Social scientists studying a representative town in America’s heartland reported widespread concern among its residents that friendship was waning, that true friends were fewer than a generation before. This report, Middletown, was published in 1929. About two decades later, The Lonely Crowd, a sociological study of American character in the post-war era, described its subjects as shallow conformists, eager to imitate others but lacking in deep connections. Twenty years after that, best-selling author Vance Packard declared in A Nation of Strangers that “personal isolation is becoming a major social fact of our time.” In the 1990s, political scientist Robert Putnam famously described Americans as bowling — and doing much more — alone. ¹ Now, in the 21st century, declarations that Americans are friendless abound. In late 2009, culture critic William Deresiewicz wrote of “faux friendships,” merely insubstantial feelings that “each of us hugs privately … in the loneliness of our electronic caves.” The 2020s phrase “friendship recession” has become popular enough to have a Wikipedia page. 

Alarms over the state of American friendship are nothing new. Over the last few decades, there has been a surge in writing about friendship in books and newspapers. ² Does this surge reflect a real crisis or simply the increasing value Americans place on friendships? Or is it just a popular cultural meme unmoored from reality? It’s possible that all three conclusions are true, depending on whether we take the long view or the short view. After all, a century ago Americans were writing about friendship as much as they do now — with perhaps as much concern and perhaps more reason. ³

A friend, as most Americans today use the word, refers to someone with whom we have a voluntary,  affectionate, supportive, and perhaps intimate relationship, and with whom we need not be entwined in any other way — not as relative, client, neighbor, fellow platoon member, whatever. Such friendship, which persists out of mutual regard and not because of social constraints or pragmatic necessity, is a modern development for ordinary, particularly middle-class people. Classical, biblical, and medieval friendships are described this way, but only among elites. Most humans for most of history lived in relationships that were defined, prescribed, and controlled by family, clan, tribe, and place — often very ambivalent, tense relationships. Relationships outside such arrangements were, like encounters with strangers, suspect, and rightly so. 

The historical argument, as put by sociologist Allan Silver, is that the growth of commerce in the eighteenth century established strictly instrumental, market relationships, thereby creating a distinctive space for non-instrumental, personal relationships. ⁴ The latter were freed from other constraints of “necessity” like patronage or family. At about the same time, the western development of a more individuated self reinforced the emerging modern friendship. That bond ideally rests on each individual being in “sympathy” with a particular other individual based on each one’s unique, decontextualized personality. A third western development, lagging a bit, was the intensification of emotion, of sentimentality in particular. ⁵ Fulsome declarations of love (philia) animated the letters of upper-middle-class friends of the nineteenth century. ⁶ In the first half of the twentieth century, “a wider range of people,” wrote historian Mark Peel, “came to regard … intimate and emotional friendship as a crucial component of a good life …” The later twentieth century, “saw the triumph” of this ideal as cultural authorities increasingly prescribed friendship, ⁷ both literally and metaphorically. In the early 1980s, the California Department of Mental Health ran a “Friends Can be Good Medicine” campaign, urging residents of the Golden State to friend up. The Surgeon General of the United States did the same in 2023 when he wrote out an Rx for Americans: “make time for friends.”

But this is a vague sort of prescription. Americans hold quite varying notions of who is a friend. Some include relatives and others do not; some insist that a friend is someone with whom you share deep secrets with while others are satisfied with laid-back companionship; and some simply call anyone a friend who is a congenial acquaintance. This variation is probably the major reason why many Americans claim to have several dozen friends and many others claim to have only a couple. Dictionaries don’t help. Merriam-Webster’s, both in the 1820s and the 2020s, listed the first meaning of friend as one “attached to another by affection” but gave a second meaning simply as one who is “not hostile.” That today’s social media uses friend to mean an online connection muddles things further.

