OR: Examing the promise of the development of impact-oriented individuals

Epistemic status: This post incorporates several sources of data and considerations including needs, skills, and bottlenecks reported in existing surveys within the EA community^[1] (~ 1000 respondents), interviews we conducted, relevant reports, scientific literature (ranging from cross-sectional studies to systematic reviews and RCTs), and own models and experiences as practitioners (primarily coaches). We mainly give quantitative point estimates (e.g., our cost-effectiveness estimates) even though confidence intervals would better reflect the underlying distribution and uncertainty. We hope this will be seen as one line of research into this area (that has many flaws) and would be keen to learn about other perspectives.

Summary

This post explores the promise of the Development of Impact-Oriented Individuals (DIOI). We present our research on the extent to which increasing the well-being, character, and productivity of high-impact individuals through interventions such as meditation courses, therapy, and coaching might be a particularly high-impact use of resources. We define DIOI as i) Cost-effective and truth-seeking personal and professional development of ii) impact-oriented individuals.

What we did

To approximate the answer to this we:

Created a model categorizing development areas for impact-oriented individuals based on surveys, interviews, research, and expert judgment. This led to the identification of five categories: Emotional Regulation and Mastery, Interpersonal Skills, Productivity, Professional Development, and Physical and mental needs. Each category has several areas of development, such as motivation, management, and depression.
Estimated the impact on productivity, the prevalence, the improvement potential, and the cost of promising interventions for the most relevant areas of development. The average productivity impact was 22%, and the average prevalence was 33%. We found that the average cost of effective interventions (largely identified via systematic reviews and RCTs) was ~$100/week investment over 21 weeks, and the average improvement potential was 39%.
Used these estimates to calculate the cost of an added FTE (per year) for an organization and for an individual. The expected cost per added FTE ranged from $25,000 to $75,000.
Estimated the impact-adjusted dollar value for every dollar invested in the development of impact-oriented individuals. The estimates ranged from $1.2 to $3.7 for every dollar invested. However, while there are many cost-effective interventions with a better return on investment than $3.7 per every dollar spent, this makes it plausible that this could be a cost-effective way of spending altruistic money within the domain of field-building.
Considered counterarguments to investing in DIOI. We think the strongest counterargument is that other cause areas and interventions are more promising and that resources are best spent there.

We expect there to be many more promising ways of examining the case for investing in interventions that seek to develop impact-oriented people in the broadest sense. These would include examining the importance of habits and strategies used by altruistic peak performers, estimating the impact of other interventions (e.g., leadership training), and estimating the impact of qualities such as character (e.g., virtues like integrity or ambition) development. Therefore, we don’t want to present this as “the case” for activities in this area and hope that readers won’t see this as conclusive findings.

What we recommend

Overall, given all the ways altruistic actors can allocate scarce resources, we don’t think this is likely to be amongst the top ways of doing so. That said, we think that some amount of altruistic resources should be allocated to contribute to DIOI as it seems to be able to further important work in top cause areas (by increasing the number of FTEs). Based on this we make the following recommendations to different stakeholders:

Leaders and HR professionals at organizations

Our model suggests that organizations and individuals could gain an FTE at the price of ~ $25,000-$75,000 corresponding to a $1.2-$3.7 benefit per dollar invested in well-tailored development of their employees. This could include things such as leadership/management training, coaching, therapy, and mindfulness-based programs. This could be done by providing a development (or mental health) budget and supporting employees in making good use of the budget.

Organizations that are low on resources can look for cheaper alternatives such as peer-coaching, peer-guided mental health programs, and mastermind groups combined with cheap and effective digital interventions.

Individuals who think they might benefit from these services

Our models suggest that a well-tailored intervention lasting 21 weeks with an average total cost of ~ $2000/person could lead to an average increase in an individual’s yearly output by 9%. This roughly corresponds to a $3.7 benefit per dollar invested in well-tailored development. We expect this to be particularly important for people working individually like individual AI safety researchers. This could include things such as (productivity) coaching, therapy, and mindfulness-based programs.

For those who are financially constrained, we expect significant benefits to be obtained via cheaper alternatives such as peer-coaching, peer-guided mental health programs, and mastermind groups combined with cheap and effective digital interventions.

