Empathy is often described as if it were a fixed possession: some people have it, others do not. A 25-year study from the University of Virginia offers a more relational picture. Care received in one relationship may become practice for care given in the next.
The study followed 184 people from age 13 into their parenting years. Mothers who showed more empathic care toward their teenagers had teenagers who later showed more empathic care toward close friends. Empathy in late-adolescent friendship predicted more supportive parenting in adulthood, and that parenting was associated with more empathy in the participants’ young children.
It resembles a relay across three generations, but the baton is not a gene or a single lesson. It is a pattern of attention, understanding and support repeatedly experienced in relationships.
Researchers watched empathy rather than merely asking about it
The project began in 1998 with 184 seventh- and eighth-grade students recruited from a public middle school in the mid-Atlantic United States. The sample included 99 girls and 85 boys; 58 percent identified as White, 29 percent as African American, 8 percent as mixed race or ethnicity and 5 percent as belonging to other groups. The researchers also accounted statistically for gender and family income.
When each participant was 13, researchers filmed the teenager and their mother discussing a problem for which the teenager wanted help or advice. Trained observers coded whether the mother stayed emotionally engaged, understood the problem accurately, validated the teenager’s feelings and offered useful support.
Then, every year from ages 13 to 19, the teenager came to the laboratory with a nominated closest friend. This time the friend described a problem and the researchers rated the teenager’s empathic response. As lead author Jessica Stern explained in the University of Virginia’s account of the research, coders looked for visible engagement, interest, emotional understanding and support.
Mothers’ care predicted how teenagers treated friends
The first link in the chain was clear: more empathic maternal support at age 13 predicted greater empathic care for close friends across adolescence. Teenagers’ empathy rose modestly on average between 13 and 19, while individuals differed meaningfully from one another.
This result extends an earlier finding from the same long-running project. In previous work, teenagers who reported more secure family relationships showed more empathic support toward friends. That earlier analysis followed the same 184-person cohort, but addressed family security rather than the three-generation pathway in the newer paper.
Friendship matters because it changes the direction of care. A parent-child relationship is unequal: the adult normally holds more responsibility and authority. A close adolescent friendship is more horizontal. Each person may need to give and receive support, notice when advice is welcome, and remain emotionally regulated while someone else is distressed.
The authors propose that friendship may therefore serve as a practice ground for caregiving. This fits broader developmental work treating adolescence as a period shaped jointly by parents and peers, not as a clean transfer of influence from one group to the other. A review of adolescent-parent relationships likewise describes family and peer development as linked processes.
Friendship empathy predicted parenting a decade later
When participants were between 22 and 37 and had children of their own, researchers asked how they would respond to common situations in which a child was upset. Supportive options included helping solve the problem, helping the child feel better and encouraging the child to express emotion. Less supportive options included punishment, minimising the emotion or becoming personally overwhelmed.
Participants who had shown more empathy toward close friends at ages 17 to 19 reported more supportive responses to their children’s distress in adulthood. The association remained after the model accounted for the original mother’s empathy, the participant’s gender and family income. Adolescent friendship empathy did not significantly predict the separate measure of non-supportive parenting.
This was the study’s most novel bridge. In an interview published by the Society for Research in Child Development, Stern described the implication as “paying forward” experiences of being understood and supported. The finding does not imply that close friends replace parents. It suggests that friendship can be one place where teenagers rehearse the emotional work later required by caregiving.
The third generation came with important limits
The children’s part of the analysis included 114 children aged 3 to 8 from 70 of the original participants’ families. Parents rated how often each child tried to understand how others felt, showed concern or tried to comfort another person. Across families, adults reporting more supportive responses to distress also reported greater empathy in their children.
The indirect pathway from observed friendship empathy in late adolescence to child empathy through supportive parenting was statistically significant, but only just: the estimate was 1.39, with a 95 percent confidence interval from 0.01 to 2.77 and p = .048. The authors explicitly described this third-generation test as exploratory.
That caution matters. The first two links used recorded conversations scored by observers. The parenting measure and the children’s empathy measure were both reports from the same adult participant. Parents who see themselves as supportive may also be inclined to see their children as empathic. The parenting and child measures were collected at the same period, so they cannot establish which came first.
The sample also narrowed over time, came from one US community and began with mothers because comparable father data were unavailable at age 13. Adult mothers and fathers were represented later, and their reported responses to child distress did not differ significantly, but that does not fill the missing first-generation evidence about fathers.
A relay is an association, not proof of causation
This was not an experiment. Researchers did not randomly assign some parents to behave empathically and compare their descendants with a control group. Temperament, culture, stress, relationship quality and other unmeasured factors may contribute to several links at once. The results show a prospective chain of prediction, not proof that one mother’s response caused a grandchild’s empathy decades later.
Nor did all 184 people travel through every stage. Only participants who became parents of children in the required age range could contribute to the last analysis. Attrition and family timing may make that smaller group different from the full cohort.
Still, the study’s design is unusually valuable. It began before the participants became parents, observed real interactions repeatedly through adolescence and connected those observations to behaviour reported much later. Most claims about intergenerational patterns rely on adults remembering their childhoods after the fact. Here, the earlier relationships were recorded when they were happening.
The practical lesson is smaller and more hopeful
Empathy in this research did not mean agreeing with everything a teenager said, erasing every difficult feeling or instantly solving the problem. It looked like staying present, understanding accurately, validating emotion and offering support suited to what the other person needed.
Those are behaviours, which means they can be modelled and practised. A parent can pause before correcting, reflect what they think the teenager is feeling and ask whether the teenager wants advice, practical help or simply company. A teenager can learn the same sequence with a friend. No single conversation determines a family line, but repeated conversations can create a familiar way of responding.
The finding belongs beside the broader evidence that close relationships shape development and health across the life course, including the long-running work discussed in our guide to the Harvard Study of Adult Development. Relationships are not only outcomes we enjoy after becoming emotionally skilled. They are among the places where emotional skill is built.
The most careful interpretation of the relay is therefore not that empathy passes automatically from grandmother to grandchild. It is that supportive relationships may give people both a model and a rehearsal space. Care received from a mother can appear again between friends; care practised between friends can shape later parenting; and supportive parenting may give the next child another place to begin.