Who Do You Call When You Need Comfort? New Study Tests Whether Friends Have Emotional Specialties

When life delivers a setback, most people do not reach for a random friend. They reach for a particular one: the person who always talks them down from anxiety, the one who can turn a bad mood around with a joke, or the level-headed confidant who helps them plot a practical way forward. Psychologists have long suspected that adults organize their friendships into something like an emotional toolbox, assigning different regulation partners to different needs. A new study set out to test that intuition with unusual rigor, and its results are a striking lesson in how easily a compelling psychological story can dissolve under careful statistical scrutiny.

The research, published in BMC Psychology by Paula Schott, Katharina Enser, and Martin Weiß of the University of Würzburg, examined what the authors call friendship-based regulation of individual emotions in networks, or FRIENDS. The central question was twofold. First, do people systematically pair specific friends with specific regulatory motives, for example turning to one friend when they want distraction and to another when they want a problem solved? Second, if such pairing exists, does it actually pay off, in the form of more effective emotion regulation, better evaluations of social interactions, higher psychological well-being, and lower social loneliness?

Earlier work had suggested that individuals select regulation partners according to how competent they perceive each person to be for a particular emotion. The Würzburg team wanted to push that idea one step further, from emotions to motives: the specific benefit a person hopes to obtain from a contact, such as soothing, distraction, validation, or problem solving. If functional specialization really is a feature of adult friendship networks, it should show up not just in who people contact but in the fine-grained structure of why they contact them.

To capture this structure in everyday life rather than in the artificial setting of a laboratory, the researchers ran a preregistered 14-day ecological momentary assessment study, a method that samples experience in real time as it unfolds. A total of 106 adults received four semi-random prompts per day, and whenever a prompt arrived they reported on a recent friend contact. Across the two weeks this produced 2,796 contact reports, each recording which friend was involved, which of six affect-improving regulation strategies the participant had hoped would benefit them, how much the contact had actually helped, and how positively they evaluated the interaction. One important limitation is worth noting: the study did not assess which strategy the friend actually used, only what the participant had sought and experienced.

Quantifying specialization required careful measurement. For each participant, the researchers computed a Specialization Score capturing how strongly particular friends were paired with particular motives, and, for interaction outcomes, a leave-one-out residual that separated motive-specific pairing from a partner’s overall involvement. This distinction matters enormously. If someone contacts their best friend for everything, that friend will appear paired with every motive simply because they dominate the contact record. A residualized metric was needed to ask whether, beyond sheer frequency, certain motives were genuinely channeled toward certain people. At the network level, the team used a Network Specialization Index, or NSI, to summarize specialization across each participant’s entire friendship web.

The statistical approach was deliberately conservative. Because partner-motive pairing could arise by chance in any individual’s data, the researchers tested their observed scores against participant-level permutation nulls, essentially asking whether the real pattern was more specialized than random reshufflings of the same contacts. The results were sobering. Under the permutation null, no regulatory motive survived correction for multiple testing. The residualized metric, which is more sensitive, did indicate motive-specific channeling for exactly two motives: problem solving and distraction. In other words, people did appear to route their desire for practical advice and their desire for diversion toward particular friends, but the evidence for the remaining motives did not hold up.

The network-level picture was similarly nuanced. When the NSI was standardized against each participant’s own permutation null, the average effect exceeded chance, corresponding to a standardized effect size of d = 0.36, a small-to-moderate difference. Yet that average concealed enormous variability: only 18 of 104 participants showed a network-level specialization that reached statistical significance. Specialization, where it exists at all, appears to be a property of some people’s networks rather than a universal architecture of adult friendship. Much of what looks like pairing in raw data may simply reflect the concentration of contact on a few favored partners.

Perhaps the most consequential finding is what did not appear at all. Neither the leave-one-out pairing predictor nor the network-level indices were associated with perceived emotion regulation effectiveness, with how positively participants evaluated their interactions, with psychological well-being as measured by the German Flourishing Scale, or with social loneliness as measured by the Social and Emotional Loneliness Scale for Adults. The null results proved robust across alternative metrics, alternative indices, ordinal models, and different random-effects structures. Even where participants showed detectable motive-specific pairing, that structure bought them nothing measurable in terms of feeling better or being less lonely.

The authors draw a pointed conclusion from these results: the findings caution against inferring functional specialization from partner-selection data alone. This warning has real methodological weight. Much of what we believe about social support rests on self-reported patterns of who people turn to and why, and this study demonstrates that such patterns can be dominated by a few highly active relationships, that apparent specialization may be small once overall involvement is accounted for, and that network indices derived from contact reports are not the same thing as the subjective experience of support. A friend who is contacted for every purpose is statistically indistinguishable from a team of specialists unless the analysis separates frequency from specificity, as this one did.

For the broader science of interpersonal emotion regulation, the study is both a refinement and a corrective. It confirms that people do, in aggregate, show some tendency to channel problem solving and distraction toward particular friends, and that network-level specialization is detectable on average when tested against individualized chance baselines. But it also shows that this structure is modest, unevenly distributed across individuals, and, crucially, unrelated to the outcomes that matter, from daily interaction quality to long-term well-being and loneliness. The intuitive image of the friendship toolbox, with one friend for comfort and another for advice, survives only in weakened form. What seems to matter more may simply be having people to turn to at all, rather than matching the right person to the right emotional job. Future work, the authors suggest, will need to look beyond selection patterns, perhaps by assessing what partners actually do during regulation attempts, to understand when turning to friends genuinely helps. For now, the study stands as an elegant demonstration that in the psychology of friendship, as elsewhere, a good story told by the data is not always a story the data can actually support.

Subject of Research: Interpersonal emotion regulation and partner-motive pairing in adult friendship networks studied with ecological momentary assessment

Article Title: 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

Article References: 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. (n.d.). https://doi.org/10.1186/s40359-026-05767-w

Image Credits: AI Generated

DOI: 10.1186/s40359-026-05767-w

Keywords: interpersonal emotion regulation, friendship networks, ecological momentary assessment, social loneliness, psychological well-being, partner selection, emotion regulation strategies, social support, EMA study, BMC Psychology, permutation testing, adult friendships