Hide table of contents

This post is a submission to Open Philanthropy's Cause Exploration Prize, specifically its Worldview Investigations prompt which looks for essays that propose an equivalent DALY-like metric for a hard-to-measure state (such as freedom, empowerment, injustice, or discrimination) that we could ultimately use to undertake similar comparisons.

Here, I make a case for Social Relationships as an important, quantifiable factor in wellbeing, and explore the use of the WHO-QOL metrics for Social Relationships.

Intro: If Students Hate School, Why Do Still They Go?

In late 2019, this author was given a grant to reform education in Singapore. Since high school, I had worked on at least five student-run projects aimed at education reform. This included a collaborative online study notes database and a subreddit for Singaporean students.

As anyone in education can tell you, positive outcomes can seem both accessible yet elusive. This time, I wanted to make sure I was solving the right problems. I posted a simple question on the subreddit:

"Do you look forward to school? Why or why not?"

This question was deliberately vague. Much like this Cause Exploration Prize, we were hoping people would interpret the question so broadly that we'd get totally unexpected answers with entirely new perspectives. After ~200 anonymous submissions[1], we got our surprise.

Despite the incredibly vague wording, anonymous individual submissions and open-ended response fields to garner diverse responses, a majority of answers said the exact same thing:

  1. Students had an overall negative perception of their academic experience.
  2. Students stated that they looked forward to school due to peer relationships (~60%), teacher relationships (20%), or both (20%).
  3. Self-assessment of social relationships was the determining factor of overall student satisfaction in all cases, regardless of academic experience.

Social Relationships: The neglected factor behind student wellbeing

The survey results were surprising, yet intuitive in hindsight. Surprising because no one had really mentioned it before. Education interventions and policy debates generally centre around student engagement with curricula, educators and academic material. Decision makers closely watch standardised test outcomes, retention/dropout /graduation rates and countless other metrics to determine whether a policy works. Peer interaction is often treated as a confounding variable, distraction or white noise, if it's mentioned at all.

The survey posited an interesting hypothesis:

Students' wellbeing was primarily determined by the self-perception of their relationships, a factor frequently neglected in education interventions.

A Quick Glossary of Concepts

Social relationships -  interpersonal connections between individuals. Includes romantic relationships, direct spouses, family and close friends.

Happiness - positive self-evaluation

Wellbeing - health+prosperity+happiness/ physical+financial+mental

Suffering - self-reported mental and physical pain

Social Relationships: A primary factor in wellbeing

As my project rolled along, I remained intrigued by that first hypothesis. It was such an interesting premise. Peer relationships are not a frequently-cited factor when assessing wellbeing outcomes, yet they could be the most important factor.

Clearly, I couldn't let this go, so I searched for other, more rigorous, data. The case for Social Relationships as a major factor in mental, physical and financial wellbeing only seemed to grow.

  • Youth happiness Data from Berkeley's Greater Good Science Centre covering >2,500 respondents found that "students’ self-reported measures of well-being increasingly improve with better ratings of relationship quality from peers".
  • Lifelong happiness An 80-year, intergenerational study of Harvard alumni found that "close relationships, more than money or fame, are what keep people happy throughout their lives [...] Those ties protect people from life’s discontents, help to delay mental and physical decline, and are better predictors of long and happy lives than social class, IQ, or even genes." Another study concluded similarly, accounting for gender.
  • Lifelong health For the elderly, who face lower quality of life due to health complications, the study found: “When we gathered together everything we knew about them about at age 50, it wasn’t their middle-age cholesterol levels that predicted how they were going to grow old,” [...] “It was how satisfied they were in their relationships. The people who were the most satisfied in their relationships at age 50 were the healthiest at age 80.
  • Mitigating illness and suffering Another study found that social relationships may directly influence elderly pain perception: "People who had happy marriages in their 80s reported that their moods didn’t suffer even on the days when they had more physical pain. Those who had unhappy marriages felt both more emotional and physical pain." [...] "According to the study, those who lived longer and enjoyed sound health avoided smoking and alcohol in excess. Researchers also found that those with strong social support experienced less mental deterioration as they aged."
  • Reducing mortality risk A meta-analysis found: "Across 148 studies, the random effects weighted average effect size [indicated] a 50% increased likelihood of survival for participants with stronger social relationships. This finding remained consistent across age, sex, initial health status, cause of death, and follow-up period. The association was strongest for complex measures of social integration and lowest for binary indicators of residential status (living alone versus with others).Other studies found lower morbidity and mortality outcomes and positive effects on cardiovascular, neuroendocrine, and immune function, even for severe stroke and HIV patients.
  • Worker satisfaction and engagement - As several Gallup polls concluded, "We've consistently found that for women and men having a best friend at work leads to better performance." Furthermore, Gallup's research "revealed that just 30% of employees have a best friend at work. Those who do are seven times as likely to be engaged in their jobs, are better at engaging customers, produce higher quality work, have higher well-being, and are less likely to get injured on the job. In sharp contrast, those without a best friend in the workplace have just a 1 in 12 chance of being engaged."
  • Worker retention And if you recoil at the mere implication of company spirit, this other Gallup finding might be of interest: "While companies often pay significant attention to loyalty toward the organization, the best employers recognize that loyalty also exists among employees toward one another. All employees have "leaving moments" when they examine whether to leave or stay at an organization. The best managers in the world observe that the quality and depth of employees' relationships is a critical component of employee loyalty."
  • Cross-culture prevalence The international World Values Survey found that "in almost all countries, people who often spend time with their friends report to be happier than those who spend less time with friends."

