RESEARCH_FILE

The Science of One-Sided Friendships

If you've ever suspected a friendship only exists because you keep it alive, the data has news — and it's not just about you.

SEE THE PRACTICE

Turn a thought this research explains into one clear move.

THE THOUGHT

“I'm always the one who texts first”

YOUR RECORDED RESPONSE

“I can notice the pattern without turning it into a bill.”

ONE PRIVATE MOVE

Open your recent messages and note, on paper or in your notes app, the last three times you started a conversation. Next to each one, write only what you actually wanted in that moment: contact, clarity, or reassurance.

When researchers actually mapped who calls whom a friend, only about half of the ties people were sure about turned out to be mutual, even though nearly everyone expected reciprocity. At the same time, decades of work on equity, tie decay, and the liking gap show that some of what feels one-sided is real imbalance — and some is a systematic misread of the other side. Here's how the science learned to tell the difference.

~53%of self-reported friendship ties in the MIT study were actually reciprocated — even though participants expected nearly all of them to be mutual 34–53%range of friendship reciprocity rates across the earlier studies reviewed in the same paper — no study found perceived friendship lists matching reality 200+ hrsof shared time it takes to turn someone into a close friend, per Hall's time-investment study — with hours at work counting far less than leisure time 18 monthswas enough for friendships without contact to measurably lose emotional closeness in Roberts and Dunbar's longitudinal study — while family ties held stable through the same silence

How the science changed

1978
Elaine Hatfield (then Walster) and colleagues formalize equity theory for close relationships: people track give-and-take even with friends and lovers, under-benefited partners feel anger and resentment, over-benefited partners feel guilt — and inequitable relationships drift toward repair or dissolution. ↗
2011
Roberts and Dunbar follow students for 18 months across the school-to-university transition: friendships that stop receiving contact decline measurably in emotional closeness, while family ties stay stable through the same gaps — friendship, unlike kinship, must be actively maintained. ↗
2013
Ball and Newman analyze directed friendship networks in US schools: unreciprocated friendships are not random noise — almost all of them run from a lower-ranked person claiming friendship with a higher-ranked one, suggesting one-sided ties encode social status, not personal failure. ↗
2016
Almaatouq, Radaelli, Pentland and Shmueli publish the MIT reciprocity study in PLOS ONE: participants expected the vast majority of their friendships to be mutual, but only about 53% actually were — consistent with earlier studies finding reciprocity rates of 34–53% — and only reciprocal ties effectively promoted behavior change. ↗
2018
Boothby, Cooney, Sandstrom and Clark document the 'liking gap' in Psychological Science: across lab conversations, workshops, and dorm roommates tracked over an academic year, people systematically underestimated how much their conversation partners liked them. ↗
2019
Jeffrey Hall quantifies the cost of friendship in the Journal of Social and Personal Relationships: it takes roughly 50 hours of shared time to move from acquaintance to casual friend, about 90 to become a friend, and more than 200 to become a close friend — and hours spent together at work count far less than leisure time. ↗

What people believe vs. what the data shows

The belief“If I feel close to them, they must feel the same about me.”

The dataIn the MIT reciprocity study, participants expected the vast majority of their friendships to be mutual — but only about 53% actually were. Misjudging who counts you back is the statistical norm, not a personal blind spot unique to you. ↗

The belief“Keeping score in a friendship is petty — real friends don't track who gives more.”

The dataEquity research shows people track give-and-take in close relationships whether they admit it or not: chronic under-benefit reliably produces anger and resentment, over-benefit produces guilt, and inequitable relationships tend toward repair or dissolution. Noticing the imbalance isn't pettiness — it's the documented signal. ↗

The belief“Real friends stay close no matter how little you talk — needing contact means it was never solid.”

The dataIn an 18-month longitudinal study, friendships that stopped receiving contact declined measurably in emotional closeness, while family ties survived the same gaps intact. Friendship runs on active investment by design — decay without contact is how the tie works, not proof it was fake. ↗

The belief“I can just feel that they like me less than I like them.”

The dataThe liking-gap studies found the opposite bias: after conversations — and even across most of a year living with a roommate — people systematically underestimated how much the other person liked them. Your gut estimate of their side of the friendship is measurably skewed downward. ↗

The belief“A one-sided friendship proves there's something wrong with me.”

The dataNetwork analysis of school friendship data shows unreciprocated ties follow a structural pattern: almost all of them run from lower- to higher-status members of the network, in every network studied. One-sided ties are a property of how human groups organize, present in half of everyone's ledger — not a diagnosis of the person holding one. ↗

TEST_YOURSELF · How well do you know this science?

01 In the 2016 MIT study of friendship networks (Almaatouq et al.), what share of self-reported friendships turned out to be mutual?
Only about 53% of the friendship ties participants reported were reciprocated, even though they expected the vast majority to be mutual — a gap consistent with earlier studies finding reciprocity rates of 34–53%. source ↗
02 According to Hall's time-investment research (2019), roughly how many hours of shared time does it take to become someone's close friend?
Hall estimated roughly 50 hours to move from acquaintance to casual friend, about 90 to become a friend, and more than 200 hours to become a close friend — with time spent joking and catching up counting more than hours at work. source ↗
03 What did the 'liking gap' studies (Boothby et al., 2018) find about how people judge their conversation partners' feelings?
Across lab conversations, a personal development workshop, and dorm roommates tracked over an academic year, people consistently rated themselves as less liked than their partners actually reported liking them. Shyness made the gap larger but the bias appeared broadly. source ↗
04 According to equity theory in close relationships (Hatfield), who experiences distress in an unbalanced friendship?
Equity theory holds that both sides of an inequitable relationship experience distress: under-benefited partners report anger and resentment, over-benefited partners report guilt — and both are motivated to restore balance or exit the relationship. source ↗
05 When Ball and Newman (2013) analyzed unreciprocated friendships in school networks, what pattern did they find?
In every network studied, without exception, a ranking of members existed such that almost all unreciprocated friendships ran from a lower-ranked person toward a higher-ranked one — a pattern the authors interpret as reflecting social status. source ↗