How did a methodological appendix in a 1997 social psychology paper become the internet's favorite love potion?
The Aron et al. 1997 study did not demonstrate that 36 questions can make two strangers fall in love. Hear me out. The paper demonstrated something narrower, more technical, and — once you actually read it — more interesting than the viral headline suggests. To see why, we have to walk the timeline backward from the 2015 moment that locked the misreading in place, and forward from the 1992 instrument that made the 1997 study possible in the first place.
October 1992: The IOS Scale That Made the 1997 Procedure Measurable
Before there could be a closeness-generating procedure, there had to be a way to measure closeness. Five years before the famous 36-questions paper, Arthur Aron, Elaine N. Aron, and Danny Smollan published "Inclusion of Other in the Self Scale and the Structure of Interpersonal Closeness" in the Journal of Personality and Social Psychology (Vol. 63, No. 4, pp. 596-612). The paper introduced a single-item pictorial measure: seven pairs of overlapping circles, progressing from barely touching to almost fully merged. Respondents chose the pair that best described their relationship with a target person.
This sounds trivial. It is not. Across the studies reported in the 1992 paper, the IOS showed test-retest stability and correlations with longer closeness inventories that were high enough to make social psychologists comfortable using a single-item tool in settings where longer batteries were impractical — settings like a 45-minute lab interaction ending with a post-test questionnaire.
Everything about the 1997 procedure assumes IOS as the dependent variable. Without IOS, the 36-questions study could not have been what it was: a controlled manipulation with a quick, validated closeness measure administered right after a short lab session. The 1992 paper is load-bearing infrastructure. We are writing this piece because almost no popular account of the 36 questions mentions it, and the silence around the measurement instrument is part of why the study's scope gets stretched.
April 1997: The Paper Actually Gets Published
The publication itself: Arthur Aron, Edward Melinat, Elaine N. Aron, Robert Darrin Vallone, and Renee J. Bator — "The Experimental Generation of Interpersonal Closeness: A Procedure and Some Preliminary Findings" — Personality and Social Psychology Bulletin, Vol. 23, No. 4, pp. 363-377, April 1997.
First, the arithmetic the viral reading always skips. The paper reports three studies, not one. Study 1 compared a closeness-generating task against a small-talk control; later studies layered on similarity and expected-liking manipulations. The closeness task itself is the part everyone quotes: 36 questions in three sets of 12 questions each, 15 minutes per set, 45 minutes total. If you assume rough parity of speaking time within a pair, each participant gets about 22.5 minutes of airtime, which works out to roughly 37.5 seconds per question. That number matters. It tells you the procedure is not a deep-dive interview — it is a rapid, structured exchange where neither person has time to rehearse polished self-presentation. And the reported effect in the paper is generated against a small-talk control running the same 45 minutes, which rules out the simple "they just spent longer together" confound.
Post-test, pairs in the closeness condition reported significantly higher IOS scores than pairs in the small-talk condition. The authors were explicit in the discussion: the procedure generates experimentally-induced feelings of closeness comparable to what participants said they felt toward their closest other relationships. They did not claim it generated romantic love. The word "love" appears sparingly in the paper. The claim is about interpersonal closeness as operationalized by the IOS.
July 2005: Aron's Neuroimaging Work Reframes the Lab
Between 1997 and 2015, Aron's lab did not spend most of its time on the 36 questions. It moved into functional neuroimaging. In July 2005, Arthur Aron, Helen Fisher, Debra Mashek, Greg Strong, Haifang Li, and Lucy Brown published "Reward, Motivation, and Emotion Systems Associated with Early-Stage Intense Romantic Love: An fMRI Study" in the Journal of Neurophysiology (Vol. 94, No. 1, pp. 327-337).
The sample was small — 17 participants who reported being intensely in love — and the study showed activation in dopamine-rich areas (the ventral tegmental area and parts of the caudate) when participants viewed photographs of their partner versus a neutral acquaintance. Effect sizes in small-n fMRI work always deserve skeptical reading, and the paper itself is careful about what functional imaging can and cannot tell us about subjective love experiences. But this is the paper that made "Aron" a name outside social psychology, and it is the paper that retrofitted the 1997 closeness procedure into the public imagination as part of a "love scientist" body of work.
The conflation matters for what happened a decade later. By the time a journalist ran across the 36 questions, Aron's lab was already publicly associated with brain-level research on romantic love. The 1997 paper was no longer sitting in a social psychology journal neighborhood where its scope was obvious on contact. It was hovering in a vague pop-science aura of dopamine stories and brain scans, one step removed from the methods section that would have kept it honest.
