The Approach

The forecast is wrong

Epley and Schroeder ran nine experiments with more than 800 commuters on Chicago trains and buses, randomly assigning them to talk to a stranger, to sit in solitude, or to commute as usual.

People assigned to connect reported a more positive journey — and no less productive one — than those who sat alone. A separate group of commuters, asked to predict, expected exactly the opposite.

The rejection fear was the sharpest error. Participants predicted that fewer than 47% of commuters would be willing to talk at all. In the event, everyone who returned a survey reported having had the conversation, and not a single person reported being rebuffed. The authors' summary is blunt: commuters thought talking to a stranger posed a meaningful risk of social rejection, and as far as they could tell, it posed none.

The move

Treat the dread as a reading, not a verdict

The feeling that this will go badly is not information about the room. It is a known, measured, directional bias — the same one almost everyone in the carriage is having about you.

Across nine experiments, participants randomly assigned to connect with a stranger reported a more positive experience than those assigned to solitude, while separate participants predicted the reverse. No participant who attempted a conversation reported being rebuffed.

Limits — Public transport and cabs, mostly in Chicago. It is a well-replicated effect, but a commuter train is a low-stakes setting with a shared context already supplied.

The liking gap

After a conversation, people systematically underestimate how much their partner liked them and enjoyed their company. Boothby and colleagues named it the liking gap.

It was not a one-room effect. It held for strangers meeting in the laboratory, for first-year students getting to know their hall-mates over months, and for members of the public at a personal development workshop.

The mechanism is asymmetric information. You have a transcript of everything clumsy you said. They have an impression of someone who made an effort.

The move

Discount your own read of how it went

Your post-conversation verdict is systematically pessimistic. That matters most in the ten minutes afterwards, when you are deciding whether to ever do it again.

Across strangers in the lab, dorm-mates over months, and workshop participants, people underestimated how much their conversation partners liked them and enjoyed their company.

Move 01 — open on the room

The commuters who connected were not using lines. They had a shared situation — the same delayed train — and that was the whole permission structure.

A remark about the person asks them to react to you, and to decide in one second what you want. A remark about the room you are both standing in asks for nothing and can be answered with a shrug at no cost to anyone.

The move

A compliment is a demand. A comment is an offer.

Say the true, boring, obvious thing about the thing you are both already in: the queue, the rain, the delay, the noise. You do not need a reason to speak. Being there together is the reason.

That the shared situation supplies the permission is our reading of how the connection condition actually worked — Epley and Schroeder measured the outcome, not the opening line. We would rather say so than dress it up.

Move 02 — go deeper

People expect that a deeper question to a stranger will be awkward, and they are wrong about that too. Kardas, Kumar and Epley found that conversations between strangers felt less awkward, and produced more connectedness and happiness, than the participants expected.

The misprediction was strongest exactly where it does most damage: people over-estimated the awkwardness of deep conversations far more than shallow ones. So the fear scales with the thing worth doing.

The proposed cause is that we underestimate how interested other people are in what we have to say.

The move

Ask the question one level below small talk

Not an interrogation — one honest question about the thing under the topic. What they are actually hoping for from the trip, not where they are going.

Deep conversations between strangers felt less awkward and created more connection and happiness than participants predicted, and the misprediction was larger for deep conversations than shallow ones.

Limits — Largely laboratory pairs given assigned questions. The direction of the error is robust; the exact size in a bus queue is not established.

Move 03 — repetition

Sandstrom, Boothby and Cooney built a week-long intervention: a scavenger hunt that required participants to find, approach and talk to strangers every day.

Fear of rejection fell substantially compared with controls, and the reduction was still there a week after the intervention ended. Participants also felt better about their general ability to talk to strangers, at the end and at follow-up. The control group did not.

This is the ordinary shape of graded exposure, and it is the reason the third move is not a mindset but a schedule.

The move

Book the reps, not the courage

One approach a day for a week does more than any amount of deciding to be more confident. The fear is not argued down, it is worn down.

A week of repeatedly approaching strangers substantially reduced fear of rejection relative to controls, and the reduction persisted at least a week after the intervention ended.

Let them leave. The whole method assumes a person who can walk away from you at no cost — a queue, a platform, a bar. On a late train, in a lift, on an empty street at night, that assumption is false, and the same opening line becomes something else entirely. If they cannot leave, it is not the place.

A stranger who does not want to talk is not delivering a verdict on you. Read the answer you get and let it be the end of it. One sentence back and no follow-up is a no, and the skill is hearing it the first time.

And note what the research does not say. It does not say every conversation will go well. It says your forecast is biased in a known direction, which is a smaller and more useful claim.