EA Fellowship
1,000+ people have completed it
- 8.7 out of 10 on how likely participants are to recommend it
- About 3 new connections per participant
- 77% feel more confident they can make a positive impact
Participant exit surveys, 2023 to 2026
Our impact
What our work adds up to: the numbers from our programmes, an independent evaluation of what changes for people, and what the wider effective altruism community has achieved.
Why we measure
Why we evaluate our impact
We ask people to look at the evidence before they give or choose a career. We should hold ourselves to the same standard.
To shape our strategy
Knowing what works tells us where our time and money do the most good, and shows us what to fix or stop.
To stay accountable
The people who fund us should be able to see whether what we achieve is worth what they give.
To help donors decide
If you're thinking of supporting us, you deserve clear, honest information about what your money does.
To set a norm
We'd like measuring impact to be normal for organisations that build communities, not the exception.
In numbers
What our programmes add up to
EAGx Amsterdam 2025
517 attendees from across Europe, over half at their first EAGx
- 10 new connections per attendee, 4 of them impactful
- 8.9 out of 10 on how likely attendees are to recommend it
- 57 Giving What We Can pledges, worth an expected $2.9 million in donations, or $436,000 after adjustments
Giving What We Can's adjustments count donations made further in the future for less, count only money that goes to highly effective charities, and take out what people would have given anyway.
Post-event survey, 159 responses
Co-working space
1,580 visits from October 2025 to September 2026
- 8.9 out of 10 on how likely users are to recommend it
- About 2 extra productive hours for each day in the office
- About 8 new connections per user in the past year, 3 of them meaningful
- In the past year, 6 users got a job, internship or grant at an effective altruism organisation, and 6 started a new project or organisation
User survey, October 2026, 28 responses
Does it work?
What changes for people who get involved
But numbers alone aren't enough, so we've also commissioned qualitative research into our work. Over the winter of 2025–26, independent researchers interviewed 24 people who'd been involved with us. They never asked about us, only about what had changed in people's lives and why. Every mention of us came from the people themselves.
- 20 of 24
- said effective altruism helped them see more clearly how to have impact
- 18 of 24
- linked us to a new job, career plan or study plan
- Over half
- had changed how they donate
What helps people have impact
Community is the multiplier
The change people mentioned most was finding like-minded people. One step led to the next: a meetup, then the Fellowship, then a conference, then volunteering or a new job.
The Fellowship is the way in
The EA Fellowship and student groups were people's main ways in. The Fellowship was also the main reason people changed how they give: more, and to charities with strong evidence behind them.
Meeting people opens doors
Several people got their first job in AI safety through someone they met at our conference or in our co-working space.
Running a group builds people
Organisers gained confidence and skills in running projects and leading discussions, and made the connections that turned into concrete career plans.
You meet […] other people that are working or thinking about the same problems as you, and it can really help you navigate not just your career but also how you think about making an impact.
How the research worked
Three groups, eight people each
Fellowship participants from 2021 to 2024, student group organisers and users of our co-working space each had an in-depth video interview between December 2025 and January 2026.
Interviewers who didn't lead
The interviewers knew as little as possible about our work and never mentioned effective altruism or EAN. They asked about changes in careers, volunteering and giving, and let people name the causes themselves.
Anonymous to us
People signed up through a form only the researchers could see, so we don't know who took part.
Every claim mapped
The researchers used the Qualitative Impact Protocol (QuIP): they coded every cause and effect people described and counted how many made each link. Eleven more AI-assisted interviews told the same stories.
Qualitative impact evaluation by Bath Social & Development Research, February 2026
Public opinion
What the Dutch public thinks
Before people can get involved, they need to know effective altruism exists. The movement is only about 15 years old, and we've only had full-time staff since 2021. Yet roughly 100,000 Dutch adults can already explain what it is, and most people who hear about it like it.
- 0.7%
- of Dutch adults, roughly 100,000 people, can correctly explain what effective altruism isUS: under 1%
- 56%
- react positively after reading a short description, and 8% negativelyUS: 63% and 8%
- 69%
- of Dutch people who know it well are positive about itUS: 51%
How people feel about effective altruism
- Positive
- Neutral
- Negative
Netherlands, after reading a description
56% positive · 37% neutral · 8% negative
United States, after reading a description
63% positive · 29% neutral · 8% negative
Netherlands, people who know it well (25 people)
69% positive · 14% neutral · 18% negative
United States, people who know it well (39 people)
51% positive · 10% neutral · 39% negative
Rethink Priorities: survey of 3,331 Dutch adults, February 2026, weighted to match the population, and Pulse survey of 4,890 US adults, 2024. “Can correctly explain” counts people whose own description of the term was clearly right.