January 2015: A Modern Love Column Rewrites the Headline
On January 9, 2015, the New York Times Modern Love column published "To Fall in Love With Anyone, Do This," by essayist Mandy Len Catron. Catron described going through the 36 questions with a university acquaintance at a bar, followed by four minutes of silent mutual gaze — a procedural variant Aron's lab had used in some follow-up work — and then, some months later, ending up in a relationship with him.
The column became one of the most-read Modern Love installments ever published. It was shared, quoted, and headlined with the framing that did the actual damage: "to fall in love with anyone, do this." Catron herself was more careful inside the piece than the headline suggests. She noted that the original study was about generating closeness, not romance. She acknowledged that she and the acquaintance already had an element of mutual interest before starting. She treated the procedure as an essayist's experiment on herself, not a repeatable recipe.
None of that survived the virality. Within weeks, "the 36 questions to fall in love" was the dominant framing. Decks of conversation cards appeared. Apps followed. Wedding toasts referenced them. The 1997 paper — a controlled manipulation comparing a self-disclosure condition against a small-talk condition, using a single-item pictorial closeness scale as the dependent variable — was now folded into a cultural object whose headline promise the authors had never made in print. This is the concession-then-teardown the piece has been building to: the column was not dishonest. It was simply read by more people than the paper ever would be, and they read the headline and not the methods.
June 2017: The Essayist at the Center of It Publishes the Book
Mandy Len Catron's memoir, How to Fall in Love With Anyone: A Memoir in Essays, was published by Simon & Schuster in June 2017, roughly two and a half years after the Modern Love column ran. The book is more interesting than its title lets on. Catron is an English instructor at the University of British Columbia, not a psychologist, and she spends much of the collection interrogating the "falling in love" framing her own viral essay created. She writes about what the procedure did and did not do, about the asymmetry between the headline and the study, and about the discomfort of being read as a participant in a science experiment when what she had actually done was a literary experiment on herself.
None of that was the part that traveled. What traveled was the 2015 framing, now calcified by book publication into the dominant cultural reference for the 1997 paper. By mid-2017, the 36 questions had appeared in apps, decks of cards, relationship workbooks, and casual references across dating media. Arthur Aron gave interviews in which he patiently explained that the procedure was designed to generate closeness in lab conditions, that the 1997 paper had never claimed to produce romantic love, and that the effect was real but bounded. The interviews were generally read as qualifying the procedure rather than as rewriting the framing. By the late 2010s, the canonical reference for "the 36 questions" was the essay and the book, not the PSPB paper. The primary source had become a supporting citation in someone else's story.
What It All Means
The 1997 paper did what it said it did. It ran a controlled comparison between a self-disclosure task and a small-talk task, measured post-interaction closeness with a validated instrument, and found a statistically significant effect favoring the self-disclosure condition. That finding is, within the limits of lab social psychology, replicable and useful. It tells us that structured reciprocal self-disclosure under instructions increases reported closeness between strangers compared to unstructured small talk. The effect is not a love spell. It is a clean demonstration that if you put two cooperative people in a room and give them a protocol for vulnerability, they will report feeling closer than if you give them a protocol for chatting about the weather.
The gap between that finding and the viral framing is not a gap of degree. It is a category error. "Closeness on the IOS" and "romantic love" are not points on the same continuum. The IOS is a single-item pictorial measure of perceived relationship overlap. Romantic love, however you operationalize it, involves pair-bond formation, sexual interest, jealousy circuitry, and — per the same lab's 2005 fMRI work — a distinctive neural signature. One can exist without the other in both directions. You can feel close to a colleague you will never date. You can romantically obsess over a person you barely know. The 1997 procedure manipulates the first variable. The Modern Love headline promised it manipulated the second.
Reading the 1997 paper today, the most striking thing is how modest its claims actually are. The authors anticipate the virality by accident: they note in the discussion that the procedure is not a substitute for the processes by which long-term relationships form, and they position it as a method — a tool other researchers can use to rapidly generate moderate levels of interpersonal closeness for subsequent experimental manipulations. That is what it is. A closeness-induction tool for lab work. The reason it became a cultural object is not that the paper overclaimed; it is that a good essay, a catchy headline, and a few hundred thousand shares will overwhelm the plain text of any methods section.
This piece does not cover the replication literature on the 36-questions procedure in any detail — the meta-analytic picture, where it exists, deserves its own treatment against the 1997 effect sizes. It does not address whether the four-minute mutual gaze component that Aron's lab used in some follow-up variants carries a meaningful effect independent of the question set. And it does not address whether the procedure is culturally transportable outside the WEIRD undergraduate samples it was developed in, which is the question most popular app versions sidestep entirely. Each of those is a separate argument, and each deserves its own careful reading of the primary sources rather than another recycled headline.