Rethink Priorities 2019 Impact and Strategy

by Peter_Hurford 10d2nd Dec 20197 min read1 comment

54


Summary

  • Rethink Priorities is an EA research organization focused on influencing funders and key decision-makers to improve decisions within EA and EA-aligned organizations.

  • Over the past year, we have produced a large amount of research on a number of topics in animal welfare, risks from nuclear war, and EA movement growth. This work includes cost-effectiveness estimates of corporate campaigns, farmed fish welfare, fish stocking, the impact of cat predation, invertebrate sentience, the likelihood and severity of nuclear winter, and a comprehensive evaluation of the 2018 EA Survey.

  • Rethink Priorities currently has a staff of 10 people and 8 full-time equivalents. We spend 80% of our time working on research relevant to farmed and wild animal welfare. In 2019, we spent about $400K.

  • We have been tracking our impact, formally in a survey and informally by consulting key decision-makers and tracking their actions. We can improve our impact by ensuring relevant parties read our work.

  • Over the next two years we plan to try to find new actionable interventions to improve animal welfare, further analyze nuclear risks, use polling to find winning policy changes, study EA movement growth, and much more.

  • We are heavily funding constrained and have the management capacity and hiring talent pool to quickly grow if given more money.

  • If funded we would bring Peter Hurford on full-time, buy more working time from our existing research team, increase our budget for surveys and other projects, invest further in our operations, give raises, and hire more staff to expand the team and work we could do.

  • Currently, our funding gap through the end of 2021 is $1.79M overall. This consists of gaps of $1.27M for animal research, $337k for longtermism research, and $177k for meta / other research respectively. We do accept and track restricted funds by cause area if that is of interest.

  • If you'd like to support our work, you can donate to us as part of Facebook's donation matching on Giving Tuesday or donate directly to us here.

Our Mission

Only a small fraction of potentially high-impact research questions have actually been investigated. Despite this, the EA movement allocates hundreds of millions of dollars and hundreds of thousands of work-hours each year to improving the world.

Our goal is to improve this allocation. We can do this by identifying neglected but potentially high-impact areas, investigating them, then ensuring our research is read and used by key decision-makers to make better decisions that impact the world. We focus on uncovering new interventions within cause areas rather than ranking cause areas against each other because this strategy is more tractable and less neglected.

In doing this research, we strive to emphasize relatively quick feedback loops, seeking tractability in research by looking for questions that we may be able to make meaningful progress on, and tracking the impact of our research on an ongoing basis in order to determine how to maximize our impact.

Throughout 2018 and 2019, Rethink Priorities has been a project under the larger Rethink Charity umbrella. After much discussion and deliberation we've decided that RP will begin the process of spinning off from Rethink Charity to form its own US 501(c)(3). We have only just begun this process and estimate this it will likely take 6-12 more months. Until that process is complete, and RP has legally moved all operations to the new organization, you can continue to donate to RP via RC. If you are planning a donation around the time of our legal separation, please let us know and we will ensure your donation is processed by the correct legal entity.

Ultimately, both RP and RC want RP to be a fully independent organization. While sharing operations staff and legal housing made sense when RP was small (2-3 people), it made less and less sense as RP grew in size and as it became a larger and larger share of RC by employees and budget. In particular, there are practical reasons around the ease of accepting donations that makes this particularly worthwhile for RP and less of a burden on RC operations.

Our Work and Impact So Far

Summary of Work Published in 2019

In 2019, we published 48 pieces of work. We think we've made some quality advancements too—within farmed animal welfare we analyzed whether corporations were likely to meet their animal welfare commitments and the cost-effectiveness of corporate campaigns to improve chicken welfare, significantly contributed to Animal Charity Evaluators's report on farmed fish welfare, and we examined the lives of fish raised to be released into the wild and rodents killed and fed to pet snakes.

In wild animal welfare, we explored life history classification in ecology, quantitatively assessed the lives of herbivorous insects, and considered the impacts of cat predation.

We've also published the results of our examination of the scientific literature for invertebrate sentience, including a three part analysis of features potentially relevant to assessing sentience, a cause profile of invertebrate welfare, and the table of the resulting data compiling all our work on 18 different taxa and 53 different features. We followed this up by reviewing the welfare of managed honey bees and considering next steps for invertebrate welfare for research, possible interventions, and in assessing attitudes and possibilities in this space.

We've also done some work on analyzing nuclear weapons as a possible extinction risk, under the supervision of Carl Shulman. We've posted the beginning of our work, including which nuclear wars seem worst in expectation, and four posts analyzing a US-Russia nuclear war: whether nuclear forces would survive a nuclear first strike, expected deaths as a direct result of an exchange, how bad nuclear winter would be, and the likelihood of such a nuclear exchange.

Our evaluation of the 2018 EA Survey included posts on cause selection, group membership, geographic differences, where people first hear of EA, EA welcomingness, EA movement growth, general retention and GWWC pledge adherence. We also evaluated the cost-effectiveness of RC Forward and the value of a grant to fund Donational. We've completed and assessed the 2019 Local Group Survey and conducted the 2019 EA survey, and will begin publishing results before the year is out.

You can view a complete list of all of our publications on our website.

Our Impact

As a research organization, we think it's particularly important to confirm that our work is important to decision-makers and key players in the areas in which we investigate. As such, we're tracking our impact in multiple ways.

We've conducted a formal impact survey, we informally track our impact via regular contacts with relevant players, and, most inconsequentially, we have done some tracking of page views and engagement with our work on our site and the EA Forum.

