This post is part of a series of rough posts on nuclear risk research ideas. I strongly recommend that, before you read this post, you read the series’ summary & introduction post for context, caveats, and to see the list of other ideas. One caveat that’s especially worth flagging here is that I drafted this in late 2021 and haven’t updated it much since. I’m grateful to Will Aldred for help with this series.

One reason I'm publishing this now is to serve as one menu of research project ideas for upcoming summer research fellowships.

Some tentative bottom-line views about this project idea

Note that these are my views on how important this sort of project would be currently and how tractable it’d be averaged across different people. But I think polling and message testing will become more important once we have a better sense of what goals/interventions would be (very) valuable if politically tractable (i.e., once we have more strategic clarity). And I think this work would be highly tractable for people with relevant training/experience and setups but would have low tractability for people that lack those things.

What is this idea? How could it be tackled?

This project would involve surveying people about their views relevant to nuclear risk reduction, various potential intermediate goals for nuclear risk reduction, or various potential interventions. This could include questions such as how high they think the odds of various conflicts are, how much they think various actors should prioritise this problem, and how supportive they are of various intermediate goals or interventions (perhaps after being given some info about the rationale, the expected costs, etc.).

This could include testing the effects of presenting people with different phrasings, arguments, information, etc.

This could focus on surveying people who currently make or advise on relevant decisions, people who are listened to by those people, people who might become such people in future, people who are somewhat representative of such people (e.g., similar demographic characteristics), voters in general, or some other group. Results from people who currently make or advise on relevant decisions would probably be hardest to gather but most useful.

This project should probably start by checking for existing polling and message testing relevant to the particular questions or topics one wishes to do polling or message testing on, synthesising evidence from that, and working out what gaps remain.

Why might this research be useful?

Many intermediate goals and interventions for reducing nuclear risk require changing policy. Many nuclear risk reduction efforts are therefore benefitted very substantially by policymakers, policy advisors, and (to a lesser extent) the public being in favour of the efforts or its objectives, or at least not strongly opposed. This project would inform us about what these people are likely to favour and oppose, why, and how to influence that (e.g., with different framings or by providing some key info/arguments).

This could help us decide which intermediate goals and interventions to (de-)prioritise investigating and pursuing, how best to pursue them, and what other things we should do to change people’s views or behaviours (e.g., getting them to devote more resources to nuclear risk as a whole or to be more supportive of particular interventions). For example, this could inform decisions about what public advocacy campaigns EAs should run/fund and what target audiences and messages those campaigns have.

This could also inform things other than policy-focused work, such as what demographic and messaging to focus on for building the longtermist or nuclear risk reduction communities.

What sort of person might be a good fit for this?

The ideal fit for this project would be someone with experience designing and analysing results from polls, message testing surveys, or other surveys. But if someone who lacks that background was interested in this project idea, I imagine that they could relatively quickly get up to speed on the necessary skills to the degree that they can do at least a decent job of this project.

Some relevant previous work

• Rethink Priorities has done/commissioned similar types of work on other topics, including testing different longtermism/x-risk related messages and polling views on various possible x-risk-related policies
• Various non-EAs have done polling and (presumably) message testing relevant to nuclear risk, but I haven’t gathered links

Should we try to convince/fund non-EAs to do this work?

I expect contracting non-EAs to do this project (given some more specific questions/topics that some EA is interested in) would work well and that their outputs would be easy for EAs to use (i.e., wouldn’t require much further vetting, analysis, or extension). However, someone who’s attempted to outsource similar work before told me it was surprisingly hard.