Local priorities research: what is it, who should consider doing it, and why

By Yi-Yang @ 2020-09-06T14:57 (+83)

Epistemic status: 60% confidence that the high level arguments are right. I’ve spent 78 hours on this write-up, and had three rounds of feedback from various EAs about it. I’ve also informally reviewed a few EA group’s community building strategies (I can’t say who these EA groups are since I did not get their permissions).

Please consider filling up this feedback form after you’re done reading this forum post. I wonder what kind of impact it has over EA community building strategy.

Summary

• EAs in later-stage EA groups and in non-EA-hubs should consider integrating “local priorities research” (LPR) into their community building work.
• LPR is about figuring out what certain things we should prioritise working on in a local context (typically non-EA-hubs) to best maximise impact beyond their geographic boundaries or within their local context. LPR work can include: local cause prioritisation, research into specific local cause areas, research into a specific career pathway, etc.
• Later-stage EA groups are likely to be in a better position to work on LPR, but you can review these six criteria to figure out whether LPR work is a good fit for your group: (a) capacity, (b) EA-knowledge, (c) comparative advantage, (d) scope, (e) risks, and (e) opportunity costs.
• There are three main reasons why LPR is valuable:
• It helps provide career insights for community members to figure out their high impact career pathways.
• It helps provide external insights into opportunities and threats, making your community building plans more strategic.
• It helps make outreach more effective, by making outreach more targeted, by improving your group’s credibility, and by attracting more EA-aligned people.
• There are three potential risks to mitigate:
• There is a risk of allocating too much resources into LPR work when there are more valuable activities to do.
• There is a risk of producing poor quality LPR work, which may be harmful.
• There is a risk of being unwelcoming.
• If conducting LPR is a significant improvement to community building, then the EA movement should consider providing more resources to build LPR capabilities.

Acknowledgements

I really appreciate the feedback I’ve received for this write-up from the following people: Catherine Low, Aaron Gertler, Vaidehi Agarwalla, Brian Tan, and Loke Jia Yuan. All opinions and mistakes in this write-up are my own.

1. Introduction

I argue that EA groups that are in their later-stages and in non-EA-hubs should consider integrating “local priorities research” (LPR) in their community building work. LPR is likely to be helpful in (a) producing local career insights to support community members’ high impact career transitions, (b) producing external insights to formulate effective community building strategies, and (c) making outreach more effective. I also caution that there are risks involved with LPR work, especially if an EA group is still in their earlier-stages.

In the next few sections, I will first define LPR, show how LPR might fit into community building work, who might be a good fit for doing LPR, explain the reasons why LPR is helpful, then address some potential risks with LPR. I will also touch on a potential idea the EA movement could work on at the end.

2. What is local priorities research (LPR)?

You’ve probably already heard a lot about global priorities research (GPR), and LPR is quite similar, except that it’s narrowed down to a certain country. While GPR is about figuring out what are the most important global problems to work on, LPR is about figuring out what are the most important problems in a local context that can best maximise impact both locally and globally.

The key difference in this geographic difference is in career pathways. Currently, GPR has already narrowed to some very important cause areas for EAs to work on, and organisations like 80,000 Hours have identified potentially very impactful organisations and jobs in those cause areas. However, not all countries have such impactful organisations and jobs available to them. Hence, LPR exists to determine the most impactful cause areas, organisations, and jobs in a local context.

For example, even if it is not as impactful as working in a top AI lab in the US, China, or Europe, LPR has helped EA Singapore identify technical AI alignment research as one of the most impactful cause areas to work on in Singapore. We also found a few potentially impactful local organisations and jobs that people can apply to in this area.

Furthermore, LPR includes not just cause areas that can create impact within a local setting (e.g. research into the effects of a certain health intervention in a country may only be generalisable in that location), it also includes ucase areas that can create impact cross border (e.g. research into biological risks can help others in different countries). LPR is also the more applied kind (i.e. figuring out which specific cause areas we should prioritise in a certain country) and less the foundational or philosophical kind (e.g. figuring out the value of the far future).

Finally, LPR is a broad catch-all term for many specific research activities that one can do in your local context. Here is a non-exhaustive list of examples:

• Local cause area prioritisation[1]
• Local problem profile research
• High impact local career pathway research
• Risk assessment
• Giving and philanthropy landscape research
• Charity evaluation
• Public policy research
• Analysis of the convergence and divergence between EA and local culture/religion/ethics

3. How does LPR fit into community building?

There are generally three categories of work under community building:

• “Community focused activities”. This is the most visible kind of community building work. It’s about introducing EA to people in a high-fidelity manner, supporting aspiring EAs in their journey to maximise impact, and ensuring their local EA community remains welcoming, inclusive, and collaborative. Examples of activities associated with “community focused activities” are: introductory meetups, EA concepts workshops, 1-1s, retreats, other themed events, etc.
• Administration. If you have grown to a certain stage, you might start spending more time on more administrative related tasks, such as accounting, legal, and human resources. They don’t directly affect your community building goals, but they can have a multiplier effect on your work.
• Strategy management. A significant part of work that comes with maximising a certain outcome, whether it be profits or impact, is strategic research and planning. It’s about figuring out the right goals and the best plan of action to achieve those goals, considering the internal capabilities and resources, as well as the external environment. Typical tools used to help formulate strategy are: SWOT analysis, Porter five forces analysis, portfolio theory, theory of change analysis, etc.