Practitioners and professionals working in this area

If you’re actively working with the development of impact-oriented individuals, you can consider familiarizing yourself with the five-factor model (see below) that categorizes development areas for impact-oriented individuals based on surveys, interviews, scientific literature, and expert judgment.

The model includes five categories: Emotional Regulation and Mastery, Interpersonal Excellence, Productivity, Professional Development, and Physical and mental needs.

We also created a more elaborate spreadsheet that includes considerations such as what might indicate that an individual would benefit from developing an area (e.g., stress) and how to measure it (e.g., the Perceived Stress Scale) (see 1.1 Area of improvement estimates (elaborate)). Most importantly, however, it seems that finding the right people and offering highly tailored solutions are key.

For those considering entering this area, we would recommend that you consider other options unless you think you have a strong personal fit for this. However, if you have a strong personal fit, this may be a top option conditional on you finding promising interventions and impact-oriented clients.

During our research, we also identified some ideas we’d like to see:

A high-quality recommender platform that can recommend promising services tailored to individual users. This would have some overlap with services like the EA mental health navigator overview of providers, Upgradable.org, and Clearer Thinking.
An eco-system of providers and services aiming to be the main provider to people pursuing high-impact careers.
High-quality content written on the topics of personal development, mental health, and similar as it relates to doing the most good.

Introduction

This post explores the promise of the Development of Impact-Oriented Individuals. The main motivation was for Sebastian to examine something he felt could have the potential to be a cause area or simply just a promising way of bringing about more good in the world. So, he decided to research this further and contracted Amine Challouf for part of the process. “We” will be used throughout the post but the majority of the post was written by Sebastian (Amine hasn’t read all parts of the post and may not agree with some parts) and all mistakes are Sebastian’s.

We present our research on the extent to which increasing the well-being and productivity of high-impact individuals through interventions such as meditation, therapy, and coaching might be a particularly high-impact use of resources. We hope this post will be valuable for the following audiences:

Grant-makers who find it hard to reason clearly about funding things in this space.
Leaders and HR leads at organizations who find it hard to evaluate the value of personal and professional development, or who see the value, but don’t know what to do about it.
Individuals who think they might benefit from services in this area find it hard to evaluate the cost-effectiveness of related initiatives—especially because it feels too fluffy and can trade off against work.
Professionals (e.g., coaches) working in this area who would like to inform their activities and have a potential way of thinking about the cost-effectiveness of related initiatives.

What is the “development of impact-oriented individuals”?

In order to clearly and rigorously explore the promise of the development of impact-oriented individuals, we need to define what’s within and outside of scope. We’ll do so by giving a brief definition, comparing that definition to a related set of areas, and giving concrete (yet simplified) case studies of what we’d consider within and outside of scope.

We define the development of impact-oriented individuals (DIOI) as:

i) Cost-effective and truth-seeking personal and professional development of ii) impact-oriented individuals.

By truth-seeking, we mean approaching this activity with epistemic humility and a Scout Mindset. By personal and professional development, we mean interventions aimed at improving things such as mental health, (eudaimonic) well-being, character, productivity, and professional skills. By impact-oriented individuals, we mean people who are taking significant action based on impartial altruism (attempting to give equal weight to their beneficiaries regardless of when and where they exist) and truth-seeking. In principle, this is community-agnostic. However, we choose to focus on the Effective Altruism community, so the definition and data (e.g., surveys) are heavily grounded in that community.

Compared to related areas

To further delineate DIOI, we’ll look at some similarities and differences between it and related areas.

Table: Development of impact-oriented individuals (DIOI) compared to related areas

Table: A simple diagram showing the relationship between the development of impact-oriented individuals and more established areas such as mental health. You can find an interactive version of this table here.

Examples of case studies

To further clarify what we mean by this area, we’ll give some simplified case studies that are within and outside of scope.