Cost-effectiveness
Towards an impact multiplier
We're painfully aware that every euro spent on our work could have gone to saving lives instead. GiveWell estimates that its top charities save a life for roughly $3,000 to $5,500. So the money spent on us has to do even more good here.
Giving What We Can estimates that every dollar spent on it leads to about seven dollars for highly effective charities. We'd like to work towards a similar estimate for ourselves. This is one approach we're exploring:
What the Dutch community achieves
People working in high-impact jobs, and money given to highly effective charities, valued in euros.
Our share
The part that wouldn't have happened without EA Netherlands.
What it costs
Everything our funders pay for, including the conference and the office.
Our multiplier
Euros of impact for every euro spent on us.
If we get there, we'd follow Giving What We Can in giving a best guess and a cautious lower bound.
Evidence that could help
The EA Survey
Rethink Priorities' survey of thousands of people in effective altruism shows what helped them get involved and have an impact.
Independent interviews
Bath Social & Development Research traced each change in people's careers and giving to everything that caused it, including things that had nothing to do with us.
Our own surveys
Office users and conference attendees tell us how much each outcome depended on us.
Comparing countries
Denmark and Belgium have national organisations that have had little funding. Comparing their communities with ours could show what a funded organisation adds.
A survey of the Dutch community
We're exploring a survey of the whole Dutch community with Rethink Priorities.
What's hard to put a number on
Some of the community's biggest wins are changes to policy and institutions. They're hard to value fairly, so they may be better described next to any multiplier than counted inside it.
The wider movement
Part of a global community
Effective altruism has been around for about 15 years. This is some of what people and organisations in it have achieved, across the problems it works on.
Global health
- Against Malaria FoundationProtected hundreds of millions of people against malaria with insecticide-treated nets
- GiveDirectlyDelivered over $800 million in cash transfers to over 1.6 million people in 12 countries
- Lead Exposure Elimination ProjectHelped bring in regulations projected to spare tens of millions of children lead exposure by 2100
Animal welfare
- Open Wing AllianceFreed hundreds of millions of hens from battery cages
- Good Food InstituteHelped secure approval for cultivated meat in the United States
- Legal Impact for ChickensOpened a new legal route for animal advocates to challenge cruelty in court
AI and other global risks
- Neel Nanda, Google DeepMindPioneered mechanistic interpretability: working out what happens inside AI models
- Nucleic Acid ObservatoryBuilt monitoring that spots dangerous biological sequences before they spread
- ForethoughtResearched what superintelligent AI could mean for humanity's future
Research and new charities
- GiveWellDirected more than $2.4 billion to evidence-backed charities
- Ambitious ImpactStarted 40+ charities serving 35 million people and a billion animals. Three people from our co-working space found it through us
- Animal Charity EvaluatorsInfluenced more than $68 million towards highly effective animal charities
Effective giving
- Giving What We CanBuilt a network of thousands of pledgers who have given more than $300 million
- Founders PledgeHelped entrepreneurs commit $1.5 billion to charity across 2,500+ pledges
- Giving GreenAdvised millions of dollars towards evidence-based climate work
Ready to get involved?
Want to start diving in?
Try our quick, free course on the world's most pressing problems. Six short units, start anytime, go at your own pace. Think Duolingo for doing good.
- About 100 minutes in six short units
- Free, with no sign-up
- Your progress saves on this device
Prefer to talk to a person first? Come to one of our events, or book a call with one of our co-directors.
The world's most pressing problems · about 100 minutes
Your problem set
- Why numbers matter14 min
- Global health and poverty13 min
- Animals and the moral circle16 min
- Our final century?15 min
- Advanced AI17 min
- Choosing, and what you can do25 min
Ready for more?
The EA Fellowship
A selective, five-week programme for people ready to go deeper. Expect about 3.5 hours a week of reading and small-group discussion. Participants score it 8.7 out of 10 on how likely they are to recommend it. It's a real commitment and places are limited, so we ask you to apply. The next round starts in early 2027.
Join the waitlist