What the data actually shows
Chopik's 2017 research analysed friendship, family, and wellbeing across the lifespan using large international and U.S. samples. A consistent finding was that the value people placed on friendships predicted better health and happiness, and that in older adults this association was particularly strong — at times comparable to or stronger than the association with family relationships. The pattern held across hundreds of thousands of participants, which is part of why it drew attention.
A likely reason is the nature of the bond. Friendships are voluntary: they are kept up by mutual choice and tend to centre on shared activities and enjoyment, whereas family ties are obligatory and carry a wider mix of support, conflict, and duty. Both forms of support matter, but the chosen, lighter-friction quality of friendship may make it especially protective for everyday mood, and the friendships that survive into later life are often the ones that were genuinely rewarding.
Zoom out and the more reliable conclusion is about relationship quality in general. The Harvard Study of Adult Development — one of the longest-running studies of adult life, led now by Robert Waldinger and Marc Schulz — found across more than eight decades that the warmth and quality of people's relationships was among the strongest predictors of long-term health and happiness. Crucially, their finding is about relationships broadly, not about whether friends or family is the 'right' category to invest in.
Why this feels different from how it actually is
It can feel like family should obviously matter most, because that is the cultural default and the relationships we are born into. So the finding that friendships carry comparable or even greater weight for some outcomes can be surprising. The research is not saying family is unimportant — it is showing that the chosen, enjoyable quality of friendship does real and sometimes underappreciated work for wellbeing.
It also feels different because friendships are easy to deprioritise. Family and partners come with built-in obligations and routines; friendships usually do not, so they are the first thing to quietly fall away when life gets busy. People often discover the cost only later, which fits the research showing friendships becoming especially important to happiness in the second half of life — sometimes after they have already thinned out.
And because the answer is 'it depends,' it resists the tidy ranking people want. The honest version — that both matter, that quality matters more than category, and that friendship is underrated relative to its weight — is less satisfying than 'family first' or 'friends are everything,' but it is what the data supports.
What the research says to do about it
The most defensible takeaway is to treat relationship quality as the thing worth protecting, rather than agonising over whether friends or family should come first. The Harvard study's long-run finding points at warmth and closeness across your relationships, not at picking a category. Investing in the ties that actually feel good and supportive — wherever they sit — is what the evidence backs.
Given how easily friendships erode, the research is a fair argument for tending them deliberately. Because they lack the built-in obligations of family, friendships need active maintenance — reaching out, showing up, keeping them alive through the busy decades. Chopik's work suggests this is not a soft nicety but plausibly linked to health and happiness later on.
It is also worth resisting the idea that one strong relationship should carry everything. A varied web of connection — partner, family, and friends — spreads the load and offers different kinds of support. None of this is a verdict on anyone's particular mix of relationships; it is simply where the broad research points.
What the research says does not help
Treating the question as a strict ranking does not help, because the data does not support one. Deciding 'family always matters most' or 'friends are all that count' oversimplifies a picture where both matter and the quality of relationships overall does the heavy lifting. The category is less predictive than the closeness within it.
Maximising the number of relationships does not reliably help either. The research points to the quality and felt warmth of connection, not the size of your social network. A large circle of shallow ties is not what predicts wellbeing in this work; a few genuinely close, rewarding relationships does more.
Letting friendships lapse on the assumption that family alone will cover your social needs is a common pattern the research gently cautions against. Because friendships fade without active upkeep and appear to matter more, not less, later in life, neglecting them tends to cost more than it seems to at the time.
Real numbers in context
The standout finding is qualitative rather than a single statistic: across very large samples, Chopik (2017) found that valuing friendships predicted better health and happiness in older adults, sometimes rivalling or exceeding family relationships. The likely mechanism is that friendships are chosen and built around shared enjoyment. Treat this as a robust pattern across hundreds of thousands of people rather than a precise effect size.
Put in context, the Harvard Study of Adult Development — running for over 80 years — found that the quality of people's relationships overall was among the strongest predictors of long-term health and happiness, ahead of many things people assume matter more, like wealth or fame. So the most reliable answer is not 'friends' or 'family' but 'close, warm relationships' — with friendship being the underrated member of that pair, especially as we age.