The science of small acts of kindness — and why they land bigger than you think

We are poor forecasters of our own warmth

The most useful finding in the modern science of small acts of kindness is not about kindness at all. It is about prediction. Across a decade of experiments — many from Nicholas Epley's group at the University of Chicago — the same pattern repeats: before a small warm gesture, people forecast awkwardness and indifference; after it, both sides report something closer to delight. We hold an entire inventory of undelivered compliments, unsent letters, and unmade calls, curated by a forecasting error.

The founding study put commuters on Chicago-area trains and buses and assigned some to do the thing everyone avoids: talk to a stranger. Those forecasting the experience predicted solitude would be more pleasant. The data said the opposite — the commuters assigned to converse had more positive commutes than those who kept to themselves. The barrier, the researchers found, was not that strangers dislike being spoken to; it was that we underestimate how interested they are in connecting.

That single miscalibration turns out to be the lock on a whole corridor of doors. What follows is the evidence for opening them, one small act at a time — and, at the end, a way to run the experiment on your own week.

[1]

The letter you never sent

Consider the thank-you letter — the one you have mentally drafted to a teacher, a friend, a parent, and never delivered. Amit Kumar and Nicholas Epley asked people to actually write and send such letters, and to predict how they would be received. Writers significantly underestimated how surprised and how positive recipients would feel, and overestimated how awkward the exchange would be. Crucially, that miscalibration was not idle: it predicted people's reluctance to send letters at all.

In other words, the inner editor telling you the letter would be excessive, mawkish, or strange is systematically wrong about the person on the other end. Gratitude, in this literature, behaves like a currency that gains value in transit — worth more upon arrival than the sender ever books at departure. We explored the interpersonal side of this in our companion article on the research behind better relationships: gratitude expressed strengthens bonds in ways gratitude merely felt does not.

[2]

Compliments, and the gap in how much you were liked

The pattern holds for the smallest currency of all. Across nine experiments, Xuan Zhao and Epley found that compliment-givers consistently underestimated how good their compliments would make recipients feel, and overestimated how awkward the moment would be — a forecast gloomy enough to stop many compliments from ever being paid.

Even after a conversation has gone well, the pessimism lingers. Erica Boothby and colleagues documented what they call the liking gap: in laboratory conversations, first-year college dorms, and workshops alike, people systematically underestimated how much their conversation partners liked them and enjoyed their company. You walk away auditing your own missteps; the other person walks away having simply enjoyed you. Both of you, statistically, are wrong in the same lonely direction — which means the kind interpretation of an encounter is also, on average, the accurate one.

Notice what these two findings do when placed side by side. The liking gap says you were received more warmly than you believe; the compliments research says your warmth would be received better than you predict. The pessimism runs both ways across the same channel, and it compounds — each unpaid compliment confirms the caution that produced it. Which is why the researchers' remedy is not confidence training but simple delivery: the forecast corrects only on contact.

[3][4]

The check-in text, and the acquaintances who count

Then there is the person you have been meaning to contact. Peggy Liu and colleagues ran thirteen experiments on exactly this gesture and titled the paper for its result: “The Surprise of Reaching Out.” Initiators robustly underestimated how much recipients appreciated a check-in text, a note, or a small gift — and the appreciation was driven precisely by the recipient's pleasant surprise at being thought of. The staleness that makes you hesitate is the ingredient that makes the gesture land.

Nor does connection run only through intimates. Gillian Sandstrom and Elizabeth Dunn tracked daily interactions and found that on days people interacted with more acquaintances — classmates, baristas, the peripheral cast of a life — they reported greater happiness and belonging. Weak ties are not diluted friendship; they are a distinct nutrient, and most of us are mildly deficient.

Taken together, the two studies redraw the map of a social life. The strong ties are the load-bearing walls; but the texture of a good day is often set at the edges — the familiar face at the counter greeted by name, the colleague from another floor, the neighbor whose dog you know better than their surname. These are gestures with no agenda and no follow-up required, which is precisely what makes them light enough to perform daily. A life warm at the periphery tends, in Sandstrom and Dunn's data, to feel warmer at the center too.

[5][6]

Five dollars, given away

Small acts also move through money, in amounts too modest to feel like philanthropy. In a series published in Science, Elizabeth Dunn, Lara Aknin, and Michael Norton combined a national survey, a windfall field study, and a randomized experiment: people assigned to spend money on others — as little as five dollars — reported greater happiness than those assigned to spend it on themselves. The follow-up work scaled the question to the planet: Gallup World Poll data from 136 countries, plus experiments in Canada, Uganda, India, and South Africa, showed prosocial spending predicting greater happiness in poor and rich countries alike — the authors call it a psychological universal.

It is worth being precise about the size of these effects, because precision is a form of respect. Oxford's Oliver Curry and colleagues meta-analyzed 27 kindness-intervention experiments covering 4,045 people: performing acts of kindness improved the actor's well-being with a small-to-medium effect (δ = 0.28), with no sign of publication bias. Kindness is not a cure and no serious researcher sells it as one. It is a reliable, modest lever — which, pulled often and cheaply, is exactly what a practice is made of.

[7][8][9]

Helping, when life is heaviest

There is a reasonable objection to all of this: small acts are easy to praise from inside an easy life. The literature has an answer, and it comes from the hardest seasons rather than the gentlest ones.

The most sobering study in this literature followed 846 adults for five years, recording stressful life events and helping behaviour towards friends, neighbors, and family. Major stress predicted mortality over the follow-up period — but only among those who did not help others (a hazard ratio of 1.30). Among the helpers, the stress–mortality link vanished (HR 0.96; interaction HR 0.58).

This is an observed association, not a prescription, and helping others is no substitute for care when life becomes too heavy to carry. But the direction of the finding is quietly radical: in this sample, generosity was not a luxury of the untroubled. It sat closest to the people under the most pressure — and appeared, in the data, to sit protectively.

[10]

Run the experiment on one week

Every study above shares a design you can borrow: take the gesture your forecast argues against, deliver it anyway, and compare the result to the prediction. A week is enough, and the discipline of a real study — hypothesis first, observation after — is what separates this from a vague resolution to be nicer. The instruments are a compliment, a letter, a text, a coffee bought for someone else: nothing you do not already own.

Run it the way the researchers did. Before each act, write one line predicting how it will land — awkward, forgettable, appreciated. Deliver it. Afterwards, write one line about what actually happened, and set the two beside each other. If your results resemble a decade of published data, the forecasts will be reliably gloomier than the outcomes — and it is hard to keep obeying a forecaster you have caught being wrong all week. Marcus Aurelius, who wrote to himself for an audience of one, closed the argument seventeen centuries before the data arrived.

For the reaching-out day, some readers borrow a page from RiseHush itself: when a line in the feed reads like someone you know, sharing it as a small piece of quote art is a check-in that arrives as a gift — precisely the shape Liu's experiments found most underestimated. However you send it, choose at most three acts below. Small enough to survive a busy week; that is the method, not a compromise.

“Waste no more time arguing what a good man should be. Be one.”

Marcus Aurelius · Meditations, X.16

[2][5][7]