If this is your first time reading, please check out the overview for Poor Ash’s Almanack, a free, vertically-integrated resource including a latticework of mental models, reviews/notes/analysis on books, guided learning journeys, and more.
Social Connection / Social Proof /Reciprocity Mental Model
If you only have three minutes, this introductory section will get you up to speed on the social connection / social proof / reciprocity mental model.
The concept in one quote:
The surest thing I took away from my Ph.D in social work is this: connection is why we’re here. We’re hardwired to connect with others, it’s what gives purpose and meaning to our lives, and without it there is suffering. – Brene Brown
(From “ Daring Greatly”)
The concept in one sentence: given humans’ unique ability to get mileage out of culture, being connected to others – and liked by them – is hugely adaptive, and thus biologically hardwired into us, which can cause some interesting side effects.
Key takeaways/applications: other people very strongly influence our behavior, often without our knowledge or consent; becoming aware of the power of this model can help us make more accurate and independent decisions, while also influencing others’ behavior (hopefully for good.)
Three brief examples of social connection / social proof / reciprocity bias:
What’s one of the keys to happiness? “Love – full stop.” Those are the words of George Vaillant, long-time director of the massive longitudinal “Harvard Men” study. This is a well-replicated finding elsewhere, as in Brown’s work, and Shawn Achor’s “ The Happiness Advantage.”
Why are popular songs popular? Because they’re popular. Confused? It’s okay; you should be. The answer is social proof – in addition to path-dependency, as I explore in that mental model.
The “MusicLab” experiment demonstrates, as do many others that we’ll discuss, that our displayed preferences are strongly influenced by those of our peers.
Why do cockatoos in the Australian outback swear at humans? Social connection and social proof are underlying drivers of culture, which – as explored in that model and touched on briefly here – is a phenomenally powerful driver of results. Jennifer Ackerman’s “ The Genius of Birds” provides a refreshing and interesting exploration of some social behavior in a non-human context.
If this sounds interesting/applicable in your life, keep reading for unexpected applications and a deeper understanding of how this interacts with other mental models in the latticework.
However, if this doesn’t sound like something you need to learn right now, no worries! There’s plenty of other content on Poor Ash’s Almanack that might suit your needs. Instead, consider checking out our learning journeys, our discussion of the schema, inersion, or margin of safety mental models, or our reviews of great books like “ Seeking Wisdom” , “ The Making of the Atomic Bomb” , or “ The Frackers” .
Social Connection: A Deeper Look
“We believe that the most terrifying and destructive feeling that a person can experience is psychological isolation… a feeling that one is locked out of the possibility of human connection and of being powerless to change the situation.
People will do almost anything to escape this combination of condemned isolation and powerlessness.”– Jean Baker Miller and Irene Stiver of the Stone Center at Wellesley College, via Brene Brown’s “ Daring Greatly”
The human need for social connection, and the consequent phenomenon of “social proof,” is one of the more thoroughly replicated findings in modern psychology.
Logic similar to the above quote, for example, is invoked by Laurence Gonzales in the wonderful “Deep Survival” to explain the unique sheer terror of being lost – that we’ll never be seen again. What is it those who are lost want? More than even food or water or safe shelter, it’s to see their loved ones again.
This comes up over and over and over again. For example, in “ The Lonely American” by Olds/Schwartz, the authors discuss how among children,
“most likely moment […] of tears or rage […] is when a child feels left out.”
They also cite solitary confinement and social shunning as examples of punishment. They clarify that it’s not so important whether someone’s actually left out – simply feeling left out is enough to trigger these sorts of feelings.
How important is social connection to our well-being? Loneliness is the new smoking. Olds/Schwartz go on to note:
“there is now a clear consensus among medical researchers that social connection has powerful effects on health. Socially connected people live longer, respond better to stress, [and] have more robust immune systems
[…] these medical benefits derive directly from the social connection itself, not just from lifestyle improvements, such as better diet, more exercise, and better medical care, that might go along with it.”
I discuss their interesting book more in the local vs. global optimization mental model.
Nor is this constrained to humans: look at man’s best friend. In “Dog Is Love,” Clive Wynne compiles evidence that our bond with dogs is powered less by IQ and more by hypersociability that shows up in the body. Functional MRI work finds many dogs prefer praise from their person to food, and hormone studies show oxytocin spikes for both species simply from friendly eye contact or reunions. Genetic work links this friendliness to variants also implicated in human sociability. Wynne also explains how severely distressed many dogs can be when left alone for long periods of time (such as when humans are at work). For both dogs and humans, social connection with each other can be as powerful as with other humans.