Then there are cross-cultural variations in “friends.” For generations, visiting foreigners have complained that Americans are very friendly but very unlikely to form deeply committed friendships. A French person feels free — indeed, feels obliged — to intervene for a friend in need, even to take control if necessary, observed a French anthropologist, while an American waits to be asked to help rather than undermine the friend’s independence. This is one of the ways in which the French (and others) experience American friendships as plentiful but superficial. ⁸  

Then there are historical changes in what role friends serve. Because Americans over the last half-century or so have married later, divorced more often, and had fewer children, today they have notably fewer relatives than their parents and grandparents. Non-kin friends, as well as paid professionals, increasingly provide companionship and support. Friends pinch-hitting for kin is notably common among the elderly who have no adult children and among LGBT Americans estranged from their families. ⁹ One sign of their evolving role are the new laws that allow friends to make medical decisions as if they were immediate kin of the patient.

If friends are increasingly taking on weightier responsibilities, they are probably also being held to higher standards. The historian Mark Peel writes, “The friendship of the [nineteenth-century] avant-garde —  intimate, reciprocal, open and time-consuming, where the self was formed and realized and in which anything could be shared — became more achievable” by the masses ¹⁰ — and, I would add, more expected of those whom we claim as “real” friends.

All this historical, domestic, and international variation in how people use the word friend makes comparing how many friends Americans have now to how many they had in the past a morass. Bowling Alone has become a catchphrase for our times, even though scholarship has challenged many of the book’s claims. Political scientist Robert Putnam’s central finding was that Americans in the 1990s were less likely to belong to organizations than Americans had been in the 1950s. Among the critiques of that conclusion was that the data Putnam used, especially in the first version of his work, relied on a checklist of organizational types that had become increasingly anachronistic and unrepresentative. ¹¹ It is probably more accurate to say that Americans became — proverbially — less likely to bowl in official leagues since the 1950s but no more likely to bowl alone. There is no doubt that the 1950s, with its exceptional combination of a baby boom and rapid economic growth, saw the founding and expansion of many organizations — suburban churches and PTA’s, for example — but we have no reason to think it was a heyday of friendship.

Another example of the pitfalls in trying to measure change in social connections is a 2006 article in the American Sociological Review which reported that the percentage of Americans who said that they had no one to talk to about “important matters” had more than doubled (from 10% to 25%) between 1985 and 2004. The report immediately gained New York Times headlines and morning show attention. It is still often cited many years later, long after scholars have demonstrated its methodological problems. For just one example, the survey asked respondents to name and describe the people with whom they are involved — a heavily taxing chore. The later in a survey this task comes, the likelier it is that interviewers or respondents skip such questions or curtail the answers. The 1985 survey asked the key questions early in the survey; the 2004 version introduced it near the end. Other issues with the survey include (but are not limited to) coding errors in the data as well as interviewer fatigue. Network scholars no longer rely on this study. ¹²

It is much more difficult to accurately track the number of friendships or measure their quality  than it is to track, say, marriage and divorce rates. Surveys obtain estimates in different ways — for example, asking respondents to estimate how many “friends” or “close friends” they have, or asking them to name specific friends. Surveys also measure the quality of friendships in different ways — how often friends get together or communicate, whether they discuss “important matters,” how much material help they provide, and so on. This makes comparison difficult. Take, the 2021 study from the American Survey Center that apparently birthed the “friendship recession” metaphor. It concluded that about half of Americans had three or fewer close friends, a marked increase from less than one-third cited thirty years before.  The researchers at the American Survey Center had made an effort in 2021 to match the question about friends that Gallup had asked in 1990: “Not counting your relatives, about how many close friends would you say you have?” However, the 2021 question was not exactly the same: “Now, thinking only about friends you are close to [italics added], not counting your relatives, about how many close friends would you say you have?” Perhaps even more important is where the question came within the survey. Gallup’s 1990 survey asked the question early, following a handful of mundane questions about friends. The ASC in 2021 asked their version later, after 34 questions focused on the intimacy of friendships, such as  “How satisfied … are you with … the number of friends you have?,” “When was the last time you … shared personal feelings with…,” and “… received emotional support from a friend?” The 2021 format seems to have significantly raised the bar on whom respondents should classify as a “close friend” compared to the 1990 one. Other technical issues arise across the spectrum of friendship research, for instance, declining response rates and online panel surveys replacing interviews of random samples. ¹³

