How you can contribute to the broader EA research project

by Robert_Wiblin 17th Feb 20164 comments

13


A common request I get working at 80,000 Hours asks how people can contribute to effective altruist related research while students, or working in another career. I often find this hard to answer because I don't have an suitable project to hand.

All of the research groups have found that doing high quality work is more difficult than expected, and those people who just dabble rarely produce breakthroughs. This is because it's challenging work that doesn't suit everyone, and for people who it does suit, they get better as it the longer they spend on it.

What can even the cleverest short-term volunteer contribute to theoretical physics? Probably nothing.

However, this is not theoretical physics, and I think a wide range of people can make useful contributions by uncovering important information we are currently ignorant of.

What are some ways?

  • Produce a shallow review of a career path few people are informed about, using the 80,000 Hours framework. We score career paths on the following criteria out of 5:
    • Earning potential.
    • Direct impact.
    • Satisfaction.
    • Advocacy potential.
    • Career capital (comprising skills, credentials, network, exit opportunities and character).
    • Ease of competition.
  • We then write up a profile with the following heading:
    • What are the pros of this career?
    • What are the cons of this career?
    • Who would be best suited to it?
    • What important things are we still uncertain about?
  • You will probably need both online research and interviews with people taking that path. I made a list of resources we regularly consult.
  • If you spend a few days looking into a career path you might want to take, using these criteria, you can probably produce a blog post other people will find useful, and that we could adapt into a profile on our website in future.
You'll want to read a bunch of 80,000 Hours career profiles first to get the hang of it. We can't promise to put what you write on our site - our own work is usually edited by 3-4 people before being published - but may well refer people to it in coaching if it seems relevant.

More difficult is the following:
  • Become highly informed about an approach for improving the world, or a problem people could work on, and assess it on the following criteria:
    • Importance (how much good might be done solving the problem?)
    • Neglectedness (who is already working on it, including indirectly, and how good are they; is there much left to do?)
    • Tractability (how hard is the problem to solve; how expensive are the possible solutions; how likely are they to work; do we have suitable skills or could we develop them?)
    • If you can combine the above into a meaningful cost-effectiveness estimate then go ahead, but if not, don't despair.
  • Write up a summary of the above information with citations, comment on whether you think the topic deserves further research (i.e. could it be better than what some people are already doing), and what the key remaining questions are.
  • Concrete numbers (even if estimates) are particularly useful inasmuch as any can be found.
  • I would have suggested evaluating charities, except that this usually requires access to staff and sensitive information which they aren't willing to provide to anyone who asks.
Ideally you'll aspire to produce something like GiveWell's cause and intervention overviews.

For example, Anj wrote about using gene drives to get rid of mosquitos, and I wrote some thoughts on taxation of the very wealthy - something I hadn't seen anyone here write about despite its potentially high importance.

These are only the first bricks in the wall on these questions, but at least they bring some facts to the table and can shape what questions full-time researchers pick up next.

Distinguishing yourself with sound write-ups is probably the single best way to sharpen your skills, or get hired as a researcher. If 80,000 Hours found someone writing career profiles we thought were as good as or better than what we were currently producing, we would seriously consider hiring them.

The third approach is just to work on a good research question within your own field of expertise. The best single heuristic I know for identifying these questions is to ask the luminaries in your area which questions they think are more important, tractable and neglected. If they could start again what would they work on? Alternatively you could try to evaluate your potential research questions on these criteria yourself. An example off the top of my head might be a statistician using publicly available data to see what metrics are leading indicators of social unrest or civil war (e.g. changes in GDP, tax rates, terms of trade, etc). An impressive post of this kind was one on malaria and labour productivity.

However, if you are young, it may be best to focus on topics that primarily build your skills and credentials, rather than bite off a very difficult though important problem and fall on your face.