Within scope

An independent AI safety researcher who rates her overall life satisfaction as ⁶⁄₁₀ but is struggling with procrastination and existential dread as she’s worried about the consequences of AI while lacking good management and knowledge of best practices around productivity.
A senior executive at an ambitious charity who’s altruistically committed and well-performing but is sorely overworked and feels on the verge of burnout. This leads to significantly decreased ability to do important work and ultimately dropping out.
A moderately depressed 20-year-old student seeking to figure out what to do with her career to make a difference in the world.
A researcher who’s making a career decision and wants to get emotional support and hold himself to high standards in the process of doing so.
Someone involved in the EA community who’s struggling to get a job and has put too much of her identity into getting a particular set of jobs and is becoming increasingly miserable. See a courageous actual example here.

Outside of scope

Severe depression among a 31-year-old traumatized refugee in Sweden. Why is this case outside of scope? While bringing about flourishing in this person clearly has the highest intrinsic value to address, it’s outside of scope as it’s too severe and he’s unlikely to become someone who can significantly affect the most pressing problems and valuable opportunities of our time. This may sound outrageous, but it’s largely because I think we lead in such a weird moral time and relatively few people have an outsized impact—if this weren’t the case, this person would likely receive more of altruistic attention.
A 45-year-old man who’s feeling unhappy and wants to chase conventional success in the form of wealth and use that to retire early and lead a reclusive life.

Why this is important

Now that we have established some clarity on what we’re talking about, let’s attempt to examine how important this area is. We’ll do so by creating a model that categorizes the development areas (e.g., depression) for impact-oriented individuals based on surveys, interviews, research, and expert judgment. We can then use this to generate relevant estimates that can be used to create some cost-effectiveness estimates.

The five-factor model of the development of impact-oriented individuals

To think more clearly about how to develop impact-oriented individuals, we contracted Amine Challouf to help us develop a model. The model seeks to create meaningful categories (e.g., emotional regulation and self-mastery) within this area. We believe that these categories are helpful to i) produce estimates that can enable us to understand the impact potential and ii) to use in a more applied way as part of providing interventions within this area.

More concretely, we formulated these categories on the basis of several sources of data and considerations including needs, skills, and bottlenecks reported in existing surveys within the EA community (~ 1000 respondents), interviews we conducted, relevant reports, scientific literature (ranging from cross-sectional studies to systematic reviews and RCTs), and own models and experiences as practitioners (coaches).

That said, this categorization is tentative and should be seen more as a starting point for further research than a well-consolidated set of findings. For instance, there were many things we didn’t include that are clearly valuable in practice and would be unfortunate not to include when needed. These include qualities such as ambition, collaborative spirit, general professional and personal experience (i.e., having a significant amount of experience from various walks of life), and conscientiousness/work ethics.

Table: Description of the five-factor model of development of impact-oriented individuals

Table: Each category has a general description of the area, the top areas of development associated with the category (mostly the items with the highest prevalence from various surveys), and examples of promising interventions related to the category.

Now that we have outlined some relevant high-level categories and areas of development, we can look for relevant estimates that can inform how cost-effective it might be to make progress in these areas. The most relevant estimates to establish are:

Estimates related to the effectiveness of the area of development:

Quantitative estimates of productivity effects associated with the area of improvement. This is mainly based on academic literature (primarily correlational) around the relationship between an area (e.g., anxiety) and productivity. Note these likely represent the upper bound as they’re often estimates of more significant versions of the different areas of development (e.g., generalized anxiety) than the average survey respondent is experiencing.
Prevalence of the area of improvement amongst impact-oriented individuals (in this case people within the EA community). This is primarily assessed based on self-reported survey data and, occasionally, anecdotal evidence. This is likely to be an overestimate due to selection effects amongst people who responded. In some cases, we’ve discounted the estimates by 50% to make up for this.
Potential improvements one can expect by addressing the area of development with promising interventions or services. We identified effective interventions (see Intervention identified) via systematic reviews and randomized controlled trials. Then, we found what proportion of the problem was solved via the intervention. For simplicity, we’ll assume that the potential improvements will persist for one year and then revert completely.

Estimates related to the cost of addressing the area of development:

Cost ($) per week of intervention/training. This is assessed based on our existing knowledge and research of promising services.
Average duration of intervention or training (weeks). This is based on our existing knowledge, promising services, and academic literature.