And it is in the “strategy management” category that LPR sits in. Instead of viewing EA community building work as 100% “community focused activities”, I suspect if your EA group has the right fit (which I will talk more about in the next section), you might want to consider allocating 10%-50% of your time in “strategy management”, including LPR. As an example, I’ve spent approximately 25% of my time this year on “strategy management” in EA Singapore. [2]

For a more concrete view of where LPR fits into community building, take a look at the following graph.

It also might be the case you have a different breakdown of community building work, but I think LPR work can still be integrated into community building regardless of the difference.

4. Who should be doing LPR?

LPR is new, experimental, and difficult so it’s likely to be a better fit for groups who are in a more mature community building stage. It might be helpful for you to think about whether to do LPR work using the following criteria:

• Capacity. You and your team have at least a combined total of 6 hours a week of free time for at least 6 months, not including time spent on “community focused activities”. This is very subjective, but I suspect for a good enough quality of research output, you would need at least a total of 100 hours of research.
• EA knowledge. You and your team have already engaged most of the core readings from a typical introductory fellowship. This is important to create alignment between team members.
• Comparative advantage. You and your team have at least completed one 6-month research project before (this can be your undergraduate research). If you don’t have a track record of conducting research, it might be quite difficult for you to do LPR.
• Scope. You and your team can coordinate and operate at a national capacity, and not just within a city or a university. I think conducting LPR for an entire country is probably more helpful than just conducting LPR for a city or town.
• Risks. In certain countries, any kind of civil society work might be highly risky. So you and your team might want to conduct a risk assessment first before conducting any LPR work or publishing any LPR work.
• Opportunity costs. You and your team will need to pick a configuration of LPR, administration, and community focused activities that has the highest total net value. It might be the case that doing no LPR work and more “community focused activities” is more impactful on the margin.

To emphasise, it’s not necessary that all EA community builders conduct LPR or spend a lot of time thinking about strategy. Not everything you do has to be at a high standard, and it’s definitely okay to work on community building casually (as long as you follow EA Hub’s recommendations). Furthermore, if you’re already capacity stretched, perhaps focus more on your personal high impact career plans or giving plans instead of community building.

5. Why consider integrating LPR into your community building work?

There are three main reasons why LPR can be valuable to your community building:

• The most important reason is that it helps provide career insights for community members to figure out their high impact career pathways in their local context.
• It provides external insights for your community building strategy.
• It helps make outreach more effective.

Here’s a chart to help you visualise this better. I will explain more in the sections below.

5.1. You gain local career insights

As more and more of your community members are seriously considering changing to a higher impact career pathway, you might find that a key bottleneck in your community building work is the lack of information regarding local high impact career pathways. If you’re in a non-EA-hub, 80,000 Hours’s job board is likely not useful, so doing LPR can contribute to finding potentially impactful local career pathways. Besides the example of Singapore’s technical AI alignment research career guide, LPR work has shown that career pathways in AI policy and developing alternative proteins are also likely to be impactful in Singapore. We’re still testing out whether these career guides will be helpful, but I would assign a 70% confidence level that such career insights will be helpful in supporting community members transition to a high impact career pathway.

As a side note, I think LPR can also be helpful for your own high impact career pathway too, and it’s not just for your community members!

5.2. You gain external insights

Insights from LPR can complement usual strategy frameworks that are used in many organisations. For example, when conducting research for your SWOT analysis, insights from local cause prioritisation research or local problem profile research can improve on the opportunities and threats section (or the O and T section) of the SWOT analysis.

Moreover, LPR work can also help you understand the risk landscape of your country better, preserving option value for your community. One example that might be helpful is the LPR work that will be done by EA Malaysia on farm animal advocacy in Malaysia. With the help of Lynn Tan who has gone through Charity Entrepreneurship’s incubation programme, it will be interesting to see what kind of risks that will come up from our analysis using the weighted factor model. Hopefully that will help us not make any bad irreversible decisions in this space, or cause any accidental harm.[3]

5.3. Your outreach is more effective

There are three ways how LPR can improve outreach: it helps make outreach more targeted, it helps improve your group’s credibility, and it attracts more EA-aligned people.

