When we plotted average engagement against age first involved, the peak was at 20. People who first got involved at age 18 were less involved on average, and had a similar average level of engagement of people who first got involved at age 40.
Just for the benefit of people who haven’t seen the graph, we also split this by cohort (which year they first heard of EA) and there was no cohort for which the peak was younger than 20.
It’s hard to know what to draw from this (younger and older people probably get less engaged because the community is less well set up for them)
I think the fact that we see this effect across cohorts is some evidence for age (when they got involved) itself driving the effect. People who joined (when they were young) in earlier cohorts will be at least in their early 20s and maybe almost 30 by now. So you might think that they will now have been in EA during the ages which, ex hypothesi, the EA community is better set up for, and it seems like they are still, on average lower in self-reported engagement. Of course, it could also be that how well the EA community is set up for you when you first hear of it is really important, and so people who first hear about it earlier than university age never recover, but it’s not clear to me what the mechanism would be there.
Of course, we are talking about a relatively small group of people who first hear about EA at these young ages: about 15% first heard of EA when they were younger than 20 (but that comfortably includes university age), but <5% first heard of EA when they were younger than 18 (and this is probably an over-estimate because age-first-heard is calculated from reported year when people first heard and their date of birth, so there’s a bit of wiggle room as to exactly how old they were when they first heard.
Just for the benefit of people who haven’t seen the graph, we also split this by cohort (which year they first heard of EA) and there was no cohort for which the peak was younger than 20.
I think the fact that we see this effect across cohorts is some evidence for age (when they got involved) itself driving the effect. People who joined (when they were young) in earlier cohorts will be at least in their early 20s and maybe almost 30 by now. So you might think that they will now have been in EA during the ages which, ex hypothesi, the EA community is better set up for, and it seems like they are still, on average lower in self-reported engagement. Of course, it could also be that how well the EA community is set up for you when you first hear of it is really important, and so people who first hear about it earlier than university age never recover, but it’s not clear to me what the mechanism would be there.
Of course, we are talking about a relatively small group of people who first hear about EA at these young ages: about 15% first heard of EA when they were younger than 20 (but that comfortably includes university age), but <5% first heard of EA when they were younger than 18 (and this is probably an over-estimate because age-first-heard is calculated from reported year when people first heard and their date of birth, so there’s a bit of wiggle room as to exactly how old they were when they first heard.