Thanks for the post! It’s great to see analysis of the LEAF data and engagement with existing EA Survey data.
much of the current research suggests these events [conferences, local groups, and educational programs] to be largely ineffective in encouraging participants to engage further with EA communities
That is not my impression of the existing data.
For example, you cite the 2019 cause prioritization report to say that:
Rethink Priorities’ analysis of the 2019 EA Survey found that 42% of respondents reported changing their primary cause area after becoming involved with an EA community. However, relatively few respondents had actually made career or behavioural changes to align with EA priorities.
I’m afraid I don’t understand the reason why you think this post suggests that claim. That post addressed cause prioritization, not behavioural changes, and I don’t think whether people changed their cause prioritization since joining EA is a good proxy for them making changes to align with EA priorities. Most respondents already supported EA causes at the time of joining EA(though many switch between causes or change their relative prioritizations over time).
In the report on Engagement from that same year, we find that large numbers of EAs are taking actions aligned with EA priorities (e.g. making EA donations, changing their career plans, volunteering or working in EA jobs, etc.).
a 2023 study of EAGx conferences, which compared the attitudes and behaviours of attendees with non-attendees, found no statistically significant differences between the two groups
Participants’ responses were tracked before and after the Leaf 2025 course to evaluate belief changes. Overall, the dataset revealed very little net change in views towards the statement ‘we should prioritise what is evidenced as best over what we emotionally prefer’.
I’ve not dug into the LEAF data in detail (and thank you again for analyzing it). But it looks like the main reason why there was very little increase in people’s agreement with this statement was because respondents overwhelmingly agreed with it even in the pre condition. Mean ratings were 6.05 out of 7 at the start of the course, leaving almost no room for the score to go up.
Thanks for the post! It’s great to see analysis of the LEAF data and engagement with existing EA Survey data.
That is not my impression of the existing data.
For example, you cite the 2019 cause prioritization report to say that:
I’m afraid I don’t understand the reason why you think this post suggests that claim. That post addressed cause prioritization, not behavioural changes, and I don’t think whether people changed their cause prioritization since joining EA is a good proxy for them making changes to align with EA priorities. Most respondents already supported EA causes at the time of joining EA (though many switch between causes or change their relative prioritizations over time).
In the report on Engagement from that same year, we find that large numbers of EAs are taking actions aligned with EA priorities (e.g. making EA donations, changing their career plans, volunteering or working in EA jobs, etc.).
I couldn’t find a post with the title you gave, but perhaps you are referring to this one? While I was very glad that they did the study, as I commented at the time, it was extremely under-powered, so finding non-significant effects was not surprising.
I’ve not dug into the LEAF data in detail (and thank you again for analyzing it). But it looks like the main reason why there was very little increase in people’s agreement with this statement was because respondents overwhelmingly agreed with it even in the pre condition. Mean ratings were 6.05 out of 7 at the start of the course, leaving almost no room for the score to go up.