University Groups Should Run Socials

This post is for university group student organizers. We start with the big reason you should run a social, three tips we’re highly certain of, and some additional thoughts. 

We think ~40% of the value of this post is in a tentative answer to the question of how university groups can best provide value to members post intro fellowship; this is a question almost all university group organizers should think more deeply about. Another ~50% is in the “Big intro fellowship” section (tip 1), which is sufficient to run a fairly successful weekly meeting, and we are highly confident that almost all university groups should adopt the “Big intro fellowship” even if you do not choose to run a social.

We both co-lead the University of Chicago group and all data comes from the UChicago group unless otherwise noted.

If this post makes you consider a change at your university group, please reach out to us ([email protected] and [email protected]); we'd be excited to talk with you about your particular circumstances.

Quick definitions to avoid confusion:

• Social: A weekly meeting where there is loose, if any, structure. The purpose is solely for members to talk with each other.
• Big intro fellowship: Rather than meeting separately in small groups according to a when to meet, there is a prescribed time when all intro fellows meet together in the same room before being split into their small groups with a facilitator. Importantly, this prescribed time is right before the weekly meeting/social to minimize frictions for intro fellows to attend the weekly meeting/social.

Why a Social

Claim: We are 80% confident that (1) doing a social and (2) doing a big intro fellowship will dramatically increase your weekly attendance dramatically (>2x) conditional on your club being relatively established and there being other EA-adjacent clubs (like AI Safety) on your campus.

University EA groups have converged on the intro fellowship, a 6-10 week reading group as the standard method of introducing new members to the concepts of Effective Altruism. While some tweaks can likely be made, we suspect returns would be low. The proof here is that across schools and across time, similar student demographics and student body populations lead to similar intro fellowship sizes and attrition. We're doing a pretty good job!

In getting those who completing the intro fellowship to continue thinking about EA, we are doing decidedly worse. And the evidence here is that similar groups have very different retention after the intro fellowship. Last year UChicago, despite have ~30 students complete the intro fellowship over the year, had attendance at weekly meetings of 0-5 members outside of leadership. Last quarter (UChicago runs on a quarter system; each quarter is roughly 10 weeks; we run intro fellowships and weekly meetings from weeks 3-8), we averaged 13--and 17 in the 3/6 weeks where we had food. More anecdotally, other groups in the midwest like WashU and Northwestern have in past years had similar intro fellowship sizes as us and also had similarly disappointing attendance at weekly meetings. They have not yet implemented the social but we hope they do and report back!

So to answer the question, the reason to do the social is that it works. And while sample size of 1 quarter is hardly persuasive, the mechanisms make sense (and I can promise that we, Noah and Avik, are not more hardworking organizers than past UChicago EA leaders):

• Big intro fellowship reduces the friction for intro fellows to attend. The default is for them to stay put and join the social rather than them having to decide to make the trek to attend each week.
• Intro fellows being at the social automatically sets a floor for attendance such that others who might attend know they won't be stuck in an awkward meeting with 3 other people.
• Socials, as opposed to general meetings, are consistent so older members can be confident they'll enjoy. We've heard this anecdotally where older members appreciate that each week is a chance to hang out with friends rather than sometimes being thrust into activities abt topics they're familiar with and thus find somewhat awkward.

Doing a Social Well

1. Have a “Big intro fellowship” right before social: We actually began our “Big intro fellowship” experiment in fall quarter. Instead of having students meet at separate times according to a when2meet, we picked a time (Thursday 6-7 for us) and had everyone meet in one big room where we split into our 3 smaller breakout groups. This was key to increasing social attendance. We had our social in the same room the hour after the intro fellowship so there was zero friction for intro fellows to attend. This, in turn, made it attractive for others to come because they knew there’d be at least 10 people at the social (the leadership team + at least half the intro fellows). 

   Yes, this means some will not be able to make the intro fellowship time, but for us, we are fairly confident this comes out to losing ~25% of intro fellows. And most of that 25% is not lost to you forever. By staying in touch with these members, we were able to get a few fall drop offs to join the intro fellowship in winter quarter when they no longer had a time conflict.

   We would also note that streamlining everything into one 2 hour block a week makes your life as an organizer way easier.

2. Text members before each social: While we don’t have numbers from the EA group on this, we do have numbers from the UChicago Rationality Group, which Noah runs. For the Rationality Group, when Noah started texting people, he went from 7 attendees on average to 12-15. We see texting as 1) a nice reminder about the social and 2) a better social nudge – receiving an email on a big list host or Slack is less social pressure than a text (you should also send the listhost email to be clear). To streamline this process, we have tried using a mass-text app called ‘hit em up’ which can send mass text messages that include the recipient's name. Honestly, both of us have reverted to sending texts individually since it allows you to account for whatever ongoing conversations you're having with an individual member, but even this takes <20 minutes for texting 40 ppl.

