Evidence status: conditional support Field experiments on public transport and several conversation studies find that people often underestimate strangers’ willingness to engage and the enjoyment, connection and care that follow. Safety, place, cultural norms, the particular partner and genuine choice constrain the result. It is not a requirement that everyone become more social. This is a version 0.1 research draft, with evidence reviewed to August 2026.
How Everyday Psychology Takes Shape · Season Two: How Do We Read Other People? · Article 7
When seated beside a stranger on a train, looking at a phone is the predictable option. A conversation could be awkward. The other person may not want it. A topic might expire after two sentences. Solitude at least has a known outcome.
This calculation is not irrational. Public space requires boundaries, and unknown people can present real risk. Yet when researchers compare anticipated conversations with conversations that actually take place, a recurring difference appears: people often expect interaction to be worse than it is, partly because they underestimate another person’s willingness to respond.
The interesting conclusion is not that introverts secretly love every conversation. The research does not show that. It is that two strangers can each wait for the other to display interest, and their mutual caution helps produce the silence that seems to confirm it.
Separating expected and experienced commutes
Nicholas Epley and Juliana Schroeder published Mistakenly Seeking Solitude in 2014. In train and bus settings, some commuters were asked to connect with a stranger nearby, others to remain disconnected, and others to travel as usual. Participants who completed an interaction reported a more positive commute on average and did not report being less productive.
Separate groups predicted how these conditions would feel. They generally expected solitude to be more pleasant and believed that strangers would have limited interest in connecting. The design therefore distinguished “how I imagine a conversation will feel” from “how a randomly assigned interaction was experienced”. That is more informative than asking whether someone considers themselves a person who likes small talk.
The studies also did not reveal a simple trade-off in which the initiator benefited by imposing an equal cost on the recipient. In the observed public-transport settings, both sides tended to report a more positive experience. That remains a result from research participation on particular routes; it is not a survey of every commuter in every city.
Subsequent work expanded the question to weak ties, brief daily encounters and deeper conversation. Research by Gillian Sandstrom and Elizabeth Dunn found that more interactions with acquaintances or strangers were associated with greater daily belonging and well-being. Correlational parts of that work can include extraversion, mood and opportunity, while experimental components add evidence that interaction itself can matter.
Would a deeper topic be worse?
Weather is relatively safe. A personal concern seems riskier. In research published in 2022, Michael Kardas, Amit Kumar and Epley asked strangers to discuss questions of different depth and to predict awkwardness, connection and enjoyment beforehand.
Participants tended to underestimate how caring and interested a stranger would be during deeper discussion and to overestimate awkwardness. The conversations generally created more connection than expected, and deeper exchanges in these tasks could produce greater connection than shallow ones. The authors argued that underestimating distant others’ care forms a psychological barrier to meaningful conversation.
The phrase “in these tasks” matters. Participants knew that the exchange was organised as research, questions were selected, and both people occupied a relatively symmetric condition of consent. Unexpectedly disclosing trauma, illness or private material to a stranger may impose emotional labour for which that person did not volunteer. Depth is not a quantity to maximise. Reciprocal, gradual willingness is part of the boundary.
Why are forecasts too pessimistic?
The most immediate problem is uncertainty about the other person’s interest. An initiator bears a visible risk of refusal. When both remain silent, nobody records the enjoyable conversation that might have occurred. Failures can enter memory, while unrealised successes produce no data.
When imagining conversation, people also focus on the opening and the feared silence. Once interaction begins, a partner supplies questions, stories and cues. Conversation is no longer a solo performance. The forecast underestimates the other person’s role in constructing the event.
Social norms work in two directions. Headphones, reading, short answers and turning away can communicate a wish not to engage; respecting these signals protects public life. The same norms may make two willing people reluctant to begin. Recognising forecast error does not require abandoning the norm. It permits one small, easily declined invitation to test it.
Not every silence needs to be broken
Safety comes before an average well-being effect. A confined place, late hour, unequal power or physical vulnerability can change the meaning of an approach. People from groups exposed to harassment may choose not to respond on the basis of real experience. No experimental mean entitles a stranger to demand friendliness.
Cultural and local norms differ. In some settings, talking to strangers is ordinary courtesy. In others, not intruding is the courteous act. Conversation may carry higher cost for someone using a second language, someone with hearing difficulty or a person who is socially exhausted. Solitude can be a need, not a bias requiring correction.
A responsible opening is therefore easy to refuse. Comment on a shared setting or ask a non-private question, then attend to the response. A brief answer, replacement of headphones, absence of a return question or turning away is enough to end the attempt. Connection cannot be manufactured by ignoring a boundary.
The same principle applies to deeper disclosure. One person can reveal a little and leave room for the other to determine whether the exchange should continue. Consent in conversation is often incremental.
If place, role or power makes refusal difficult, apparent continuation does not establish equal willingness. Under uncertainty, the important question is not only whether the approach felt friendly to its initiator, but whether the other person retained a real and low-cost way to leave.
A relationship that never occurred can still be an outcome of prediction
The studies reveal more than the slogan “talking makes us happy”. Expectations change whether a person speaks. Speaking determines whether new evidence becomes available. Without interaction, the original pessimistic model is not corrected. Silence may represent both people’s authentic preferences, or it may become a stable equilibrium because each underestimated the other’s willingness.
There is a plain lesson for understanding others. When we imagine whether a stranger wants to talk, much of the available information comes from our own sense of risk. A low-cost, optional invitation obtains actual interpersonal evidence. The response belongs to a shared world rather than only to a private model.
The version 0.1 conclusion is that, in safe public settings where both parties can freely exit, people can be moderately sceptical of the prediction that talking to a stranger will necessarily be unpleasant. Many interactions in research were more enjoyable than expected and produced a brief sense of connection. The result neither requires a larger social quota nor licenses treating another person’s silence as an invitation to persist. Its best application is not forced extraversion. It is occasionally leaving room for a beginning to which the other person can comfortably say no.
Primary research sources
- Epley, N., & Schroeder, J. (2014), “Mistakenly Seeking Solitude”.
- Sandstrom, G. M., & Dunn, E. W. (2014), “Social Interactions and Well-Being”.
- Kardas, M., Kumar, A., & Epley, N. (2022), “Overly Shallow? Miscalibrated Expectations Create a Barrier to Deeper Conversation”.
- Epley, N., Kardas, M., Zhao, X., Atir, S., & Schroeder, J. (2022), “Undersociality”.
Series navigation: How Everyday Psychology Takes Shape — Season Two overview
Discover more from Geoffrey Chen
Subscribe to get the latest posts sent to your email.