Firsty, LPR helps make your outreach more targeted. For example, knowing that technical AI alignment research is likely to be impactful in Singapore, I’m figuring out ways to make outreach more targeted towards those who are likely to be in this field in the future. This could mean approaching university computer science departments if I could advertise an introductory EA fellowship to their students.

Secondly, LPR helps improve your EA group’s credibility[4], and credibility is attractive. Owen Cotton-Barratt argues:

“... if the ratio of direct work to movement growth activities is too low, this could hinder the growth of the movement, either because it gives a natural line of criticism, or because it is harder to tell a clear story of what the movement is about. In such cases direct work may become the most effective form of movement growth at the margin – albeit with a slant towards direct work that can be communicated clearly or sets a good example.”

Furthermore, a community research report done by Jah Ying Chung found that one of the most positive moments for EAs, that inspired them to get more deeply involved in EA-related work, were the moments where they meet “really smart EAs doing ambitious work”. We can infer that meeting incredibly thoughtful EAs doing amazing work (which makes them look highly credible) motivates people to pursue EA-type goals further.

I argue that LPR can be analogous to direct work, and hence be a source of credibility. I’d imagine if someone finds 80,000 Hours’s research in some problem profiles and career guides helpful, then they’re likely to view good quality LPR work as somewhat helpful too.

Lastly, LPR work might also be great at enhancing the best of EA’s values (or culture)⁠—the pursuit of maximising impact by conducting some form of EA-type research. That means LPR can attract people who have strong alignment with such values.

As a community builder of a new community, you will be the most obvious “brand” for the community. The work you do, the words you say, and the actions you take subtly signal to others what kind of community you’re building. In Start with Who, the author recounts a failed community building experience—the goal was to attract and build a community of Xs, but because the founders were all Ys, the community eventually became filled with Ys instead of Xs. The author thinks that “founders effect” (or homophily) is at play here.

“People often join a community because they resonate with its vision. But they will stay (or not) because of its people. I have seen many communities fail that had a beautiful Why but didn’t tend to the Who.”

Moreover, actual or perceived similarities play a large part in attracting people and continuing to attract them, whether it be the founder’s purpose, behaviour, interests, or gender.

As a side note, I suspect you will personally feel more satisfied with the quality of the community when they are quite value-aligned with you. This may not seem important as compared to the other reasons above, but when you look at your community, you want to be able to say, “these are my people!”. I know I felt that way when I first encountered the EA movement.

6. Risks

6.1. There is a risk that LPR work is less valuable than other options

I think a majority of EAs, including me, generally have a better comparative advantage in “community focused activities” than in conducting research in a relatively new field. I can definitely sympathise with those who think LPR is a no-go for them, and there may be better options than just LPR.

My past experience trying to juggle LPR among my other responsibilities have definitely been difficult. I’ve spent approximately 270 hours on it already this year (see Appendix A for a brief history of my experience with LPR and a fuller breakdown in hours of the outputs I’ve produced). I’ve had to reduce the quality of research, in favour of publishing them out within their deadlines. It’s definitely not easy. I wish I have more time to not only work on these very important research projects, but build the necessary career capital (mostly research skills) to better work on them.

To mitigate the risk of doing LPR work when it’s not valuable, definitely use the criteria in section 4 to check whether your group is a good fit for LPR work. I suspect more later-stage communities will benefit more from LPR than communities that are still in the earlier stages. There might also be other criteria or reasons I might have missed which could be helpful in determining the value of LPR work versus other options.

6.2. There is a risk that poor quality LPR may cause accidental harm

It might be the case that bad information from LPR may make people overconfident about certain choices which can cause accidental harm. For example, someone who may be a bad fit for policy work entering the policy research space after seeing some poor quality LPR work on it.

There are a few ways to mitigate this:

• First and most importantly, you should be sure that you and your team are a good fit to work on LPR. You can review section 4 above again on this.
• Secondly, get feedback from other EAs about your LPR work.
• Thirdly, always emphasise your biggest uncertainty in your published LPR outputs and to include cautionary messages to readers.
• Lastly, it might be helpful to include workshops and discussions on accidental harm in your community focused activities too.

6.3. There is a risk of being unwelcoming

It’s important that EA remains a welcoming and inclusive community. However, due to limited resources, there is a tendency to focus on a few individuals than many. A large part of conducting LPR is figuring out what are the most pressing priorities in your local context, and it comes with figuring out what specific roles and skills are best suited to work on these priorities. With that, there is a risk that you might be unwelcoming for some people that may not fill these roles well.

I think this risk is still manageable, as long as you still keep “community focused activities” welcoming and inclusive for all. I think it’s important to have an explicit diversity, equity, and inclusion plan written out too. [5]

7. Conclusion

EA groups that are in their later-stages and in non-EA-hubs should consider conducting “local priorities research” (LPR) together with other community building work. LPR can (a) give local career insights that might benefit your community members’ high impact career plans, (b) give external insights to improve your group’s community building strategies, and (c) make your outreach efforts more effective.