Despite these problems, scholars try their best to separate signal from noise as they map change. In my review of data from 1970 to 2010 (reported in Still Connected, 2011), I drew on several surveys, such as the General Social Survey, the Gallup Poll, and the commercial survey that Putnam used in his book. In each case, I tried to match like to like — the same question asked by the same survey organization but different years. I concluded that rates of friendlessness and the median number of friends Americans claimed had probably not changed over those four decades. But the rhythms of friendship did shift. Americans saw friends less often at home (the number of dinner parties really plunged), but saw them more often outside the home. Overall communication with friends increased in volume. And Americans’ average satisfaction with friendships probably didn’t move. This year, sociologist Lane Kenworthy updated my book and returned with essentially similar conclusions. Another study found that the proportion of Americans who said that they could expect material help from friends did not drop. Some scholars, of course, would dissent. One study, for example, concluded that, while Americans became no more likely to have no close bonds, they reported fewer close ties on average. Overall, however, the conclusion on friendlessness, like that on loneliness (see below), is that there has not been a long-term deterioration in friendship.

The short term, however, may be different. After decades of insubstantial change — all those alarms notwithstanding — a few indicators point to a genuine drop in numbers of friends, or activity with friends in the 2010s. One or two continuing surveys — which are better designed to accurately assess change — have pointed to a decrease in how many friends respondents claimed. Between 2014 and 2018 the proportion of General Social Survey respondents reporting spending frequent “social evenings” with friends fell sharply. How much leisure time Americans reported spending with friends also dropped. Even the degree to which Americans said they valued friendships slumped. ¹⁴ (COVID, of course, disrupted social life, but these downturns appeared before 2020.) Importantly for our understanding, the drops in reported friendship activity were heavily concentrated among unmarried young adults. 

Some argue that greater use of smartphones explains the sudden change, but I am skeptical. The messaging, posting, and phone functions of smartphones promote more social interaction, not less. (That’s also what Americans believe about mobile technology.) Nor is there evidence that e-communication has led to a decline in relationship quality — we mainly use our phones to talk to or text with the same people we know in person. Others point to the rise of gaming, especially popular among unmarried young adults. Gamers describe their pastime as a social activity, but gaming does not require being physically together. ¹⁵ Here’s another possibility: In the last thirty years but especially during the 2010s, many more young adults, especially men, have been living with their parents — from seven percent of 25-to-35 year-olds in 1970 to 10 percent in 2000 to 17 percent in 2020, ¹⁶ and even higher among non-college graduates — perhaps because of high housing costs and stagnant income. Just under one-third of 25-year-olds in Gen Z live at home. Living with mom on a limited budget may put a crimp in seeing friends.

It will take a while to sort out what happened to friendships in the 2010s (and to whom and why) and to see whether the 2010s indicated merely  a temporary fluctuation in an otherwise steady or even intensifying culture of friendship — recall, for example, close friends becoming legally next-to-next-of-kin  — or whether the 2010s foretell an extended shrinkage in friendship ties for some Americans. Novel elements have emerged in the last decade or two. There is the Netflix effect, reducing the desire to get off the couch for entertainment (reinforcing what television did in the 1950s); the Amazon effect, eliminating the need to leave the house to shop; the COVID work-from-home effect, which precludes opportunities for seeing people in the office; and the iPhone effect, being always connected to everyone you care about, diminishing the need for — though not necessarily the actual frequency of — in-person contacts. We will have considerably more data to address such speculations looking forward than we have had looking backward. Meanwhile, we can provisionally conclude that, over the last half-century or more, friends have remained roughly constant, probably even expanding their roles in Americans’ lives. Yet, as we saw, that long history has usually been accompanied by repeated alarms about the loss of friendship. The American discourse that describes the loss of friendship as a cost of modern life is deeply rooted and intuitive.

