The 36 Questions: How 45 Minutes of Escalating Talk Made Strangers Close

By Noam Shemla · The Studies

TL;DR This is the paper behind the famous “36 questions”. Pairs of strangers spent 45 minutes taking turns through questions that grew steadily more personal, and came out feeling far closer than pairs who spent the same 45 minutes on small talk. Two more studies then tested the obvious explanations and found that none of them mattered: not being matched on attitudes, not being told they would like each other, not even being told to get close. What mattered was the conversation.

• Paper: Aron, A., Melinat, E., Aron, E. N., Vallone, R. D., & Bator, R. J. (1997). The experimental generation of interpersonal closeness: A procedure and some preliminary findings. Personality and Social Psychology Bulletin, 23(4), 363–377.
• Setting: Psychology classes at a US university; students were paired with a classmate they did not know and did the tasks during a class session.
• Samples: Study 1: 50 pairs; Study 2: 71 pairs; Study 3: 69 pairs; plus a demonstration with 222 students at a second university.
• Procedure: Three sets of question slips, 15 minutes each, read aloud in turn and answered by both, growing more personal from set to set.
• Outcome: Closeness straight afterwards: the Inclusion of Other in the Self scale and a two-item closeness index, averaged.
• Design: Study 1 compared the closeness questions with matched small-talk questions; Studies 2 and 3 varied attitude matching, expected liking and whether closeness was an explicit goal.
• Headline: Escalating, reciprocal self-disclosure produced substantial closeness in 45 minutes; nothing else the authors varied did.

Friendships usually take months. In 1997 the psychologists Arthur and Elaine Aron and their colleagues published a way to produce something like one in under an hour, on purpose, in a classroom, so that closeness could finally be studied as an experiment rather than only observed.1 The appendix of that paper became one of the most widely shared pieces of psychology on the internet, as "the 36 questions that lead to love". The paper itself makes a narrower claim, and a more useful one.

The procedure

Two students who did not know each other were paired and handed an envelope of slips in three sets. They took turns reading a slip aloud; both answered it; then the other person read the next one. After 15 minutes the experimenter called time and they moved on to the next set.

The escalation is the design. The first set invites easy, pleasant self-disclosure (the kind of question you might ask at a dinner party, about ideal days and dinner guests). The second goes to values, memories and ambitions. The third asks for things people rarely say to a stranger: something embarrassing, a regret, what they have come to like about the person opposite. Some slips are tasks rather than questions, like finishing a sentence about the two of you.

The control condition had the same structure, the same timing and the same turn-taking, but its slips were small talk: a favourite holiday, a recent television show, what you did this summer. That is what makes the comparison fair. Both groups spent 45 minutes talking in turns with a stranger; only the content differed.

Study 1: disclosure against small talk

Closeness was measured the moment the conversation ended, with two instruments. The best known is the Inclusion of Other in the Self scale, which the Arons had developed a few years earlier: seven pairs of circles labelled self and other, from barely touching to almost completely overlapping, and the respondent picks the pair that describes the relationship.2 The second was a two-item index asking how close the relationship is compared with all of one's other relationships, and with other people's. The two correlated strongly (a median of .77 across the studies), so the authors averaged them.

The difference, 4.06 against 3.25, is an effect size of .88 standard deviations, which the authors note counts as large by the usual standard. It was significant even analysed conservatively, with each pair treated as a single data point. Nothing else in the study mattered: not whether the pair was mixed-sex or all-women, and not the attachment styles the researchers had matched them on.

How close is "about 4"? The authors offer a striking comparison. In a separate sample of 296 students at the same university, people used the same circles to describe their closest, deepest, most intimate relationship, and averaged 4.65.

The authors are careful about what that comparison means, and so should we be. People may use a scale differently in a lab than about a real relationship. But the second measure points the same way: asked to compare the new relationship with all their others, participants put it at about the midpoint, roughly as close as their average relationship.