There are a few risks involved with LPR, primarily the risk of doing LPR work when it's not valuable, the risk of producing poor quality LPR work, and the risk of being unwelcoming. However, I think there are effective risk mitigation activities you can do to ensure your community building isn’t hampered by such risks.

8. Implications

If conducting LPR helps improve the effectiveness of community building significantly, the implication is that there should be more resources for EA community builders in non-EA-hubs to conduct LPR. Beside “community focused activities”, I think LPR can be a new Task Y[6] that should be better streamlined and supported.

With more and more EA communities appearing in various parts of the world, we may lose out on very impactful opportunities to grow local communities sustainably (or to start impactful local programmes), if we do not proactively support them. I envision a programme where an aspiring EA can apply for: beginning with a virtual introductory fellowship, followed up by an advanced one, coupled with workshops, resources, and 1-1 support on LPR and strategy management. This may not be the best idea (if you have any, please comment), but I suspect testing it early might be helpful.

There are definitely a lot of uncertainties in my arguments, and I hope to get more feedback before deciding whether I want to continue working on this (this investigation and write-up took me 78 hours). One potential project that could be helpful to develop a more nuanced methodology to formulate community building strategies is to build a collection of EA community building case studies from all around the world.

9. What will change my mind?

• Evidence that LPR work is better suited for other kinds of organisations and not a good fit for community building type of organisations.
• Evidence that there are more valuable strategy frameworks in formulating community building strategy compared to LPR.
• Evidence that credibility is not helpful in attracting people.
• Evidence that the effect of homophily is not significant (or if there is evidence that the effect of heterophily is stronger).
• Evidence that even if you did not do LPR, community building work naturally gets the right people to work on the right problems (or cause areas).
• Evidence that there’s a very strong chance that people will produce very poor quality LPR work, which may be harmful to the community.
• The value of information gained from LPR is worth less than the cost of conducting LPR.

Appendix A

Here’s a non-exhaustive list of things I’ve done for LPR in this past year, and why I think it’s not easy. I’ve spent approximately 270 hours so far on LPR work, and some of the hours spent are not listed below.

• Jul-Dec 2019. Loke Jia Yuan, an EA currently based in Singapore, brought up in a EA Singapore retreat that Singapore might have comparative advantages on AI risks and alternative proteins. This made me think more broadly about doing LPR work and I’ve brought the idea up amongst other EAs in Asia.
• Dec-Feb 2020. 43 hours. I seriously considered doing LPR work in conjunction with my community building work, and wrote a project proposal for EA Singapore.
• Mar 2020. 12 hours. I’ve produced a minimal viable cause prioritisation analysis for Singapore using a mix of the INT framework and Jah Ying’s framework. I didn’t spend much time working on this, as I’m already convinced we have enough evidence to focus on AI risks and farm animal welfare.
• Apr 2020. 20 hours. I’ve produced a minimal viable high impact job board for Singapore.
• Mar-Jul 2020. 106 hours. I’ve dived into technical AI alignment research as a cause area that is potentially impactful in Singapore and produced a career guide.
• Apr-current. 36 hours. I’m currently completing a less comprehensive AI policy career guide for Singapore.
• Jul-current. I’m currently getting a few EA Malaysians to conduct a Malaysian cause prioritisation project using Charity Entrepreneurship’s research framework with the help from Lynn Tan, who just completed Charity Entrepreneurship’s incubation programme.

Footnotes

[1] Not to be confused with LPR which is a lot broader, as this specific research activity is about figuring out which certain cause areas we want to prioritise on.

[2] That is roughly 319 hours of work on “strategy management” over a total of 1229 hours so far for 2020. Out of the 319 hours spent on “strategy management”, I’ve spent around 270 hours on LPR. See Appendix A for a clearer breakdown.

[3] These articles on option value and accidental harm might be helpful if you’re not familiar.

[4] There are two kinds of credibility that are important for EA. The first is credibility in terms of expertise, e.g. having relevant and competitive credentials, being able to solve crucial problems, showing competent work and accomplishments, articulating reasonable and evidence-based insights, and being the go-to resource for a specific issue. The second is credibility in terms of values, especially those related to EA’s guiding principles, e.g. being honest, thoughtful, considerate, open-minded, evidenced-based, and radically altruistic.

[5] I highly recommend reading Kelly Anthis’s write-up on this.

[6] “(1) Task Y is something that can be performed usefully by people who are not currently able to choose their career path entirely based on EA concerns. (2) Task Y is clearly effective, and doesn't become much less effective the more people who are doing it. (3) The positive effects of Task Y are obvious to the person doing the task.” (This excerpt is from this write-up but I think the Teach First or Teach for All analogy makes the most sense.)