Meanwhile, as Johann Hari explains in Lost Connections, a lot of depression can be traced to ruptures in belonging rather than to purely chemical glitches. Hari synthesizes research on how loneliness elevates stress hormones, and how communities with stronger bonds show better health and resilience independent of lifestyle confounders.
He extends the discussion in Chasing the Scream, where he contends that addiction is only partly chemical, and is far less likely to occur with stable social connections. Hari discusses “Rat Park,” Bruce Alexander’s experiment that reframed compulsive drug use as an adaptation to a barren cage rather than a chemical destiny. When rats had company, space, and stimulation, morphine lost its grip, which fits your argument that environment and belonging steer behavior at least as much as willpower. Important caveat: Rat Park is debated and has been hard to fully replicate, so treat it as one suggestive data point among many rather than as a silver bullet. Nonetheless, it’s intriguing to think about.
Social Connection x Incentives / Utility x Culture, Part 1: Social Proof
“[We’re often] reactively liv[ing] the scripts handed to us by family, associates, other people’s agendas, the pressures of circumstance […]
these scripts […] rise out of our deep vulnerabilities […] our needs for acceptance and love, for belonging, for a sense of importance and worth, for a feeling that we matter.”
– Stephen Covey in “ The 7 Habits of Highly Effective People”
Covey, here, covers one of the two main reasons I believe the social proof tendency exists. That first reason would be that people typically like us more when we are like them, and behave in a manner that is familiar and acceptable to them.
This is important because of the discussion below of reciprocity bias (risk reduction), as well as the benefits of culture and marginal utility – all of which require other people.
The second reason is that, in aggregate, there’s usually signal in what other people are doing. Take the classic “wisdom of the crowd” experiment: Philip Tetlock discusses this in “ Superforecasting” . He notes that, in aggregate, the crowd gets everyone’s utility from different pieces of information, with the biases canceling out.
So, if a lot of other people coalesce on an answer, that’s a reasonable starting point or base rate (though it is not, in any scenario, right all the time).
Jennifer Ackerman attacks this idea from a different angle in “ The Genius of Birds” : she observes how copying other birds is a low-cognitive-expense form of a/b testing a new idea:
“This kind of social learning – copying fellow birds in a local environment – say the researchers, might be a quick and cheap way of acquiring successful new behaviors without undertaking potentially risky trial-and-error learning.”
What’s the net-net of this? Humans display a very strong tendency to conform to others’ behavior.
Dogs follow a similar principle. As Clive Wynne explains in Dog Is Love, shelter dogs with no special training can learn to follow a human pointing cue in minutes; street-dog studies show that prior friendly contact makes them more willing to approach the same person later.
The social-proof phenomenon shows up in so many books that I can’t list them all, so I’ll just go with one of my favorites: Cass Sunstein and Richard Thaler’s “Nudge” . Chapter 3 in that book does a great job of reviewing a lot of the research.
Here are some of the more interesting bits: social proof influences everyone from pregnant teenagres to broadcasters to federal judges, so as with all the other cognitive biases covered on this site, we’re not exempt just because of our intelligence or education.
In fact, social proof can make our brain go totally nanners: Sunstein/Thaler cite some experiments, replicated 130 times in 17 countries ranging from Zaire to Kuwait, which find that:
“When everyone else gave an incorrect answer [on a test when the answers are obvious]… people erred more than one-third of the time… notice… people were responding to the decisions of strangers, whom they would probably never see again. They had no particular reason to want these strangers to like them.”
Sunstein/Thaler go on to note a really interesting wrinkle, which explains why the Internet (and particularly comment boards) can be such a horrible place sometimes:
“On the other hand, social scientists generally find less conformity… when people are asked to give anonymous answers. People become more likely to conform when they know that other people will see what they have to say.”
This can be triggered quite easily: put a pair of eyes above a communal coffee-fund collection jar, and you’ll get more donations. Sunstein/Thaler go on to cite a lot of other fascinating examples, like a windshield-pitting “epidemic” in Seattle that proved to just be social proof.
Application / impact: social proof is powerful. Telling people that others like them behave in a certain manner is one of the most effective ways to gain compliance with desired behavior.
Social Connection x Incentives / Utility (x Dose-Dependency xProbabilistic Thinking x Luck), Part 2: Reciprocity Bias
Taking social proof a step further is the idea of reciprocity bias: our tendency to play “tit for tat,” whether the “tat” is good or bad.
Reciprocity bias is often conceptualized in the sense of favors – i.e., “you scratch my back and I’ll scratch yours” – but by inversion, it can also be found in Old-Testament standards of justice: “an eye for an eye, a tooth for a tooth.”
Reciprocity bias is famously covered in Robert Cialdini’s classic “Influence: The Psychology of Persuasion.”
That book observes, sagely, how we have to be careful not to fall for it – i.e., not to let other people engage our reciprocity bias by giving us a gift we don’t really want or need.
