What Activities Do Local Groups Run

Speaking to a number of group organisers and movement builders at the recent EA Summit, it became clear that, collectively, we don’t have a great sense of exactly what most EA groups are actually doing and how they are doing it. For example, some people I spoke to thought that most EA groups spend most of their time trying to do mass outreach, where others thought that almost no EA groups engage in mass outreach. Similarly, some thought that most groups engage primarily in unsystematic and unstructured activities (such as unstructured socials), whereas others thought that most groups ran largely structured activities. Few of us on either side had very much other than than our respective vague senses of what groups were up to back up our intuitions.

This highlights the need for more systematic empirical investigation of the EA movement and local groups specifically, so that our view of the movement can be informed by something better than our vague impressions.

In order to offer some insight into the question described above I report data from last year’s Local Groups Survey (run as part of the LEAN Impact Assessment), which was not included in our previous reports.

Group Activities and Events

Out of the 98 group organisers who responded to our survey, 78 (79.6%) responded to questions about the number of events they ran of different types and the number of attendees. 

Table 1 shows the number of groups running activities of different kinds.

Event Type

Number of Groups

% of Groups

Discussion Groups



Speaker Events






Committee Meetings



Coworking Sessions



Career workshops






Research Projects






Table 1

These results suggest that discussion groups, speaker events and socials are the most popular events (for groups to run) by some margin, with the former two being significantly more popular than socials. Committee meetings were also run by a slim majority of groups, while the remaining event types lagged far behind with only 9-27% of groups running them at all.

All in all, I take this to be somewhat suggestive that groups are primarily running events that at least have some structure and substantive content rather than only only socials. Of course, it remains to be seen how much structure and quality content is included in ‘discussion groups.’

At first glance it might appear strange that only 67-76% of groups are running socials, discussion groups and speaker events (seemingly core EA group activities). However, in fact, almost all (75/78) are running at least one or some combination of these events. 2 groups in the sample had, thus far, only held committee meetings (and no public events), and 1 group reported specialising only in running research project events.

Table 2 shows the total number of events of different types run across groups, rather than the total number of groups running certain types of event. For example, it might be that the same number of groups might run discussion groups and socials, but that all groups run many more social events than discussion groups.


Event Type


% total events run

Discussion Groups






Committee Meetings



Speaker Events



Coworking Sessions



Career workshops



Research Projects









Table 2

This table shows a similar ordering of events, with discussion groups still dominating. Notably speaker events, despite being run by many groups, were run in smaller numbers. This is perhaps unsurprising given that speaker events are may be more costly in time and effort (and sometimes expense) to run. Notably our earlier qualitative interviews with group leaders suggested that lack of access to speakers for events was a significant bottleneck for groups. Nevertheless, organisers reported running 317 such events, which is still considerable compared to the other top event types.

It is also worth noting the strikingly high number of coworking sessions run by groups. Despite only 27% of groups running these sessions (about 1/3rd as many as the number of who run discussion groups or socials), the number of coworking sessions run was more than 1/2th the number of discussion groups and social events run. This is explained by the high number of coworking events run by a small number of groups.

We also collected data on the number of attendees at events across groups, shown in Table 3.

Event Type

Mean Attendees

Speaker events


Discussion groups






Committee meeting




Research projects


Career workshops




Table 3


These numbers are pretty intuitive, with speaker events, EAGx’s, career workshops and fundraising events receiving higher numbers, committee meetings, co-working and research projects smaller numbers and discussion groups and socials receiving similar (middling) numbers of attendees. Notably, this shows that it is not the case that socials dominate other events due to receiving much higher rates of attendance. It is also worth noting that the typical size of speaker and discussion group events are both quite small, with the median number of attendees for speaker events (which might often be considered more public outreach events) being 25 and 9 attendees for both discussion groups and socials.

How Structured are Group Discussion Events? 

To gain some insight into how structured events typically were, I ran a poll in the EA Group Organizer’s Facebook Group asking how structured their discussion groups were. I focused on discussion groups specifically, because these were the most common event run by the largest number of groups, and because other events seemed intrinsically structured (e.g. speaker events, career workshops) and I stipulatively excluded ‘purely’ social events from consideration in the poll.

I included 4 options in the poll: 

1) Highly structured: following extensive, detailed pre-prepared materials (e.g. a planned curriculum or discussing an EA article in detail)

2) Fairly structured: following loosely prepared materials (e.g. a significant number of specific questions about a topic are prepared, to guide discussions)

3) Fairly unstructured: a general topic (Giving Now vs Giving Later) is specified, but little more is prepared

4) Highly unstructured: we meet to discuss 'EA' but don't plan anything else in advance

Responses fell almost exclusively within the middle of the range (either fairly structured or fairly unstructured) with a clear predominance of more structured responses (21 fairly structured vs 14 fairly unstructured and 3 highly structured response).

Significance of Different Events for Group Development

In the Local Groups Survey, we also asked organisers "Are there any specific opportunities that had a large impact on your group's development?" (open comment). This had not been intended as explicitly discerning the importance of events relative to other factors, however, when I coded these responses, out of 35 organisers who cited some factor, 25 (71%) cited an event that they had run. Of the remainder, 3 cited external events that they had been able to attend, and 7 cited other non-event factors.

Among those citing events, 8 cited speaker events, 4 giving games, 4 career workshops, 3 pledge events and 3 discussion groups. Notably, 10 explicitly mentioned the importance of outside EAs visiting their, of which 5 mentioned Peter Singer specifically. 

Besides suggesting that events are typically highly important for groups' development, it also seems to suggest that a wide variety of different kinds of events are critical to the development of different groups.


Overall these numbers show that a variety of activities are supported by local groups, including primarily discussion groups, speaker events, as well as socials, with a substantial number of coworking sessions run by a small number of groups. This seems to suggest that groups are running a large number of events that are somewhat structured and contain substantive content, rather than simple socialising. Moreover, the events that groups run tend to be more structured than unstructured, though more research is required to get a clearer picture of the details of how different events are run, which events are more effective and whether events are targeted more at newer or more established EAs.


5 comments, sorted by Click to highlight new comments since: Today at 3:22 AM
New Comment

Coworking sessions sound interesting. The fact that few groups utilize them, but those that do do it apparently very frequently, seems to suggest that it may be underrated. Could people from groups that do this on a regular basis elaborate on the format? Is it about organizing the group itself, i.e. preparing events etc.? Actively working on research topics? Or just generally people from the group meeting to work on things they personally need to get done? Would you say this specific setup increases productivity substantially?

I've attended several co-working sessions with EA London. In my experience, they've been 2-5 people working independently on stuff they need to get done. For example, on a typical day Holly might be responding to messages and emails, David might be writing the EA London newsletter, and I might be doing a literature review for my dissertation. It's nice because we occasionally bounce ideas off of each other, and because eating lunch together is much more pleasant than working and eating alone.

My cautious guess would be that a bottleneck is groups actually identifying something to work on. Once they've set their course and found a good opportunity, from then it flows much more easily.

This was an interesting read, thank you.

I have one methodological question: are the 98 group organisers that responded to the survey all from different groups? Or are a few of them from the same group, and if so, what does the data look like after correcting for this? (I fear the data may over-represent bigger groups, as they'd have more members responding to the survey.)

EDIT: found it, from the 2017 survey report: "Where groups had more than one organiser, they were asked to nominate one organiser to complete the questions designated for organisers." So never mind, then!

Yes, as you saw.. we separated out data from organisers and data from members. For groups with more than one organiser, we asked for them to nominate just one respondent to answer as 'organiser' and the others to fill out the survey as normal 'members'.