them.
We had some internal confusion between teams about the process of reaching out to Black and Latino students after the Student Summit; for a time, we weren’t sure whether to prioritize having conversations ourselves, or finding a person who would be well-suited to having many such conversations in the future.
Events enable attendees to make new connections, learn about core concepts, share and discuss new research, and coordinate on projects.
This report covers:
EA Student Summit
Early results from this event were covered in our annual report.
Some additional updates:
There were 1.07 meetings per student through our event app (848 meetings in total), but people who filled out our survey reported 4.3 new connections (average). We think this is mostly because people who filled out the survey engaged more with the event, although some people were also booking meetings outside of the event app.
94% of participants in our mentorship program that responded to our survey intend to take at least one significant action (e.g. Giving What We Can pledge, career change) in the next 6 months.
89% of highly-engaged EAs (who attended the event mostly to advise students) report that their expectations for the event were met or exceeded.
EAGxAsia-Pacific, held virtually in November, evolved out of in-person conferences planned by organizers in Singapore and Australia. We also aimed to promote connections between community members in the Asia-Pacific region and community members in other places. Unlike our previous virtual events, the event was scheduled for the convenience of people in Asia-Pacific time zones. 526 people attended the event. As the event was held recently, we do not yet have an in-depth analysis of the outcomes.
Attendees reported 4.4 new connections in the post-event survey (on average). We succeeded in attracting >200 attendees from outside the Asia-Pacific region, which we hope will build international connections.
Attendees seem to be increasingly more demographically diverse than EA Survey respondents; we’ve seen diversity increase as events have gone virtual and become more accessible to people in a variety of countries.
Attendees’ likelihood to recommend Student Summit to other people was lower than the average for in-person events, while EAGxAsia-Pacific scores were more in line with in-person events:
Total hours of engagement across events for 2020:
Content engagement seems to drop quite a bit over the course of virtual events — it seems harder to maintain viewership, compared to in-person events. We will consider how to remedy this for future virtual events (for example, with shorter events, fewer talks, or more interactive sessions).
The Events Team is increasingly focused on designing events around specific types of users, in support of CEA’s org-wide goals. We hope that this exploration will help us to find more impactful event formats. The team has spent the last few weeks designing a portfolio of events for 2021. We also continue to receive advice from experts on which in-person events (if any) will be sensible to hold in 2021 given COVID-19.
The Effective Altruism Forum (EA Forum) aims to be the central place for collaborative discussion about how to do the most good.
Progress this quarter
We continued our experiment with open threads by adding monthly progress threads in addition to welcome threads. These have been moderately successful (usually getting a dozen or so comments). Open threads can be found here.
80,000 Hours posted an open careers thread which drew dozens of participants, many of whom had never posted on the Forum before.
Pablo Stafforini received a grant from the Long Term Future Fund to create a Wikipedia-like collection of content focused on EA. We’ve been working with Pablo to host his work on the Forum’s tagging/wiki system. We began uploading his work this quarter. We hope that the combination of Pablo’s full-time work and the Forum’s network effects will allow this wiki to be maintained and used much more consistently than past efforts of this kind.
We imported the following features from LessWrong:
You can comment on wiki/tag pages.
The new text editor is now the default option.
You can embed content from Elicit, YouTube, and Metaculus.
EA Forum-specific features:
Posts now appear on the front pageby default, rather than having to be moved there by moderators. We expect this to create a more pleasant experience for both authors and moderators.
A post’s social media preview image is now the first image in the post rather than the Forum logo (we’ve passed this feature along to LessWrong). We hope that this will drive more engagement from social media.
We link to the EA Hub’s groups page in the sidebar (thus hopefully increasing referrals to local groups).
Our primary metric is the number of hours engaging with content on the Forum. This metric is up 13% (30-day trailing average view hours is 3452 on December 14 versus 3066 on October 1). Most of that growth comes from a couple of unusually successful posts in early December. From the beginning of October to the end of November, views actually dropped slightly.
You can see other metrics here. We ended the quarter up on most other metrics, particularly the number of comments, notably due to the 80,000 Hours career advice open thread (which has 247 comments as of this writing). 80,000 Hours mentioned they might do these open threads quarterly if there was interest.
We had planned to set up OAuth in Q4 (to allow a single sign-in for the Forum and EA Funds/Giving What We Can, and maybe other things in the future). We decided not to do this because the LessWrong team was already working on it.