Instead of finding these estimates for all relevant areas of development (as listed in the survey), we identify five priority areas of development (one from each of the categories). You can see them here:

Table: Estimates related to the cost and effectiveness of priority areas of development

Table: Please refer to the page above for a detailed explanation of the different estimates. For a more detailed table (including links to the relevant references for each estimate) and other areas, please refer to this spreadsheet.

Overall, we can see that, amongst the priority areas of development, there’s a notable prevalence (33%) and impact on productivity (22%), on average. Our model suggests that, with a significant investment of ~ $100/week over 21 weeks, there’s a potential for improving the area of development by 39%, on average. While the averages are somewhat strange estimates, due to the variance, we use them to get a sense of the general cost-effectiveness of the “portfolio” of interventions. Ideally, we would use confidence intervals, instead of point estimates, to reflect the underlying distribution and uncertainty better. However, that’s beyond the scope of this post and all but a few of these estimates are based on relevant resources (primarily scientific studies) so we feel reasonably confident that these estimates are directionally correct, though there will of course be significant case-by-case differences.

For a case study of how some of these areas of development might affect individuals’ productivity over the long run, please see Howie Lempel’s account of his struggle with depression and anxiety.

How many additional FTEs can we expect to gain by investing in the development of organizations and individuals? (Cost-effectiveness)

We can use the averages of estimates of the areas of development above to approximate the costs and effectiveness of investing in addressing these areas of development. There are many relevant effectiveness measures, but we think the most reasonable is full-time equivalent (FTE) of work done in a priority path per year. We can estimate FTEs based on the averages in the following way:

Quantitative Estimated Impact on Productivity*Prevalence*Potential improvements

We can do this for the following groups of impact-oriented individuals:

An impact-oriented organization with around 10 people who are representative of the community as a whole. Here, we won’t use a discounting factor as we think it’s reasonable to expect that the organization would self-select to get benefits like this and thus not have well-tailored ways of developing itself.
Impact-oriented individuals who would benefit from these areas of development and aren’t already making use of well-tailored interventions. An example of this is an individual AI safety researcher.

Table: Benefits (full-time equivalent) per year of investing in DIOI at an organizational and individual level

	Expected cost per FTE (for one year)	Expected FTE based on the average investment	Number of people developed	Prevalence (average)	Discounting factor	Expected total cost
Impact-oriented organization (10 staff members)	$76,127	0.29	10.0	33.20%	NA	$21,871
Impact-oriented individual (relevant individual)	$25,274	0.09	1.0	NA	NA	$2,187

Table: Please see this spreadsheet for a detailed breakdown of the estimates.

Thus, we estimate the expected cost per FTE (per year) is ~$75,000 and ~ $25,000, respectively.

Put differently, these estimates suggest that a well-tailored intervention lasting 21 weeks with an average total cost of ~ $2000/person could lead to an increase in FTE of:

0.3 (for an “average” organization).
0.1 (for an individual).

We need to convert FTE to a dollar value to interpret how good of a deal this is.

What’s the return on investment for investing in this area? (Cost-benefit)

How much is an FTE (year) worth?

This will depend on the specifics of the role, but one way of assessing this is to approximate the willingness to pay for an employee amongst relevant organizations. Based on looking at the salaries of different job openings, we can get a first estimate for different levels of seniority:

Junior (<3 years of experience): ~ $50,000/year
Senior (3-10 years): ~ $75,000/year
Executive: ~ $100,000/year
Average: ~ 75,000/year

However, these will be underestimates of the willingness to pay, as the organizations pay more than the salaries (such as operations, employer tax, and benefits) so we’ll add 25% to these estimates:

Junior (<3 years of experience): ~ $63,000/year
Senior (3-10 years): ~ $90,000/year
Executive: ~ $125,000/year
Average: ~ $93,000/year

With these estimates, we can now estimate the impact-adjusted dollar value for every dollar invested in a well-tailored intervention:

Table: Cost-benefit estimates of investing in DIOI at an organizational and individual level

	Cost-benefit (impact-adjusted $ value per $ invested)	Expected benefits per year	Expected net benefits per year (impact-adjusted dollar value)	Value of FTE (per year)	Expected total cost
Impact-oriented organization (10 staff members)	$1.2	$26,718	$4,853	$93,000	$21,865
Impact-oriented individual	$3.7	$8,048	$5,861	$93,000	$2,187

Table: Please see this spreadsheet for a breakdown of the estimates.