Impact Survey

We have published the results of a formal impact survey of our work. That survey found that a quarter of those who responded had read more than half of our work and the vast majority of our work was judged to be useful by at least 50% of those who read and rated it. 57% of respondents stated that our work had changed their beliefs. Our work did not influence any of these key players to make hiring decisions or start new charities. Instead, the majority of changes reported were in 22 unspecified ways (which we would be eager to find out more about in follow-up work), 13 instances of inspiring research in a topic, and influencing 7 donations.

Due to the small sample of 47 respondents, the disproportionate importance of some of these respondents, and the ability to highlight comments from only those who opted to share their responses publicly, the precise results should not be taken too seriously.

Informal Impact Tracking

Informally, we've discussed our work and role with a number of key organizations and decision makers. We've actively sought feedback in general, but also specifically engaged these parties as to what they have and have not found helpful in what we've done, what we can do in the future that would be useful for them, and ask in depth about ways we can improve. These discussions have generally been in line with what we've found in the impact survey.

In addition to these private conversations, Charity Entrepreneurship cited our work on companies keeping their commitments as part of their analysis which made increasing the follow through rate of cage-free pledges one of their top charity ideas (though no one started this charity).

Google Analytics

We've tracked engagement with our posts on our site and the EA forum via Google Analytics. Thanks to CEA, you can view the page views and engagement with our posts on the EA Forum from our start in Jan 2018 until late September 2019[1]. While the relative popularity of some of these posts aligns somewhat with our other impressions, we do not believe this information is particularly important for judging the impact of any of our projects.

Increasing Our Impact Through Distribution

These methods of tracking have led us to believe we think we could do a lot better at making sure all of the relevant parties read our work and improve their decisions. As such, we are ramping up our newsletter to be sure to capture all the work we produce, and to be sure all the key players see our work by directly reaching out to them to ensure they are on our mailing list. We are also engaging in direct outreach to organization leaders and key funders in relevant areas to be sure they see our work.

Our Current Plans

Initiative Updates

We believe we are well positioned to produce a large amount of impactful, high-quality research over the coming year. Key initiatives we may take on in 2020 potentially include:

  • Finding actionable, neglected interventions to improve animal welfare

    • Estimating the number of vertebrates kept in captivity by humans
    • Assessing the value of legal and political interventions (including forthcoming work on ballot measures) to improve animal welfare via analysis, modeling and polling
    • Conducting case studies of species welfare to understand wild animal lives
    • Investigating possible farmed invertebrate interventions
    • We think we could potentially find millions of dollars of funding opportunities not currently known to grantmakers that are at least as impactful as the current grantees, thus unlocking massive new opportunities for impact and allowing grantmakers to scale their giving without sacrificing impact
  • Using polling to find winning policy changes

    • Running multiple high-quality surveys to better understand people's attitudes toward animals given a partnership with a data analytics firm, and our existing work using the Prolific online survey platform
    • Systematically identifying promising policies, evaluating their expected impact if enacted, forecast their likelihood of passing, and recommend the optimal policies in the optimal electoral locations by expected value
    • Studying attitudes towards wild animals, with surveys and focus groups
    • We think this work could identify the next best campaigns to run after the existing cage-free campaigns and inform a higher impact political strategy for animals
  • Evaluating risks to the long term future of humanity

    • Further evaluating nuclear weapons arms control and potentially exploring the lessons learned for AI policy
    • Analyzing the likelihood of civilization recovering from a population collapse
    • We've been developing strong relationships with the Future of Humanity Institute to understand how we can contribute to longtermist work
  • Analyzing EA movement growth

    • Continuing to analyze EA Survey data for 2019 to identify movement progress
    • This year, we have created a strong partnership with CEA and 80,000 Hours, including commissioned research for CEA, to ensure our results are useful and impactful
  • Assessing mental health and subjective well-being

    • Outlining research paths for incorporating subjective well-being into global health metrics
    • Considering the steps necessary to apply subjective-well being into DALYs and QALYs

Funding Status

We are heavily funding constrained and have the management capacity and hiring pool to quickly grow if given more money.

Our budget for 2020-2021 is between $1,174,000 and $2,384,000 over two years. If funded at the minimum level, we would cover our existing operations through 2021, pay for necessary operational support and buy 10 more hours a week working time from our existing research team. At the higher level, we'd be excited to provide a salary to allow Peter Hurford to quit his day job as a data scientist and co-lead the organization full-time, increase our budget for surveys and other projects, invest further in our ops staffing to streamline the organization, give raises, and to hire more staff to expand the research team and work we could do further. We think there are many highly talented researchers we could hire and neglected high impact research questions that we would be able to pursue if only we had more funding.

We have already raised $429k towards the minimum for 2020, and $84k per year towards the high end of our budget for 2020 and 2021. Thus our existing funding gap, the amount we are hoping to raise, is smaller than our budget through 2020 and 2021. That funding gap is as follows:

If you'd like to support our work, you can donate to us as part of Facebook's donation matching on Giving Tuesday or donate directly to us here.

Credits

This essay is a project of Rethink Priorities. It was written by Marcus A. Davis and Peter Hurford. Thanks to Daniela Waldhorn, Jason Schukraft, Saulius Simcikas, and David Moss for helpful comments. If you like our work, please consider subscribing to our newsletter. You can see all our work to date here.


  1. This data may be partially incomplete if not every version of each url was captured. Posts with multiple URLs have been manually combined to create total page view numbers. Additionally, we have not fully screened out visits from our own employees and properly doing so may be prohibitively difficult. ↩︎

54