Most of us know the feeling of having one friend who makes us laugh and another we call when something needs thinking through. A new diary study set out to test whether people really do choose particular friends for particular emotional purposes. It found limited, test-dependent evidence that they do, and no sign that choosing a certain friend led to better outcomes.
The FRIENDS study, led by Schott, Enser and Weiß, appeared in BMC Psychology on 14 October 2026. It was preregistered, which means the plan was set out before the data were analysed. Participants answered short prompts on four occasions each day for 14 days, with data collected between 6 June and 18 July 2025.
Of 124 people recruited in Germany, 106 adults were kept for analysis. Their average age was 24.72 years. The group included 80 women, 24 men, one nonbinary person and one person who preferred not to say, and the majority were young students or had finished high school.
Of the 5,936 prompts scheduled, 5,049 were answered, a response rate of 85.1 per cent. Once checks were done, 5,042 valid prompts were analysed, along with 2,796 reports of contact with a friend. About 70.5 per cent of contacts were in person and 29.5 per cent were virtual.
Each time, people gave a reason for the contact. The most common were distraction (34.1 per cent) and humour (27.4 per cent), next came problem solving at 17.5 per cent, then reappraisal at 9.1, encouragement at 8.4 and touch at 3.6. The total is affected by rounding. On average, participants reported a network of 8.90 friends, and 5.67 close friends on the baseline measure.
The question was whether people paired particular friends with particular motives more often than chance would predict. The answer depended on the method. An initial benchmark overstated how much of this selection was going on. Under permutation tests, no single motive remained significant once the analysis corrected for running many tests at once. A second set of tests, based on residualised scores, supported only two motives: problem solving and distraction.
An exploratory index of how specialised each person's network was came out higher than chance on average. Yet when each person was looked at alone, just 18 of 104 participants (17.3 per cent) exceeded chance. An average across the group does not mean that most people had a clearly specialised set of friends.
The second question was whether such pairing made a difference. The researchers did not find that pairing a friend with a motive significantly predicted how well people regulated their emotions, how they rated the interaction, their wellbeing or their social loneliness. For interaction rating the estimate was b = -0.02 (95 per cent confidence interval -0.07 to 0.02, p = .305), and for effectiveness b = -0.03 (p = .383). In plain terms, no statistically significant link was found.
That absence of a link is the core of the story. Even where people seemed to turn to particular friends for particular purposes, nothing in the study showed that this pairing led to a better interaction or a better emotional outcome. It also did not test the idea experimentally: nobody was assigned a friend or told whom to contact.
The study captured the benefit people hoped for and the motive they picked, but not what they truly needed at the time or how the friend responded. The forced-choice list of motives, and the way separate contacts were combined, limit how much can be said about the complexity of real conversations.
The participants were young and lived in Germany. Older adults, people in other countries, or people with different friendship habits may look quite different, and the study cannot say.
Nothing here suggests that having a "go-to" friend for each need is a mistake, or that it is a requirement. If you notice that you tend to call one friend for distraction and another to talk a problem through, that is only an example to reflect on. It is not a participant's story and not something this research established for you.
Treat the idea as a soft prompt for thinking, not a rule. Do different people come to mind when you want a laugh, some practical thinking or simple company? Do you get enough of the contact you value? Friends need no labels for their role, and nobody has to react in a set way. One easy step is to get in touch with someone you trust for the sort of chat you feel like having, and see where it goes.
Source: FRIENDS - friendship-based regulation of individual emotions in networks in daily situations: An EMA study on partner-motive pairing and social outcomes in adult friendships DOI: 10.1186/s40359-026-05767-w
Illustration of different forms of emotional support within friendships. AI-generated editorial illustration. This article summarises published research in our own words and is general information, not personal health advice.
Please Login or Sign up to be able to comment