Overall, multiple studies across different contexts and demographics suggest a wide-ranging and significant correlation between social relationships and wellbeing.  Or, as Dr Emma Seppala puts it:

Social connectedness generates a positive feedback loop of social, emotional and physical well-being.

Quick Analysis: Did all Social Relationships affect wellbeing?

To summarise/oversimplify the findings:

  1. Significant correlations occur with the presence of any strong social relationships. Most of the studies cited above have tested and claimed causal relationships.
  2. Magnitude of positive/negative sentiment within a relationship seems to matter more than specific type. A close, positive and strong platonic work relationship seems to impact wellbeing measures more than mildly positive romantic relationship. This hasn't been cross-tested across different contexts, however.
  3. Self-perception is generally accurate. People generally concur on mutual perceptions of relationship quality.
  4. Negative relationships can negatively affect wellbeing.[2]
  5. Multiple studies reiterate that quality matters more than quantity. Having just one "exceptional"/"intimate"/"best" relationship seems to result in a higher self-assessment/"peak" rating, which translates to better wellbeing outcomes than having many mildly positive relationships.

Measuring Social Relationships

Disclaimer: There are a lot of ways to measure Social Relationships. There is absolutely no metric I can suggest that will be 100% comprehensive. If you're a subject matter expert, I fully welcome critique and discussion.

The most widely applicable diagnostic tool I could find is the World Health Organisation's Quality of Life (WHOQOL) survey. In particular, Domain IV assesses the respondent's social relationships, comprising their personal relationships, social support and sexual activity. The weighing the survey used was as follows:

Social relationships domain = (PERSONAL RELATIONSHIPS score + SOCIAL SUPPORT score + SEXUAL ACTIVITY score)/3

In particular, here are the relevant questions, each asking for a ranking of 1 to 5:

How satisfied are you with your personal relationships?
How satisfied are you with your sex life?
How satisfied are you with the support you get from your family?
How satisfied are you with the support you get from your friends?
How satisfied are you with your ability to provide for or support others?
Do you feel happy about your relationship with your family members?
How would you rate your sex life?
How important to you are relationships with other people?
How important to you is it to support others?
How important to you is your sexual life?
How important to you is it to feel included socially?
How important to you is your home environment?
How important to you is it being able to get adequate social help?
To what extent do you feel guilty when you need the help and care of others?

Methodology of the WHO-QOL at a glance


  • Comprehensive: Covers different aspects of social relationships.
  • Intuitive: The rating system simplifies fairly complex judgements.
  • Culture-agnostic: By focusing less on the details of the specific relationships and more on overall self-perception, it sidesteps certain cultural differences. For example, many Western-centric romantic relationship surveys assume monogamous norms and Western expressions of affection, while this survey does not require such assumptions.
  • Easy to measure: As part of a broader 100-question survey, these dozen short questions can be added without inducing survey fatigue seen in longer questionnaires.
  • Widely tested, in clinical and non-clinical settings. Translations in ~80 languages are available on the WHO website.


