Can the seat next to you change your friendships? At a university in Germany, new students got their seats by lottery at their first meeting. One year later, researchers asked who their friends were.

Students who sat next to each other by chance were more likely to be friends. Students in the same row were, too. So sitting close to someone can be the start of a friendship.

Friendship research has a common problem. Do people become friends because they sit close together, or do friends choose to sit together? A German research team solved this problem with a lottery. At the first introductory session for new university students, seats were randomly assigned, and the students rated their first impressions of one another.

One year later, the researchers checked who had become friends. Students who had happened to sit next to each other had stronger friendships than those who sat farther apart, and the same was true for students in the same row. A similar pattern was found in the 1970s at an American police academy, where rooms and seats were assigned in alphabetical order.

A persistent difficulty in friendship research is separating cause from effect: proximity may encourage friendship, but friends may also choose to stay close to each other. A German research team addressed this problem with a simple yet powerful design. At the first introductory session for new university students, seats were assigned by lottery, and students rated their initial impressions of one another. One year later, the researchers assessed who had become friends.

Students who had been randomly seated next to each other had stronger friendships than those seated farther apart, and a similar effect appeared for students in the same row. Because seating was random, pre-existing similarity cannot explain the result, which suggests that a single chance encounter can shape relationships a year later. This finding echoes a 1970s study at a US state police academy, where rooms and seats were allocated alphabetically and trainees' closest friends tended to have surnames near their own.

However, the German study involved a small sample from a single university, so its generalizability is limited. Proximity probably works by creating opportunities to talk, increasing time spent together, and building familiarity.