As with most content platforms, our success is driven by the popularity of our users’ posts, which have a relatively fat-tailed distribution. Regular open threads seem like a promising approach (particularly if we can sustain the type of engagement we got in the career advice thread).
So far, tracking view hours seems like a good metric; it seems consistent with other metrics and our subjective sense of how the Forum is doing
Notably, page views reported by Google Analytics have increased 74% this quarter, so we are no longer in the situation where our internal metric is growing a lot but page views are flat.
Effective Altruism Funds (EA Funds) aims to be an easy and effective way for EA-aligned donors to donate.
EA Funds is now operating independently from CEA, more information here.
The Operations team provides the finance, legal, HR, grantmaking, office management, and fundraising support that enables CEA, 80,000 Hours, Forethought Foundation, EA Funds, and Giving What We Can to run efficiently.
Progress this quarter
We managed CEA’s annual review and fundraising process, publishing our annual review much earlier than normal and reaching out to many potential donors. We also ran financial analysis for all five organizations within the CEA legal entity (80,000 Hours, Forethought Foundation, etc). This process was much easier than last year, because our financial systems are now much more efficient.
We moved out of our former Oxford office, and pushed forward on the preparation of the new office. The project has been delayed by 6 weeks due to some complications with internal electrics, and we expect to take possession of the building on the 14th of January.
We improved the user interface of our grant management system. Our grants contractor is now the only person managing that system, which has freed up capacity for the rest of the team.
We began to scope out a project to develop a CRM for CEA’s projects.
Average score from key internal users: 8.7/10 (measured annually in November)
Grantee satisfaction with grant experience: 4.6/5
Grantee satisfaction with time taken to process grant: 4.7/5
Projected values combining US and UK (USD), as of December 14th.
Metrics and CRM
Development of our CRM has slowed this quarter, as we focused on the annual review and following up with leads from the Student Summit. Josh is now leading on developing the CRM, and is exploring next steps.
Joan began to manage all groups work, while Katie switched to a new interim role as head of community health. The main motivation for this was to make sure that each of our organization-wide goals has a clear owner. I also think that this will make good use of their skills: it will free Joan up to pursue her vision for recruitment and university groups, and I think that Katie will be strong at supporting and project-managing community health work.
The following people are now reporting to me and leading on well-specified org-wide goals:
Joan: recruitment goal
Ben: retention goal
Katie: risk reduction goal
Caitlin: people operations
I feel like we now have more clarity about our managers’ goals and responsibilities than we’ve ever had before, and I’m hopeful that this will allow CEA to make faster progress.
We’ve launched a new professional development strategy which is centered around a set of core competencies.
We’ve held biweekly all-staff discussion on recent publications, podcasts, and talks. We hosted two social events near the end of the quarter.
We ran an all-staff training session on survey design in collaboration with Faunalytics, and may do more expert-led training for staff.
Huw Thomas will join CEA as Groups Associate.
Morale has averaged 6.61/10 over the past 3 months, which is slightly lower than our annual average of 6.66.
We released our annual review and plans on the EA Forum. In December, we published an update about the Funds/Giving What We Can spinoff.
In 2021, we aim to help onboard 125 students and young professionals, 30 of whom are from 15 particular highly-ranked universities and 10 of whom are Black/Latino. Based on 2019 EA survey data, we currently estimate that there were around 100 new highly-engaged EAs in 2020, 18 from highly-ranked universities, and 6 who are Black/Latino. We are currently gathering additional data on this from the 2020 EA Survey, and we may adjust our estimate as we go. Although we’re uncertain about this target, we think that concrete and specific goals are clearer and more motivating. ↩︎
All students who attended an introductory fellowship, our introductory events, or the EA Student Summit in 2020. ↩︎
Weighted towards the most engaged and underrepresented groups. ↩︎
The event was specifically marketed toward group organizers, though we didn’t verify participation. ↩︎
The Forum’s front page is quite inclusive; any post with a clear connection to effective altruism will appear there (though most posts fall off after a few days, unless they get a lot of upvotes). The prior default option, “Personal Blog”, is used only for posts without a clear connection to EA, and was rarely the proper designation for posts. ↩︎
The expansion budget is the budget we aim to use, which allows us to make additional hires. ↩︎