  • Less specific: While robust in diagnosing broad trends, it sacrifices specific insights found elsewhere. For example, the Gallup worker engagement survey found significant differences between having work friends and a best friend at work. This tradeoff is somewhat inevitable. More specificity also provides more room for misinterpretation. A child working on a rural family farm may not share a common understanding of work bestie.
  • Subjective: Compared to more common measures of wellbeing (such as GDP, HDI, DALYs etc.), any attempts at measuring social relationships is subjective at its core.  That said, in the absence of known, widespread and identifiable flaws, self-assessments of relationships thus far seem an accurate, intuitive and robust proxy as to whether someone has good social relationships.

Of course, I've simply cherry picked the questions relevant to social support. Nothing's stopping surveyors from adding more detailed questions. For reference, here's the next seven questions in the WHOQOL Survey:

To what extent are you concerned about your HIV status breaking your family line and your future generations?
To what extent are you concerned about how people will remember you when you are dead?
To what extent do any feelings that you are suffering from fate or destiny bother you?
How much do you fear the future?
How much do you worry about death?
How bothered are you by the thought of not being able to die the way you would want to ?
How concerned are you about how and where you will die?

If needed, one could design/add more specific questions assessing relationships between specific individuals, ranking them in detail or using adjacent metrics such as frequency of social contact. This is merely the most common tool used that allows easy cross-comparison with data from other organisations.


  • Causation/correlation - As with all factors in wellbeing, it is difficult to isolate cause and effect. For example, studies also show that financial stress can negatively impact relationship quality. Previously mentioned studies did demonstrate that positive social relationships can directly mitigate suffering from other physical and emotional causes. Overall, it's a feedback loop.
  • Limits vs material wellbeing - Social relationships can affect resilience, individual and groups perceptions and responses. It can even mitigate health risks and improve survival during serious health ailments. However, it should not be expected to fully overcome challenges to physical/material wellbeing.
  • Too good to be true? - Popular psychology sometimes presents a very satisfying and intuitive hypothesis that is accepted as true, parroted as common knowledge and later called into question. Social scientists may recall Greatest Hits such as The Stanford Prison Experiment, "Your Brain Isn't Fully Developed Until You're 25" and the Stanford Marshmallow Experiment. As always, skepticism is warranted.

Implications for interventions and EA

If we accept the premise that social relationships are a major factor in wellbeing, what does that mean for interventions and, more specifically, Effective Altruism (EA)?

Social Relationships as a Controlled Variable

If Social Relationships have such a statistically significant impact on emotional and physical wellbeing, it may be important to control for this factor in future large studies. In the same way studies now account for socioeconomic background, social relationships may be equally important to isolate in future studies to draw more accurate conclusions.

Supplementing global health interventions/approaches

Leveraging the role of Social Relationships could improve the efficacy of existing interventions. We may understand that relationships are important, yet disregard it as a "serious" measure to focus on. We could look at causes already considered Most Effective or Nearly Most Effective and reassess what happens when you apply a relationship-centric iteration of the existing approach. For example, could dedicated social support interventions supplement other health interventions? A meta-analysis of RCTs for psychosocial support interventions found an average 20% increase in survival rates and 29% increase in survival time compared to controls that underwent standard care regimes, suggesting benefits to supplementing existing interventions.

Alternatively, could anti-smoking, anti-obesity or infant health interventions increase in efficacy to a significant degree when they specifically involve the recipient's support network?

Assessing new cause areas

More ambitiously, we may even examine new cause areas, factoring in the ripple effect of fostering positive relationships. If we can quantify the impact of strong social relationships on wellbeing, could we then quantify the wellbeing impact of socially-focused approaches? Could promoting certain forms of social media constitute a compelling cause area, given the increasing ubiquity of digital relationship building? Could we further quantify existing initiatives aimed at reducing societal tensions, stigma and discrimination, which were previously overlooked due to their health impact being primarily social?

Improving EA's Appeal

As Open Philanthropy has noted elsewhere, personal relationships have consistently been a key driver of EA's growth. This mirrors studies of other social movements, which have identified that those who report having close friends within the movement are more likely to be highly engaged volunteers.