For an “average” organization, the cost-benefit seems to be $1.2 for every dollar invested. One way of improving the cost-benefit would be to offer well-tailored services specifically to the proportion of people who are likely to benefit the most from it and are the most senior. This seems particularly feasible given that some organizations offer their employees mental health and development budgets.

For a selected individual, the cost-benefit looks much better at $3.7 for every dollar invested. While there are many cost-effective interventions with a better return on investment than $3.7 per every dollar spent, this makes it plausible that this could be a cost-effective way of spending altruistic money within the domain of field-building.

Why this isn’t important (negative effects)

Now that we have explored some reasons for the importance of DIOI, let’s consider some counterarguments.

This risks being self-centered and at odds with doing the most good

A counterargument is the concern that prioritizing personal and professional development may lead to self-centeredness and is fundamentally at odds with doing the most good as that involves putting others first. This could be particularly problematic if DIOI were viewed primarily as intrinsically valuable rather than instrumentally valuable (means to improve the lives of others). In some sense, it seems wrong to take care of oneself when there’s a lot of suffering right now and too high probability of risks such as Existential risks. We do see a failure mode of “overdoing” investing in oneself. However, as the survey results indicate, many impact-oriented individuals have areas of development (including not being depressed or being highly inexperienced managers) that are clearly impeding their ability to do good. Relatedly, increasing well-being and developing one’s character are likely to be part of what ultimately matters (including if it’s done for oneself). Overall, we see this as a counterargument and something to watch out for but, on the margin, we don’t think this is a strong counterargument to investing in oneself.

Other cause areas and strategies are more promising

Given the number of pressing cause areas and promising interventions and the big difference in impact between the best and the worst, one should have a very low prior that any one area or intervention is amongst the most promising. We think this is the strongest counterargument and we don’t think DIOI is amongst the ten most important cause areas and think people should think twice before pursuing this professionally and rather consider direct work (e.g., AI governance). However, as suggested by the cost-effectiveness estimates, DIOI might be a good way of directing valuable resources (i.e., FTEs) toward important cause areas.

Conclusion and recommendations

We’ve explored the case for considering the development of impact-oriented individuals as an impactful way of improving the world.

We define DIOI as i) Cost-effective and truth-seeking personal and professional development of ii) impact-oriented individuals.

To understand the value of pursuing interventions in this area, we

Created a model categorizing development areas for impact-oriented individuals based on surveys, interviews, research, and expert judgment. This led to the identification of five categories: Emotional Regulation and Mastery, Interpersonal Skills, Productivity, Professional Development, and Physical and mental needs. Each category has several areas of development, such as motivation, management, and depression.
Estimated the impact on productivity, the prevalence, the improvement potential, and cost of promising interventions for the most relevant areas of development. The average productivity impact was 22%, and the average prevalence was 33%. We found that the average cost of effective interventions (largely identified via systematic reviews and RCTs) was ~$100/week investment over 21 weeks, and the average improvement potential was 39%.
Used these estimates to calculate the cost of an added FTE (per year) for an organization and for an individual. The expected cost per added FTE ranged from $25,000 to $75,000.
Estimated the impact-adjusted dollar value for every dollar invested in the development of impact-oriented individuals. The estimates ranged from $1.2 to $3.7 for every dollar invested. However, while there are many cost-effective interventions with a better return on investment than $3.7 per every dollar spent, this makes it plausible that this could be a cost-effective way of spending altruistic money within the domain of field-building.
Considered counterarguments to investing in DIOI. We think the strongest counterargument is that other cause areas and interventions are more promising and that resources are best spent there.

Recommendations

Overall, given all the ways altruistic actors can allocate scarce resources, we don’t think this is likely to be amongst the top ways of doing so. That said, we think that some amount of altruistic resources should be allocated to contribute to DIOI as it can further important work in top cause areas (by increasing the number of FTEs).

Based on this we make the following recommendations to different stakeholders:

Leaders and HR professionals at organizations

Individuals who think they might benefit from these services

For those who are financially constrained, we expect that significant benefits can be obtained via cheaper alternatives such as peer-coaching, peer-guided mental health programs, and mastermind groups combined with cheap and effective digital interventions.