Studies 2 and 3: the explanations that failed

The procedure had come with other ingredients, borrowed from what attraction research said should matter. The team tested them one at a time.

What the three studies varied, and what made a difference

These null results are the most interesting part of the paper. Similarity and the expectation of being liked are two of the best-established drivers of attraction, yet neither changed how close these pairs became. Neither did telling them that becoming close was the point. The authors' reading is that the disclosure process is powerful enough to swamp them, at least in this setting.

Study 3 held one quiet exception. When closeness was not mentioned as the goal, extraverted pairs ended up closer than introverted pairs; when it was the explicit task, the gap essentially disappeared. The authors float an interpretation: introverts are not worse at getting close, they are less comfortable with strangers until the situation gives them permission, and a structured task does exactly that. They label this finding tentative, and so do we.

Did it last?

The procedure was designed to produce closeness in the moment, not friendships. Still, in Study 3 the team could send a short questionnaire about seven weeks later, at the final exam. Of the 58 pairs who responded:

• 57% Had talked again. at least one later conversation
• 35% Did something together. outside the study
• 37% Sat together in class. in the weeks after

What 58 pairs from Study 3 reported about seven weeks later. There was no small-talk comparison at this stage, so the figures show that some closeness carried over, not how much more than chance.

Has it held up?

Unusually well, as these things go. The procedure, which came to be known as Fast Friends, has been used in a great many studies since, and it has been tested head to head again. In 2021 Susan Sprecher randomly assigned new pairs to the closeness questions, to the original small-talk questions, or to an unstructured getting-acquainted conversation; the closeness questions produced more closeness and other positive outcomes, especially compared with small talk.3

What this study cannot tell you

• Closeness was a feeling, measured at once. The authors say it themselves: the procedure is unlikely to create loyalty, commitment or dependence, which take time and shared history.
• The participants were students in a class, doing a structured exercise their instructor had endorsed, with a partner they knew they would see again.
• The later contact has no control group, so it shows carry-over, not an effect over and above small talk.
• It is not a love potion. The internet version promises romance. The study measured closeness between classmates, many in same-sex pairs.

What a speaker can take from it

The mechanism the study isolates is reciprocal, escalating self-disclosure: each person gives a little more, the other matches it, and the pace is set by structure rather than by nerve. That is not only a technique for first dates. It is what a good talk does to an audience: it moves from the safe to the personal at a pace the room can follow, and it gives something real before it asks for anything.

It also carries a warning for anyone who fills silence with small talk. In this study, 45 minutes of pleasant, well-structured small talk produced markedly less connection than 45 minutes of disclosure. Filling the time is not the same as using it.

This article summarises a published paper. Every number above is the authors’ own, from the results of Studies 1–3 and the general discussion. The scale is redrawn from Aron, Aron & Smollan (1992); the task examples are described in our words, not quoted.

Measured

• Closeness right after the conversation, on two established self-report scales
• Follow-up contact reported by the participants about seven weeks later (Study 3)

Inferred, not measured

• That disclosure, rather than time spent together, drives the effect rests on the comparison with matched small talk in Study 1
• The introversion finding in Study 3 is described by the authors as tentative

References

1. Aron, A., Melinat, E., Aron, E. N., Vallone, R. D., & Bator, R. J. (1997). The experimental generation of interpersonal closeness: a procedure and some preliminary findings. Personality and Social Psychology Bulletin, 23(4), 363–377. https://doi.org/10.1177/0146167297234003
2. Aron, A., Aron, E. N., & Smollan, D. (1992). Inclusion of Other in the Self Scale and the structure of interpersonal closeness. Journal of Personality and Social Psychology, 63(4), 596–612. https://doi.org/10.1037/0022-3514.63.4.596
3. Sprecher, S. (2021). Closeness and other affiliative outcomes generated from the Fast Friends procedure: a comparison with a small-talk task and unstructured self-disclosure and the moderating role of mode of communication. Journal of Social and Personal Relationships, 38(5), 1452–1471. https://doi.org/10.1177/0265407521996055