A clearer understanding of the benefits of social relationships will benefit both EA cause beneficiaries and EAs themselves.  Multiple studies have found that providing social support and engaging in pro-social behaviour could sometimes be more beneficial than receiving it. This is an even more under-explored phenomenon, and further research could more rigorously assess positive effects on EAs and adoption of EA as a movement. For example, many non-EA interventions with lower quantifiable impact remain more popular, to the confusion of many EAs. If this adoption gap can be explained by the stronger impacts of relationship-focused causes on givers, EA organisations can make more informed decisions to boost the appeal of EA to donors/volunteers/stakeholders.

EA cause areas have substantial literature devoted to the long-term impact of interventions on recipients. However, less attention is given to the long-term impact on givers (i.e. EA-aligned donors/stakeholders). This matters because givers fund interventions. While recipient-centric approaches maximise relative utility (impact per dollar), this doesn't necessarily maximise absolute utility (total impact for total dollars). A suboptimal but popular intervention funded by more people may have higher absolute impact than the most optimal intervention funded by a few. Applying EA-level analytical rigour to boost receptiveness to EA ideas and causes could be a neglected approach.

I am not saying EA should maximise personal utility. I am saying EA should methodically study and double down on altruism. Prosocial systems operate on different incentives and norms, and that's a good thing.[4] Anecdotally, EA culture is uniquely open to helping new EAs, EAs they've never met and strangers in general. Social dynamics in EA are unique. We should try to understand why, so we can build on these strengths even more.

In short, EAs should seek to understand why the general public is more socially engaged with non-EA cause areas, and how we can present the best version of our values.

Summary: Better relationships = better wellbeing outcomes.

Okay. That was a lot. Let's summarise:

  1. Strong social relationships generate a positive, lifelong feedback loop of social, emotional and physical well-being.
  2. Social relationships can be adequately gauged through self-perception surveys.
  3. Accounting for the effects of social relationships can inform more effective intervention approaches and facilitate movement building.

That's it. The premise is broad and simple. Better relationships=better wellbeing outcomes. Now, there's no easy way to wrap such an incredibly all-encompassing concept in a nice bow, so I'll address the remaining questions in this last section.

Other Questions

Why is this so neglected?

Honestly, I had this question myself. Every study I saw cited social support/ relationships as a significant factor in practically every measure of wellbeing I looked at. It was like some incredibly obvious open secret among researchers. It's not news to the public, either. I hardly had to explain this concept to the dozen people I pitched this to, EA or otherwise. Anyone can anecdotally recognise the value of friends, family and loved ones.

And yet, you'd hardly see it discussed in an intervention context. Most people just don't actively associate Social Relationships with systemic interventions. Other EAs have discussed similarly neglected areas such as socioeconomic status, diet and exercise, and sleep (twice). Countless studies have established strong correlations between such factors and general wellbeing, yet it's normal to have entire policy-level discussions without ever considering their impact. One could speculate on why, but that's a much more complex discussion.[5]

How does accounting for this metric improve Open Philanthropy's grantmaking process?

  1. The impact of Social Relationships on DALYs could potentially be isolated and identified, reducing uncertainty of impact assessment for other factors.
  2. Existing top recommendations could be reevaluated, testing and leveraging the impact of Social Relationships to significantly increase efficacy.
  3. Possibility of assessing new cause areas that heavily rely on Social Relationships (mental health, destigmatisation, networking, family-focused interventions etc.)
  4. More targeted approaches in Open Philanthropy's stakeholder outreach.

Final thoughts: Why I Wrote This

Despite being EA-aligned since middle school, and being a climate and education advocate, I haven't interacted with the wider EA community until now.

EA Singapore was never as established as the US or UK. Still, I saw so many asymmetric opportunities to do good that would never pass a formal EA evaluation. Whether it was education reform, creating hyper-scaleable support groups, radical climate protests or helping someone seek an adult ADHD diagnosis, I followed EA principles and metrics, but would never dare use the term. Even today, I have friends working in AI policy, sustainable agriculture and global health who I'd struggle to pitch EA to, because I don't want to imply their work is Ineffective Altruism.

I hope this will help broaden EA's criteria of Doing Good Better. I estimate >90% of promising, prospective EAs might feel excluded from EA, because they feel their work doesn't fit institutional EA criteria. It's a weird cultural problem to have for such a well-meaning movement.[6]

By measuring the role of companionship - a fundamental human need - I hope to finally acknowledge the immense work done by those closest to us.

