After many months of hard work from everyone on the .impact team, this year's EA Survey results are finally available.
It's a long document (~25 pages), however, so we've put it together in an external PDF.
Introduction
In November 2015, a team from .impact released a survey of the effective altruist community. The survey offers data to supplement and clarify our perception of people who identify as Effective Altruists, with the aim of better understanding the community and how to promote EA.
The results should be useful to anyone involved in movement-building, analysing the impact of the Effective Altruism community as a whole (especially with reference to donations), or anyone who’s interested in a snapshot of what Effective Altruism looked like in 2015.
Summary of Important Findings
- 2904 sincere people took the survey, and out of those 2352 people would consider themselves EAs. All the following results consider only the people who’d consider themselves EAs. This is three times as many people as last year!
- The most popular way for the people we sampled to hear about EA was Less Wrong (20%), followed by ‘personal connection’ (11%) and ‘Book/article/blog post etc.’ (11%), but 20% of people didn’t answer this question. More people heard about EA for the first time this year than any other year.
- 37% of EAs sampled identified Poverty as the ‘Top Priority’ cause area. The next-most-popular top priority cause was prioritisation, with 9.4% of EAs sampled identifying this as the Top Priority.
- 885 of the EAs sampled donated money to an EA or EA-recommended organisation. The most popular organisations to donate to were AMF, SCI, and Give Directly.
- Total donations (in 2014) from EAs sampled were $6,765,244, with the median being $330; this is very skewed by large donors.
- We recorded 56 donating both last year and this year, and the median increase in donation amount was $296
- 436 (37% of those who answered the question) said yes to ‘Do insecurities about not being “EA enough” sometimes prevent you from taking action or participating more in the EA community?’
- 717 (64% of those who answered) said that EA was welcoming, 103 (9%) said that EA was unwelcoming.
The Full Document
You can read the rest at the linked PDF! -->
A Note on Methodology
One large concern is that we used a convenience sample, trying to sample as many EAs as we can in places we knew where to find them. But we didn't get everyone, and those who replied are not likely to be representative.
This year we initially launched the survey on the EA Facebook page under strict instructions not to share it further, and so we can be fairly sure that the initial group of people sampled were all members of the EA Facebook Group, although not necessarily representative ones. This gives us a benchmark to compare the other subpopulations against.
As we said last year,
"It’s easy to survey, say, all Americans in a reliable way, because we know where Americans live and we know how to send surveys to a random sample of them. Sure, there may be some difficulties with subpopulations who are too busy or subpopulations who don’t have landlines (though surveys now call cell phones).
Contrast this with trying to survey effective altruists. It’s hard to know who is an EA without asking them first, but we can’t exactly send surveys to random people all across the world and hope for the best. Instead, we have to do our best to figure out where EAs can be found, and try to get the survey to them.
We did our best, but some groups may have been oversampled (more survey respondents, by percentage, from that group than are actually in the true population of all EAs) or undersampled (not enough people in our sample from that subpopulation to be truly representative)."
This is a limitation that we can’t fully resolve, although we have tried to make some headway by using the staggered-release mechanism described above.
At the bottom of this report, we include a methodological appendix that has a discussion of the limitations of convenience sampling, and a comparison of the different subpopulations in the survey, ultimately concluding that the data we have doesn't allow us to detect a statistically significant difference between different subpopulations in donations and primary cause choice, although certain demographic indicators - such as meat consumption and gender - are different between the subpopulations.
-
Overall, this is the most comprehensive survey yet of people who identify as Effective Altruists, and should help to inform discussion about the movement for the next year.
I have two questions.
(1) The number of donors per charity was included and even broken down by referral type, but the total amount of money moved per charity was not (I believe) disclosed. I'm wondering if this is data you intentionally left out of the analysis (for privacy reasons) or just happened not to include. I would be interested in that data since it can help convey the seriousness of donations. If privacy considerations are an issue, you might wish to disclose the number only for the charities that got at least ten donors.
(2) Somewhat related to Dan Keys' point: the feedback time between the year of donations (2014) and the time the results are released (July 2016) is larger than ideal. You mentioned that the reason for having to do it this way was that the survey was conducted prior to the end of 2015. I'm wondering if it might make more sense to conduct the survey around the end of January, and then have the results released a little before the Giving Season of the next year. Is this something you've considered, and/or are there other ways you expect the feedback loop to be shorter in subsequent years?
(1) We didn't ask people how much money they donated to individual charities, that's right. The data is available in the github repository for the project - search for 'github' in the report
(2) I agree that conducting the survey at the start of the calendar year would be better. Whether we would do that depends to some extent on whether we'd want to wait six months until we start the next survey. We are tightening up the feedback loop - we're improving the code used to analyse it every time. This year the survey was handed around quite a few people - we hope next year to have a dedicated person who can focus entirely full-time on it.