With a lot of investment in intentionally growing the EA movement this year, I'm curious if some concrete activities have caused EA to grow in noticeable ways.
In particular, I'm interested in looking at the impact of...
- The publication of “Doing Good Better” by Will Macaskill
- The publication of “The Most Good You Can Do” by Peter Singer
- EA Global
- GiveWell announcing its new top charities + associated mass media
- Giving What We Can’s pledge drive
- Google searches for “effective altruism”
- Visits to the EA Forum
- Amount of posts on the EA Forum
- Number of people who have joined the EA Facebook group
- Number of posts on the EA Facebook group
- Visits to the GWWC website and sign-ups to their newsletter
- Number of Giving What We Can pledges
- Visits to the GiveWell site and sign-ups to their newsletter
- Visits to 80,000 Hours website and sign-ups to their newsletter
- Visits to effectivealtruism.org and sign-ups for the EA newsletter
- Number of mentions in the press by people unaffiliated by EA
- Number of people who have created an EA Profile or put themselves on the map of EAs
- Visits to TLYCS website
- Signups to TLYCS newsletter
I think doing this analysis could provide some good time-order evidence of causality between certain EA activities and the growth of the EA movement in different areas. It might also allow us to make further judgements about (a) whether certain effects are just “flashes in the pan” or sustainable and repeatable ways to grow the movement or (b) whether certain actions grow some parts of the EA movement but not others.
Of course, doing such an analysis is a bit reductive and the raw numbers might not be good enough to capture all the ways in which the EA movement grows, especially on key metrics we might care about, like growing existing EAs, growing donation totals, coming up with new EA ideas that improve the movement as a whole, and creating significant career changes. Still, I think having this analysis could be helpful for deciding where to invest in the EA movement.
Furthermore, it’s possible that some of the dependent variables (e.g., visits to the EA Forum) might be affected by things outside the event of interest (e.g., during the same time of EA Global but independently of EA Global, there is a campaign to post EA Forum articles on Reddit, which brings in traffic). This is a limitation of any time-order but non-experimental analysis, but the hope is by tracking enough variables that this limitation can be smoothed out.
I unfortunately don’t have much time to think about this or implement it myself, but I think it is important enough to put some money on it. I’m willing to offer...
- $2 to anyone who suggests an event or variable to look at that I merit is worth considering
- $10 to anyone who improves this study design itself in a way that causes me to update this document
- $100 to whomever successfully implements this study and publishes the results publicly (though email me at email@example.com if you intend to do so, to avoid duplication of effort).
All the data needed for the study should be readily available -- contact me by email if you need help accessing a particular data source and I can tell you where to look or whom to ask for the data.
The money will come in the form of either me giving the money to you over PayPal or through me donating to a charity of your choice. The money is coming from my personal savings and would counterfactually not have been used for any donations.
Update: Eric Yu did this, writing "Effects of major events on EA activity". (Here's my $100 donation.)
The main things I've noticed at 80k:
We also get a lot of benefit in being able to introduce new people to members of the community, and my sense is that the quality of the introductions we're able to make keeps improving.
I'd suggest further factors:
Good idea. I'll add that.
I'll add that too. Is GWWC active on Twitter though? I think visits and newsletter sign-ups are likely good enough.
Also a good idea. Do you know which FB page they have? Are they active on Twitter? Again, I think visits and newsletter sign-ups are likely good enough.
My best guess is that these local groups would be driven mainly by LEAN's activities and less by the growth of the movement as a whole. I'd expect to be an important metric of movement growth but more of a lagging metric or an input rather than an output metric to track.
Which newsletter is that?
This sounds really hard to track.
My guess is that these would be lagging metrics that would be harder to track in time, but they'd definitely be good things to have on any EA growth dashboard.
(Let me know if you want me to PayPal you $2 or donate it wherever you wish.)
Here's the GWWC Twitter: https://twitter.com/givingwhatwecan
Here's the GiveWell Twitter: https://twitter.com/GiveWell
Here's the GiveWell FB: https://www.facebook.com/GiveWell.org/?fref=ts
Here's the link to the EA local newsletter: https://eahub.org/groups/resources/newsletter
Donate to Intentional Insights: http://intentionalinsights.org/view/donate
Emailed you about it - I'm interested. Is anyone else already planning on doing this?
Replied. You're the first person to offer to do the analysis.
Good idea. I'd also be interested in data on how many meetups and talks EA give in different cities and what impact that has. If more meetups and talks in a city translate into significantly more growth that would seem to be a good reason to put more resources into such activities.
That seems pretty plausible to me but I don't know how to collect that data. Do you have any ideas?
LEAN is capturing it in our annual questionnaire for group organizers.
Great! What patterns have you spotted so far?
We'll share the full results soon, including all non-confidential raw data. :)
One approach is to run an experiment. You choose a number of similar cities and then randomly increase the number of meetups and talks in one of them. Then you measure, e.g. whether this translates into more members in the local EA FB group (provided such a group exists), more people at physical meetings, etc. (I'm sure there are other ways to measure this; that's just off the top of my head.)
Such an experiment could also be used to test the much-discussed idea of whether fast growth is risky. If it turns out that the faster-growing chapters cause problems in some way, that might be a reason to think that the EA movement should not grow fast, or at least that it should not try to grow fast using the method used in the experiment. If, on the other hand, they don't cause any problems, that is a reason for the movement to try to grow faster. (I favour faster growth, I should add.)
On Meetup you can get data on how many members have joined, how many members have looked at the page in the last three months and how many have said they are going to events.
Do you know if it's possible to get that data across all EA meetups?
LEAN is capturing it in our annual questionnaire for group organizers, and also our monitoring and evaluation of meetup.com accounts in particular (many of which we provide).
Although certainly not the comprehensive or rigorous information you are looking for, google trends can provide some interesting insights on a broad scale. For example, while effective altruism is showing strong growth in terms of search frequency, it is still searched for 7 times less frequently than Peter Singer: http://www.google.com/trends/explore?hl=en-US&q=Effective+altruism,+/m/05xnb&cmpt=q&tz=Etc/GMT%2B8&tz=Etc/GMT%2B8&content=1
May be you could look on translation of main articles into different languages?
Like the rate at which that happens? What kind of data would you envision collecting?