  1. ^

    We collected far more data, compiled and analysed it in a 28-page report presented to the Singapore Ministry of Education. I could provide the report if anyone asks, but I'd rather not accidentally break data privacy laws yet.

  2. ^

    While this means that Friendship Isn't Magic and relationships can be harmful, for the purposes of designing interventions, we're concerned about isolating confounding variables and leveraging potential new approaches. That harmful relationship effects exist do not diminish the importance of understanding these effects. After all, it's not like GDP is useless because poverty exists.

  3. ^


  4. ^

    An incredibly insightful book on prosocial/altruistic behaviour I'd recommend is Does Altruism Exist? Culture, Genes, and the Welfare of Others
    by David Sloan Wilson

  5. ^

    I would elaborate more, but I've just written 3,000 words during a school holiday so I'm just gonna be lazy here.

  6. ^

    Another example: I convinced 6 EA friends to spend a day brainstorming ideas for this prize, for a guaranteed $200. We thought of Road Safety and Youth Mental Health in LMICs, both of which would receive far better-researched submissions hours before the deadline. We all gave up on suggesting cause areas, and I wrote this mostly to vent.

Sorted by Click to highlight new comments since:

Great post, thanks for writing it!

Some reasons social relationships might be wrongly neglected:

  1. Quality of Life measures and QALYs probably underestimate the effect of social relationships on quality of life and well-being

  2. It seems difficult to measure cost effectiveness of interventions here

  3. The idea of government policy very explicitly and openly trying to affect people’s social lives would feel like an overreach of power to some people (but interventions via NGOs probably isn’t a problem to most people)

Yeah, I didn't want to speculate on why social relationships are underrated at first. That's worthy of a multidisciplinary PhD thesis on its own, and I was stretching my arguments a bit as is.

Quality of Life measures and QALYs probably underestimate the effect of social relationships on quality of life

Yeah, that was a weird hole I found when examining the scientific and policy literature. There's no explicit objection to it, it just never showed up to begin with. What's interesting is that I dug through Vietnamese and Chinese sources, which did make reference to family and community bonds in planning! This makes me wonder if social relationships were more regarded in East Asian societies, and faded out of public discourse as people looked up to more Western empirical/academic approaches. Not to romanticise, of course. There's a reason I left.

It seems difficult to measure cost effectiveness of interventions here

Agreed, haha. This post was less meant to say "here's a really good intervention, and I can prove why you should fund it right now" and more "looking more closely at this metric could greatly improve impact assessments on a huge range of existing causes". I also suggested taking causes already considered Most Effective or Nearly Most Effective and reassessing what happens when you apply a relationship-centric iteration of the approach. So for example, did anti-malaria interventions increase even more in cost-effectiveness when critical patients received frequent social support from their family? Or could bottlenecks in efficacy during interventions or fundraising be bypassed with a bit of peer pressure?

Anecdotally, if you've ever had a chronic condition, you might recognise how a supportive community makes a huge difference. I have late-diagnosed ADHD in a country with low mental healthcare coverage and awareness. >90% of the quality of life improvements from professional treatments would not have happened without peer advice and support, mostly online. I've helped others seek treatment, and all of them said that me just mentioning ADHD improved their decades of their quality of life (ADHD also has a quantifiable disease burden). So to me, there seems to be an untapped reservoir of very effective interventions resulting from social support, that are hard to pin an exact number on, but worth attempting to value because they're just that valuable.

The idea of government policy very explicitly and openly trying to affect people’s social lives would feel like an overreach of power to some people (but interventions via NGOs probably isn’t a problem to most people)

Well, if you frame it like that, I'm compelled to agree. In practice though, there's plenty of interventions with existing social dimensions. Pretty much anything involving minors relies on peer or parental pressure. Heck, the WHOQOL survey I used has been implemented for over a decade with no issues to speak of.

Again, I think this reflects a cognitive dichotomy between family/friends and policy, and people find it weird for the two to intersect even though those concepts blend together all the time without us noticing. Family and friends' financial outcomes has a huge impact on individual outcomes, yet I hardly see it mentioned in tax reform debates.

Curated and popular this week
Relevant opportunities