   We understand this might seem rude to you at first, but we are pretty confident you won't offend anyone. Only one person has ever asked us not to text them (and they weren't pissed; we just took their name off the list and they still get the email and sometimes come to socials). On the other hand, both of us have received numerous thank you's for doing the text reminders. And if done thoughtfully, we've found it even helps build relationships.

3. Invite other clubs (and especially their leaders): UChicago has four EA-aligned clubs on campus: AI Safety, Alt-Protein, Rationality Group, and the EA club. People in these clubs disproportionately enjoy the kinds of conversations and activities we have during EA socials. While we appreciate a principles-first approach to EA (which some of them share), having more interesting people at a social generally improves the vibe—both because of the diversity of conversation topics, the joy of meeting like-minded new people, and the novelty that new faces bring. Because of this, we’ve found immense value in reaching out to individuals in these clubs and asking leaders (especially EA-aligned leaders) to regularly promote EA events and socials in their group chats or during meetings (Noah does this, for example, during Rationality Group meetings). We’ve also had success doing this with less explicitly aligned clubs—like economics or philosophy groups—as this helps both with recruitment and with filling up space at socials. People who regularly attend socials tend to deepen their engagement over time (e.g., by doing an intro fellowship or going to EAG), and socials are a great on-ramp for intellectually curious people to slide into EA ideas without commitment to a fellowship.

   This was honestly our original motivation for the transition to the social. By moving away from more EA specific content, it was easier to convince students from other clubs to come.

Some additional thoughts

• You might have some concerns. A few that we think can be mitigated:

  1. How about continued member education?

     If you're concerned that if the social replaces a general weekly meeting where you can expose newer members to ideas that go beyond the intro fellowship, this is fair. By definition, a social is unstructured, so while conversations will often connect current events to EA ideas and cause members to go read blogposts they heard referenced, you cannot easily ensure specific ideas get covered.

     Luckily, this can be mitigated by having one on ones. You should be having one on ones with intro fellows regardless and if you have these check-ins, that's a great time to check if there are any gaps and talk through/recommend reading materials for whatever topics you deem prudent.

     (Also, we think that continued member education is hard in general. Even if people do show up to listen to a lecture, and even if they do consistently, people don’t easily apply the ideas to their own lives. In this way, we prefer one on ones and the discussions that happen in socials because they tend to be pointed towards concrete actions or at least concrete claims.)

  2. But what if my group isn't established?

     This is also a very fair critique where we are less confident of the answer. If your group is all new members and you don't have other EA-adjacent clubs on campus, a weekly meeting might not be much worse than a social in terms of generating attendance.

     All we can recommend here is that you try both and report back.

• You should probably get food: We ran 6 socials last quarter. The 3 we did with food both had stronger attendance and ran longer (our socials are 7-8 but usually run over until conversations die down). 

  	With dinner provided	With only snacks
  Average Time (min)	85	68
  Average Attendance (excluding leadership)	17	9

• The data above is not sufficient to make us very confident about the claim that food increases attendance because we were trying many new things during last quarter and there could easily be confounding variables. We initially thought metrics like retention could be useful, but these were mostly functions of the total number of people who showed up. We did notice:

  • Intro Fellows were more likely to stay when they saw food. We heard a few times that because this obviated the need to otherwise go to the dining hall, they were more inclined to stay.
  • Mentioning food in texts did get some people excited

  We are more confident that food increases quality of conversation. There was both qualitative difference in the excitement during the social and people could stay longer (in part because they didn't have to grab dinner).

• Some loose structure to your socials is good and easy: We experimented with completely unstructured socials (people came and sat down and started talking without guidance) as well as structured socials (we had a few topics around the room or some game or something for people to at least initially start with). In general, we heard some feedback that the structured socials were easier for the intro fellows to engage with.

  Some particular ideas we tried (if you want to implement any of these, we'd be happy to talk more about this)

  • Chalk talks: interested students write a claim (on a chalkboard or a post it they stick to the wall) and then talk about it with anyone else who wants to engage.
  • Speed Updating: everyone reacts to a list of printed out prompts and ranks their agreement 1-5. They then show their sheets to eachother and discuss their largest disagreements.
  • Career talk: a group leader discusses concrete career decisions they're thinking over and how they're evaluating different choices for impact.