Practitioners and professionals working in this area

If you’re actively working with the development of impact-oriented individuals, you can consider familiarizing yourself with the five-factor model categorizing development areas for impact-oriented individuals based on surveys, interviews, research, and expert judgment.

The model included five categories: Emotional Regulation and Mastery, Interpersonal Excellence, Productivity, Professional Development, and Physical and mental needs.

We also created a more elaborate spreadsheet which includes considerations such as what might indicate that an individual would benefit from developing an area (e.g., stress) and how to measure it (e.g., the Perceived Stress Scale) (see 1.1 Area of improvement estimates (elaborate)). Most importantly, however, it seems that finding the right people and offering highly tailored solutions are key.

During our research, we identified some ideas we’d like to see:

A high-quality recommender platform that can recommend promising services tailored to individual users. This would have some overlap with services like the EA mental health navigator overview of providers, Upgradable.org, and Clearer Thinking.
An eco-system of providers and services aiming to be the main provider to people pursuing high-impact careers.
High-quality content written on the topics of personal development, mental health, and similar as it relates to doing the most good.

If you’re interested in making something happen within this area, feel free to reach out to Sebastian.

Acknowledgment

Several people have contributed to this via informal conversations and feedback on the post (though they do not agree with all claims and methods). Thanks to Magnus Vinding, Jan Brauner, Henning Bartsch, Paul Rohde, Anna Weldon, Tee Barnett, Gidon Kadosh, James Norris, and Vilhelm Skoglund for valuable feedback and conversations around this topic more generally. Special thanks to Lowe Lundin and Toby Tremlett for their valuable and critical feedback.

As mentioned above, this post was mainly written by Sebastian Schmidt with substantial contributions from Amine Challouf (who hasn’t read all parts of the post and may not agree with all parts) - all mistakes are Sebastian’s.

Appendix

Summary of the data around needs, skills, and bottlenecks use

The 2021 EA mental Health Survey (N=326) outlines the following as the top 10 mental health problems in EA:

The 2023 EA mental health survey (n=167), asked the question, which of these do you struggle with or would like to improve the most?

The following were listed by >20% of the participants:

Mood
Sleep
Exhaustion/fatigue
Concentration/attention
Self-esteem
Self-compassion
Control over behavior
Energy levels
Motivation
Work-life balance
Fear of rejection and judgment
Perfectionism
General anxiety
Procrastination
Burn out

In the 2020 Effective Altruism survey (535 respondents) respondents listed these as the top 10 most frequent bottlenecks:

Finances
Network
Productivity
Uncertainty
Mental health
Job opportunities
Direction
Time
Career change
Education

Summary of top needs identified from interviews with Existential risk researchers

Developing a clear understanding of risks and a clear understanding of the impact they’re having
Training in skills, and having access to mechanisms that ensure they do not amplify risks
Optimizing both work and rest time sustainably
Addressing x-risk anxiety and developing resilience
Combatting isolation, especially for independent x-risk researchers (and the difficulty in sharing their ideas with their “outgroup”)
Building the capacity to deal with uncertainty, ambiguity, and confusion
Validation of their work & sense of belonging
Easily meet physical and emotional needs
Training in meditation, emotional regulation & cognitive skills
Training in epistemics
Maintaining full focus on tasks
Persuasion skills, clear communication, and argumentation
Fundraising
Bottlenecks on ambition
Being able to work independently

Charity Entrepreneurship’s EA Training Report (2021) identified the following professional development needs for EAs:

Communication and interview skills:
Earning to give skills (negotiating, fundraising, rising in the corporate ladder)
Policy Skills
Entrepreneurship
Grantmaking
Self Care/Burnout

More concrete subskills are listed in the Appendix

More general skills gaps:

Epistemics
Reasoning ability
Good judgment
Handling feedback
Senior researchers
Social skills
Formal measurement and evaluation experience
Volunteer management and utilization

^
The 2021 EA mental Health Survey (n=326), the 2023 EA mental health survey (n=167), and the 2020 Effective Altruism survey (n=535). See more in the appendix.

Should we invest more in supporting altruists to be at their best? (Mental health + productivity)