Yeah, fair point. No doubt there's sub-groups that signs up because they like signing up to things, or liked the idea at the time but lost interest etc. These folks won't engage no matter what time horizon.

But equally there are sub-groups that are insufficiently engaged at the moment due to circumstances constraining them - major life changes (job change/loss, marriage/divorce, adding or losing family members). These folks would be coming out of such circumstances eventually and at that point would be looking to engage. My over-1-year-late reply is an example of this!

Well, between relocating and having kids, one of those decisions is far more irreversible so should be more carefully made. It's one of those rare one-way doors, and you won't pass through many of those over your life.

I summarised Elliot's post in my mind as a form of reflecting on how a decision affects you over differing time horizons, as well as the second-order consequences of a decision. I would add to his bullet list of reasons to return home if you plan on having kids: going home for kids retains more of your discretionary time.

Having kids is a transformative experience and as such by its very nature you can't fully predict how you'll come out the other side of it. Elliot's post does a good job of stepping that out for us, and lilly's comments ask some valuable questions about messaging: appealing to older folks that are not EA-aware could be better reached. I resonate with both paragraphs of lilly's comment. At the time of my reply I can see three detractors from lilly's comments -- I'd like to have read their responses here why they disagreed.

Thanks James and Miles, I appreciate your summary at the start of this post. Good to read this and important work to attempt to validate the core aims of the conference. Couple of ideas for survey tweaks:

1. Push to get higher response rate for control group, too. This should give a more representative sample of non-attendees. A 5% response rate is probably the self-selected “keen beans” who may be more likely to stay the course in their views and actions (rather than decay over 6 months), and who may just be too busy for that particular year’s EAGx.
2. Add (or adjust) a survey step much later than 6 months. In my view, 6 months is too short to lose social connections for the control group. Such connections are lost over longer periods (1.5+ years), and especially less likely to be lost by the “keen beans” who respond to the survey. These may be more likely to be conscientious or proactive respondents who stay in touch with connections more often.