Becoming a close friend takes more than 200 hours together, according to University of Kansas research — and simply working beside someone does surprisingly little of the job. The hours that really predict closeness are the ones people voluntarily choose to spend together.

Two colleagues can sit within a few metres of each other for years and never become close. They know each other’s coffee order, attend the same meetings and exchange messages most days, yet the relationship stops at the office door.

Jeffrey Hall’s University of Kansas research offers a useful explanation. Time mattered for friendship, but hours imposed by a workplace or classroom carried less information than hours people chose to spend together.

In the Journal of Social and Personal Relationships, Hall estimated that good or best friendship usually required more than 200 hours together. This is one study, not settled consensus. Its numbers are modelled thresholds from particular samples, not a timetable that every relationship follows.

The 200 hours came from two different studies

Hall began with 355 American adults who had moved at least 50 miles within the previous six months. Their average age was 32.9, with participants ranging from 19 to 75. Each person named someone they had met since moving, estimated how many hours they had spent together and placed the relationship on a scale from acquaintance to best friend.

A second study followed 112 first-year students at a Midwestern public university. Each student identified two people they had recently met. They reported their cumulative time together and relationship status three, six and nine weeks into the semester, allowing Hall to observe how both changed.

Combining the two studies produced the ranges that travelled widely: around 40 to 60 hours to move from acquaintance to casual friend, 80 to 100 hours to become a friend, and more than 200 hours for a friend to become good or best friend. The University of Kansas summary published those rounded estimates in 2018.

The underlying figures were less tidy. In the adult study, the modelled chance of moving from friend to good or best friend passed 50 per cent at about 219 hours. In the early student data, the corresponding cut-off was 119 hours, while Hall’s synthesis argued that 200-plus hours over six weeks was a more likely requirement for the fuller transition.

Two hundred hours is therefore a useful scale, not a deadline.

Work introduced people, but did little of the later work

The paper did not find that workplaces are irrelevant to friendship. Among the relocated adults, 25.1 per cent had met the person they described at work, making it the second most common meeting place after the neighbourhood. A job can put potential friends repeatedly within reach.

Hall then separated what he called “closed system” relationships, such as coworkers and classmates, from relationships of choice. Total time predicted closeness, but the association between hours and closeness was stronger when people were choosing the relationship rather than merely sharing an institution.

The way time was spent also mattered. A higher proportion devoted to work or class was associated with less closeness. Hanging out, including eating or drinking together, predicted more closeness, as did watching films or television and playing games.

That negative work coefficient needs careful handling. Participants divided their shared time among six categories that had to add up to 100 per cent. A larger work share necessarily left a smaller share for leisure. The study did not show that an extra hour in the office makes two people less friendly; it showed that relationships dominated by obligatory tasks tended to remain less close.

Voluntary time contains information that proximity does not

Being scheduled into the same meeting says almost nothing about whether either person wants a relationship. Agreeing to eat together, visit someone’s home or keep talking after the task has ended is different. Both people are allocating time they could have used elsewhere.

Hall interpreted a change of context as part of friendship development. Someone first encountered as a coworker becomes a person another colleague chooses to see beyond the role. That choice does not guarantee closeness, but it makes the relationship less dependent on the organisation that introduced them.

More recent evidence points in a similar direction. In 2026, Benjamin Meagher reported three studies in the Journal of Environmental Psychology examining where friends interacted. More frequent contact in someone’s home was associated with greater closeness and reciprocal knowledge, even after accounting for contact in other settings. In a longitudinal student sample, weeks containing more residential interaction were also associated with higher closeness.

That paper cannot prove that homes make people close. Friends who already trust each other are more likely to be invited inside, and the frequency of virtual interaction was a strong predictor across its studies. The contribution is more modest: voluntarily changing the setting may tell us something about a developing relationship that raw physical proximity misses.

Conversation type mattered more than conversation share

Hall also tested a familiar assumption about talking. The overall proportion of time that participants spent in conversation was not associated with closeness in either study. Simply filling more minutes with words was not enough to distinguish the relationships that deepened.

Among the students, several kinds of exchange did predict change. Catching up, checking in, joking around and meaningful conversation were associated with increased closeness beyond the contribution of total hours and the relationship’s previous closeness. Small talk between weeks six and nine was associated with a decline.

This does not establish a conversational formula. The study did not assign people to joke, disclose or make small talk, and its categories came from participants’ own reports. A relationship that remains on safe topics may do so because it is already stalling. Small talk may also be useful early, before either person has earned the trust required for anything more personal.

The distinction is still useful. “Talking” bundles together exchanges with very different meanings. A project update can last an hour without revealing much about either person; a brief check-in can signal attention to a life outside the shared task.

Time and closeness probably reinforce each other

The nine-week student design improves on a one-off survey. Hall controlled for closeness at the previous wave, and additional hours still predicted later change. Students whose relationships moved from friend towards good or best friend added large amounts of time between surveys.

Causality can still run in both directions. People make more time for someone because the relationship is becoming rewarding. That time creates more shared experiences, which can make the relationship closer and justify another round of time. Relationships that cooled received less investment.

The adult study had its own limits. Participants reconstructed hours across weeks or months, which invites recall error. Time may have kept accumulating after the relationship had already crossed into a new category, inflating the cut-offs. The design also captured relationships that remained available to describe more readily than attempts at friendship that had already failed.

Friendship labels are subjective as well. One participant’s “good friend” may be another’s “friend”, and students living near each other can accumulate hours at a pace that working adults with families cannot. The difference between the adult and student thresholds shows how much context changes the estimate.

The workplace implication is narrower than a return-to-office slogan

The study offers no basis for claiming that colleagues cannot become close friends. Work was one of the main places adults in the sample met new people. It does place a limit on the idea that co-location itself creates meaningful connection.

Adding office days can increase encounters, shared references and the chance of an invitation. Those are conditions under which friendship may begin. The data suggest that the relationship develops when some contact becomes voluntary, and when two people keep allocating scarce unstructured time to each other.

That scarcity remains visible in newer data. The American Friendship Project’s 2022 and 2023 surveys found that roughly a third of adults agreed it was difficult to maintain friendships at their current stage of life, even though about three quarters were satisfied with the number of friends they had. Wanting more friends and having enough time for closer ones are different problems.

The 200-hour estimate helps explain why people can meet easily and still form few close relationships. The first encounter may be accidental or assigned. The hours after it are much harder to come by.