The Approach

Your prediction is wrong

Epley and Schroeder ran nine experiments with more than 800 people: commuters on Chicago trains and buses, cab passengers, and people in a lab waiting room. Each person was told to talk to a stranger, to sit alone, or to do what they normally do.

The people who talked to a stranger enjoyed the journey more than the people who sat alone, and they got just as much done. Another group of commuters was asked to predict the result, and they predicted the opposite.

The biggest mistake was about rejection. People predicted that fewer than 47% of commuters would be willing to talk. In fact, everyone who returned a survey had the conversation, and nobody reported being turned away. The authors put it plainly: commuters thought talking to a stranger carried a real risk of rejection, and as far as the study could tell, it carried none.

What to do

Treat the fear as a guess, not a fact

The feeling that it will go badly tells you nothing about the people around you. It is a known bias, it has been measured, and it always points the same way. Almost everyone near you has the same fear about talking to you.

In nine experiments, people who were told to talk to a stranger had a better experience than people who were told to sit alone. Other participants had predicted the reverse. Nobody who tried a conversation reported being turned away.

Limits of the study: The studies were done on public transport and in cabs, mostly in Chicago. The effect has been repeated many times, but a commuter train is a low-pressure place where people already share a situation.

The liking gap

After a conversation, people consistently underestimate how much the other person liked them and enjoyed their company. Boothby and colleagues named this the liking gap.

It showed up in more than one setting. It held for strangers meeting in a lab, for first-year students getting to know their hall-mates over months, and for members of the public at a workshop.

The cause is that you and the other person have different information. You remember every clumsy thing you said. They remember someone who made an effort.

What to do

Do not trust your own verdict on how it went

Your judgement after a conversation is more negative than the truth. That matters most in the ten minutes afterwards, when you are deciding whether to ever try again.

With strangers in the lab, dorm-mates over several months, and workshop participants, people underestimated how much the other person liked them and enjoyed their company.

Step 1: Talk about what is around you

The commuters who talked to strangers did not use clever lines. They were in the same situation, such as the same delayed train, and that gave them a reason to speak.

A remark about the person asks them to react to you, and to work out in one second what you want. A remark about the place you are both in asks for nothing. They can answer with a shrug, and nobody loses anything.

What to do

Comment on the situation, not on them

Say something true, ordinary and obvious about the situation you are both in: the queue, the rain, the delay, the noise. You do not need a reason to speak. Being in the same place is reason enough.

This is our interpretation. Epley and Schroeder measured how the conversations turned out, not how they started. The idea that a shared situation gives you a reason to speak is ours, and we want to be clear about that.

Step 2: Ask a deeper question

People expect a deeper question to feel awkward with a stranger. They are wrong about that too. Kardas, Kumar and Epley found that conversations between strangers felt less awkward than people expected, and left them feeling more connected and happier than they expected.

The error was biggest for deep conversations. People overestimated how awkward a deep conversation would be much more than a shallow one. So the more worthwhile the conversation, the more the fear exaggerates.

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

What to do

Ask one question that goes below small talk

This is not an interrogation. Ask one honest question about what is behind the topic. For example, ask what they are hoping for from the trip, not only where they are going.

Deep conversations between strangers felt less awkward than people predicted, and created more connection and happiness than they predicted. The prediction error was larger for deep conversations than for shallow ones.

Limits of the study: Most of the studies paired people in a lab and gave them questions to ask. The direction of the error is well supported. Its exact size in a real bus queue has not been measured.

Step 3: Do it every day

Sandstrom, Boothby and Cooney ran a week-long programme. It was a scavenger hunt that made people find, approach and talk to strangers every day.

Fear of rejection dropped a lot compared with a control group, and it was still lower a week after the programme ended. Participants also felt more able to talk to strangers, both at the end and a week later. The control group did not change.

This is how graded exposure normally works. It is why the third step is a schedule and not a mindset.

What to do

Plan the practice, and the courage follows

One approach a day for a week does more than any amount of deciding to be confident. You do not argue the fear away. You wear it down by repeating.

A week of approaching strangers every day reduced fear of rejection a lot compared with a control group. The reduction lasted at least a week after the programme ended.

Let them leave. This method only works when the other person can walk away easily, as in a queue, on a platform or in a bar. On a late train, in a lift, or on an empty street at night, they cannot, and the same sentence feels threatening. If they cannot leave, do not approach.

A stranger who does not want to talk is not judging you. Accept the answer you get and leave it there. One short reply with no question back means no. The skill is hearing it the first time.

Notice what the research does not say. It does not say every conversation will go well. It says your prediction is too negative, which is a smaller claim and a more useful one.