This research roundup gathers the ten most important studies on friendship published in the last twenty years, spanning Jeffrey Hall's time-investment research, Robin Dunbar's layered friendship model, Daniel Cox's friendship-gap surveys, the Harvard Study of Adult Development on relationship quality, Nicholas Christakis's social-network research, and the networked-individualism work of Rainie and Wellman. Each entry explains who conducted the study, what it found, why it matters, and how to cite it. Friendship research has expanded rapidly as loneliness has become a public health concern. The studies below form the essential canon for anyone trying to understand why adults struggle to make friends, how many friends people actually have, and what qualities distinguish strong friendships from weak ties. Jeffrey Hall at Kansas quantified that it takes roughly 200 hours to form a close friendship, and Robin Dunbar at Oxford identified the layered structure of human friendship networks. Daniel Cox at the Survey Center on American Life has tracked the collapse of close friendships since 1990. The Harvard Study shows friendship quality predicts health at 80 better than cholesterol. This roundup will be useful to anyone thinking about their social life, trying to build friendships, or researching social connection. Citations include original venues. The convergent finding: friendship is more structured, measurable, and consequential than popular culture assumes.

1. What Did Jeffrey Hall Find With His 200-Hour Friendship Study?

Jeffrey Hall at the University of Kansas published a 2019 study in the Journal of Social and Personal Relationships quantifying how much time friendship formation requires. Analyzing data from 355 adults and 112 university students, he found that acquaintance-to-casual took 50 hours, casual-to-friend took 90 hours, and friend-to-close-friend took 200 hours. It matters because it is the first empirical answer to a universal question. Citation: Hall, Journal of Social and Personal Relationships (2019).

2. What Did Dunbar's Number Research Reveal?

Robin Dunbar at the University of Oxford proposed Dunbar's number (around 150) in 1992 as the cognitive limit on stable social relationships, based on primate neocortex ratios. Subsequent research confirmed the number across hunter-gatherer tribes, military units, and corporate divisions. Dunbar later identified layered subgroups: five intimate friends, fifteen close, fifty good, and 150 meaningful. It matters because it gave friendship a quantitative ceiling. Citation: Dunbar, Journal of Human Evolution (1992).

3. What Did the Survey Center Friendship Gap Report Show?

Daniel Cox at the Survey Center on American Life published the 2021 State of American Friendship report, finding that 12 percent of Americans have no close friends, up from 3 percent in 1990. Among men the rate was 15 percent. The median number of close friends had dropped from 3 to 2. It matters because it quantified the collapse of American friendship over a generation. Citation: Cox, Survey Center on American Life (2021).

4. What Did Waldinger and Schulz Find About Friendship and Longevity?

Robert Waldinger and Marc Schulz directed the Harvard Study of Adult Development and published The Good Life (2023) synthesizing 85 years of longitudinal data. Relationship quality (not wealth, fame, or career) at age 50 was the strongest predictor of physical and mental health at 80. Close friendships were central to the effect. It matters because it is the most definitive evidence that friendship affects longevity. Citation: Waldinger and Schulz, The Good Life (2023).

5. What Did Christakis's Social Network Studies Show?

Nicholas Christakis at Yale and James Fowler at UC San Diego analyzed the Framingham Heart Study social network data in a series of papers starting in 2007 in NEJM. They found that happiness, obesity, smoking, and loneliness all spread through social networks up to three degrees of separation. Friends of friends affect you. It matters because it showed social ties have measurable contagion effects. Citation: Christakis and Fowler, New England Journal of Medicine (2007).

6. What Did Parigi and Henson Find About Friendship Loss?

Paolo Parigi and Warner Henson published research analyzing friendship turnover, finding that Americans replace about half of their close friends every seven years. The churn is higher for younger adults and during major life transitions. It matters because it shows friendship is dynamic, not static, and life transitions (moving, parenting, divorce) are the largest source of loss. Citation: Parigi and Henson, Annual Review of Sociology (2014).

7. What Did Rainie and Wellman Establish With Networked Individualism?

Lee Rainie of Pew Research and Barry Wellman at University of Toronto published Networked: The New Social Operating System (2012), arguing that modern friendship is no longer rooted in geographically bounded groups but in personal networks spanning digital and physical space. Their research described how people manage multiple overlapping networks as individuals rather than members of tight-knit communities. It matters because it explained how friendship structure has changed in the internet era. Citation: Rainie and Wellman, Networked (2012).

8. What Did the 150 Friends Ceiling Research Demonstrate?

Follow-up research on Dunbar's number by Russell Hill and Robin Dunbar in 2003 and later studies using mobile phone records, Facebook data, and Twitter networks confirmed the 150 ceiling even in digital contexts. Tamas David-Barrett's 2020 paper replicated the finding across 27 million mobile users. It matters because it showed the social brain's limit is biological, not technological. Citation: Dunbar et al., Royal Society Open Science (2015).

9. What Did Small-World Friendship Network Research Show?

Stanley Milgram's 1967 six degrees of separation experiment was revisited and confirmed with digital data by Duncan Watts and others in the 2000s. Jure Leskovec's 2008 Microsoft Messenger study of 240 million users found an average of 6.6 degrees. It matters because it showed that despite small close circles, wider networks give everyone short paths to nearly anyone else. Citation: Leskovec and Horvitz, Communications of the ACM (2008).

10. What Did Dunbar's Layered Friendship Model Establish?

Robin Dunbar's layered friendship model, refined in papers through the 2010s, identified distinct layers: the intimate five (closest confidants), the sympathy group of fifteen (close friends), the affinity group of fifty, and the active network of 150. Each layer requires different time investment, with the closest layer taking the most. It matters because it gives friendship a precise architecture rather than a vague sense of closeness. Citation: Dunbar, Philosophical Transactions of the Royal Society B (2018). These ten studies together tell a coherent story. Friendship is bounded (roughly 150 people), layered (with five intimates at the core), time-hungry (200 hours to close), declining in America (12 percent report none), contagious (effects ripple three degrees), and the most reliable predictor of health and happiness across decades of longitudinal data. If you want to build friendships, the research suggests you need to invest time in person, tolerate the churn of life transitions, and understand that even the closest layer is only five people. If you want to argue that friendship is optional, the Harvard Study data will disagree.