I’m not going to catalog examples, since they tend to be fairly obvious and widespread, although I discuss one from Dr. Jerome Groopman’s “ How Doctors Think” in the contrast bias mental model.
I have seen some evidence cited – although I’m not hugely familiar with it – that “TIT FOR TAT” is actually the evolutionary equilibrium strategy. Whether or not that’s true, there’s definitely an underlying logic.
Megan McArdle does a great job of exploring the underlying logic of reciprocity bias in “ The Up Side of Down.” In high-luck, probabilistic endeavors like big-game hunting or ice fishing, there’s high marginal utility in sharing, if there’s a mechanism for enforcing that sharing – which reciprocity bias provides.
It may not be a uniquely human trait, either. In what may be a case of convergence, for example, it turns out that crows display something resembling reciprocity bias – see Jennifer Ackerman’s “ The Genius of Birds” for some fascinating discussion of this.
That said, one of the things that I think people have a really hard time understanding with regards to reciprocity bias is that while it’s a strong, biologically-engrained tendency, that doesn’t mean it applies in every situation, nor that everyone will participate. For example, Kip Tindell, author of “ Uncontainable” and co-founder of The Container Store, mentions in Uncontainable that:
“Rarely does someone, in any walk of life, just take, take, take and not reciprocate. Most people instinctively know to try love – it will come back to you in spades.”
I haven’t found this to be true whatsoever: that is not a rare behavior. It is, in fact, a common one. Many of us have had experiences with people who are quite content to take and take and take without giving anything back.
Some of this is due to self-justification (as discussed in the contrast bias mental model, riffing on Tavris/Aronson’s “ Mistakes were Made (but not by me)”), while other bits are inexplicable.
Brad Stone’s “The Upstarts” provides one good example of this, exploring an Uber competitor.
IBM veteran John Wolpert, the founder of “Cabulous,” another yellow-cab platform, turned down Bill Gurley of Benchmark because, according to John Wolpert:
Wolpert was conceiving of a service that empowered those yellow-cab drivers. It would make the traditional taxi businesses more efficient and help drivers boost their earnings. This was his fatal mistake […] Cabulous was doomed by civility.
“I tried to be the nice guy. was very into the win-win in those days. To a fault. I’ve learned a lot about negotiation since then. [Gurley wanted to invest but the round was already full.] I was a Boy Scout. I was going to go with the date that brought me.”
Some of the evolutionary material I was talking about discusses how “cheating all the time” is a strategy that dominates “tit for tat” if you allow it to develop – which explains Uber, in a sense. It is also related to McArdle’s path-dependency discussion of capitalism vs. corruption in “ The Up Side of Down” .
Indeed, this sort of take-take-take, anti-reciprocity-bias behavior is so prevalent a theme among love songs throughout the ages that it’s a trope. Take Gotye’s “Somebody That I Used To Know”:
“You didn’t have to cut me off… make out like it never happened and that we were nothing.”
Or Rise Against’s “Audience of One” off their 2008 album Appeal To Reason:
“I brought down the sky for you,
but all you did was shrug.
You gave my emptiness a name…
when you ran away.
Now all my friends are gone.
Maybe we’ve outgrown,
all the things that we once loved.”
Ultimately, reciprocity is a real and strong phenomenon that will work, given a high enough sample size – but that doesn’t mean it will work in every individual situation.
Application / impact: be aware of the power of reciprocity bias: both on yourself, and on your interactions with other people.
Social Connection: The AI Angle
AI is interesting in the concept of social connection in several ways. The first is in its original use case, as a “chat” bot that you can literally have a conversation with. LLMs can appear empathetic because they are capable of modeling the mechanics of compassion or having a personality, even though they are just synthetic constructs. I’ve had conversations – with both Grok and ChatGPT – where I temporarily forgot that I was talking to a computer; it felt very real, and I had some of the same warm-fuzzies that I get from conversations with real friends, even though I fully knew I was talking to a machine.
This simulation of human traits introduces a critical challenge; the concept was brought into sharp focus with the release of GPT-5. Widespread public backlash was, at least in part, attributed to many people forming parasocial relationships with an LLM which they had anthropomorphized. When the new model was introduced, the drastic change in the “personality” of ChatGPT felt to many users like the loss of a friend (an example of deprival superreaction syndrome).
This kind of relationship with AI can potentially crowd out rich, real-life social connections that data shows are important for mental well-being; teens and kids, particularly those in need of social connection, might be especially vulnerable to forming these relationships. Without proper guardrails, AI could amplify their worst impulses rather than steering them back on track.
Conversely, because AI is so good at handling rote mechanical work, that can free up our time – less time to spend on emails, more time to spend with people.
