Anonymous asked: How do you currently run your effective altruist groups? What did you do to increase their size that worked? What did not work?

Our group started out by meeting once a week with a discussion topic. Some groups are mostly project-focused rather than discussion-focused; it seems to me that those groups have a harder time attracting members and a harder time building a community, because projects feel like work and it’s hard to get college students to show up for extra work. It’s easy to get them to show up for food and interesting conversation (the food was a late addition).



Every week I write a discussion questionnaire that usually has about ten questions related to the topic. The questions are meant to set up the possibility for the discussion to go in several different directions, and to hit on lots of potentially interesting things so everyone can look at it and see something they want to talk about.

The top of the questionnaire has a prompt we’ll do for introductions at the start of the meeting. This was my effort to compromise between ‘introductions make sure everyone gets a chance to talk and to know each other’ and “i hate hate hate being forced to come up with something interesting about myself on the spot’. Since it’s on the worksheet, you have lots of time to think about it.

Example questionnaire:

topic: flowthrough effects in global poverty introductions: what is something you believe that most of us probably disagree with you about? questions: GiveWell’s estimates of the effectiveness of the top global poverty charities looks mostly at the direct effects (saving one person from malaria counts as doing one person’s worth of good, etc). But lots of people give in the hope that they’re fundamentally changing a situation. How much do you weigh flowthrough effects?



Imagine AMF and simplified!AMF, where simplified!AMF does the same things as AMF but is stipulated to have no flowthrough effects at all. How much better do you think AMF is than simplified!AMF? Some possible sources of flowthrough effects from health interventions are economic growth (less disease means people can be more productive), family dynamics changes (lower infant mortality means less need to have lots of kids), reduced transmission of the targeted disease overall, reduced susceptibility to other diseases, and more education (less disease allows people to stay in school and means they can perform better). Can you think of others? Which do you think are most significant? What sort of evidence would you look for? There are also negative flowthrough effects: crowding out of local economies in favor of foreign aid, promoting antibiotic resistance, the rich meat-eaters problem, risks of economic growth. Can you think of others? Which do you think are most significant? What sort of evidence would you look for?

We all introduce ourselves, which takes around 30min. The group is fiercely divided over whether that’s far too long or the best part of the meeting. I’m experimenting with gentle ways of speeding the introductions along, but I’m on side “best part of the meeting” because it usually turns into a chance for everyone to share a tidbit of something that fascinates them.

Then someone proposes a questionnaire point that they want to talk about, or someone argues that a question on the sheet is poorly-formed or completely wrongheaded (this happens pretty often), and makes a case for a better question, and we let the discussion proceed naturally from there. I intervene only if

1) the discussion has turned into a back-and-forth between two people

2) the discussion has gotten jargon-y and inaccessible and people are looking confused

3) the discussion is inappropriate

In the case of 1) I say “A, B, I think you have an empirical disagreement, I think you’re not going to settle it, we’re moving on to C’s point from earlier.” In the case of 2) I say “A, want to give us the 60-second summary of the Julia Wise post you just referenced? I doubt everyone has read it.” 3) happens very rarely and ‘A, I’m not cool with this conversation, let’s talk about X’ always resolves it fine.

I think most people will be inclined to intervene a lot later and a lot less than optimal for a lively discussion group. I think discussion leaders should be pretty fast to jump on the above problems. People seem to be pretty naturally high-variance in how comfortable they feel doing this, so your discussion leader should be someone who feels comfortable doing this and is widely respected in the group. The fact I write the questionnaires helps me lead the discussions because it’s an obvious source of legitimacy.

After the meetings, which last 1.5 hours, we all go to dinner together. The people who aren’t Stanford undergrads get guest meals from the people who are. Dinner has done a lot to help the group grow; food is always appealing, it’s informal, it lets the discussion splinter into several separate ones, and it lets people feel closer to each other. Some people skip the meetings and show up for dinner. This might be a good thing, because the meetings are consistently at the upper limits of a good size for a discussion group. (Sometimes we split into two discussion groups of ten. This helps make the discussions more manageable, but I personally can’t concentrate when there are multiple conversations going on so I wish I had a better solution.)



Once we were starting to approach that upper limit for the Sunday meetups, we added another meetup, on Fridays, which is purely social (no discussion topic). There’s a lot of overlap between the two groups, but also lots of people who prefer one or the other. As a result this has been a good way to grow the group. If I weren’t worry about sustaining momentum when I graduated, I think I’d start a third event - a group study hall or something - once both of the existing groups are as large as they can reasonably get.

As for getting people to show up - if the group is talking about cool stuff, they show up. We’ve tried doing anonymous surveys about how people find the group environment, and gotten no responses; we’ve tried asking people, and gotten pretty neutral ones. But if the meetups are fun and interesting, people keep coming back. It also works well to reach out to people individually and say “hey! I think this week’s topic will interest you!”

One thing I’d like to do better on is making sure people know their presence is appreciated and that we want to see them again - maybe by messaging them on Facebook post-event to say “hey! the conversation moved on before I got a chance to say so, but I really agree with your thoughts on [whatever].” Likewise, recruiting works best on the individual level - find a friend, say to them “hey! I suspect you’ll disagree with us about lots of stuff, but honestly we all disagree with each other about lots of stuff, and I bet you’ll find the discussion really interesting. Come with me this week?” Most efforts at large-scale recruiting don’t seem to work, unfortunately, but individual-level works quite well.