Self-disclosure and the thirty-six questions protocol

Arthur Aron and colleagues published in 1997 a procedure in which two strangers converse for forty-five minutes answering a series of increasingly personal questions, arranged in three sets.

The comparison group converses for the same time about ordinary topics. The main measure is reported closeness at the end of the session, collected through scales on which each participant places the relationship that has just begun.

The design belongs to an older tradition. Sidney Jourard had made self-disclosure an object of study as early as the 1950s and 1960s, and the social penetration theory of Irwin Altman and Dalmas Taylor, in 1973, already described the gradual deepening of exchanges.

Reported closeness is markedly higher in the personal questions group. The result has been replicated and underpins much research on rapid tie formation.

The study's contribution is methodological: it provides a way of producing a state of closeness in the laboratory, which then allows its effects on other measures to be studied. That is what researchers had lacked, having until then been obliged to study relationships already formed, whose history escaped them.

It does not show that a lasting relationship forms. The measure is taken immediately afterwards, and follow-ups at a distance are rare and inconclusive. Closeness felt at the end of a session and a friendship built over years are two distinct objects.

Media coverage, notably a widely read 2015 article on the romantic use of the procedure, went far beyond what the data support, as the authors themselves have pointed out.

Work on self-disclosure regularly finds that one-sided disclosure does not produce the same effect. It is the alternation that is associated with reported closeness. The 1994 meta-analysis by Nancy Collins and Lynn Miller already showed that the link between disclosure and liking runs in both directions.

Disclosure that is too early or too intense, without reciprocity, is associated in several studies with a less favourable evaluation, which contradicts the idea that saying more would suffice. The norms governing that pace also vary across social milieus and cultures.

It is used in research on prejudice reduction, where members of groups in tension are brought into conversation. The measured results bear on short-term reported attitudes, and sometimes on physiological measures collected during subsequent sessions.

Here too, there is a large gap between the established effect, measured after the session, and the social use made of it, which presupposes a lasting effect on real relationships. A well-established laboratory protocol should not have to carry promises it never made.