Our thinking about friendship is shaped by both personal experience and contemporary culture. Most adults recall the easy friendships of childhood and the many, albeit often fraught, friendships of high school. Then, college, careers, moving, partnership, marriage, and children sunder some friendships and thin out others. ¹⁷ For example, a young woman in a survey my colleagues and I conducted explained why she had stopped listing a former friend among her intimates: “She found a new boyfriend … She just didn't have time for me.” Another said of a couple she had counted as friends only a year earlier, “I’m not as close to those guys as I was. They have two kids and they’re just really busy … we just drifted apart.” Even when we make new friends, we have still experienced loss. We then read our personal biographies as national history. Thus, Atlantic columnist Jennifer Senior wrote after discussing her own stories of friendship loss, “Were friendships always so fragile? I suspect not. ... One could argue that modern life conspires against friendship …”

Another distortion arises from a simple mathematical oddity pointed out by sociologist Scott Feld in 1991 and since memorialized in Wikipedia as the “Friendship Paradox”: Most of us have friends with more friends than we do. Why? “The most popular people appear on many ... people’s friendship lists, while people with very few friends appear on relatively few people’s lists.” Therefore, most of us see many people among our friends who have more friends than we have, and so we end up feeling comparatively friendless. All the more so given that this statistical illusion is greatly magnified on and by social media.

There are also social forces that make us feel friendless however many friends we have. For one, the last generation or two of Americans have grown up in a “therapeutic culture.” A practice once mainly of elites, more Americans have come to engage in self-exploration, self-improvement, and self-fashioning guided by religion ¹⁸ or increasingly guided by psychologists, both through media and in person. ¹⁹ Intensive examination of one’s friendships — or their seeming absence — is part of this self-examined life.

Paralleling this development has been the explosion of “service” or “lifestyle” journalism across books, magazines, newspapers, television, and the internet, all helping fuel the friendship talk that I described at the start of this essay. Much of service journalism focuses on clothes, food, fitness, and the like, but increasingly much also addresses personal issues, including friendship, as in this from The New York Times: “Why Is It So Hard for Men to Make Close Friends? American men are stuck in a ‘friendship recession.’ Here’s how to climb out.”

These developments build upon a canonical belief, centuries old, in the “loss of community.” It is the widely-shared conviction, inscribed in our literature, politics, and folk wisdom, that modern life has degraded intimate relationships. Decades of social historical research challenging this story have done little to weaken its hold. ²⁰ So, for instance, a film-maker who wants viewers to feel warm fellowship can deploy a small-town setting and one who would have viewers feel cold alienation a big-city one, sure that the image evokes the canon. 

These personal and cultural influences driving the friendship discourse are deep and long-lived. Social science can conclude that American friendship has thrived over the long term, but most Americans will still find that the traditional story of lost bonds feels intuitively true.

The friendship alarms are relatively muted compared to the sirens blaring an “epidemic of loneliness.” ²¹ The phrase is not entirely new — for example, it appeared in a couple of 1978 Los Angeles Times stories about widows and widowers. ²² The first instance that I have found was in a 1971 book by an evangelical church leader named R. Eugene Sterner, titled Being the Community of Christian Love. The first chapter is “Loneliness is an Epidemic.” It was in the 2010s, however, that the loneliness epidemic gained global attention. Culture critic Stephen Marche used the phrase in a 2012 Atlantic article about Facebook, while The New York Times first used it in a 2013 op-ed by Ross Douthat. The key moment came in 2017 when Obama’s Surgeon General Vivek Murthy, whose concern about guns nearly sank his appointment two years before, warned about the loneliness epidemic and would later prescribe friendship as a cure. By 2018, Republican senator Ben Sasse had written a book in much the same vein. But I would argue, as other scholars have, that what we really see are periodic scares about loneliness.

I reviewed, in the first of two 2023 blog posts, the research literature on trends in Americans’ reports of loneliness and concluded that — much as in the case of friendlessness — the data on loneliness over years did not support claims of a long-term increase, with the possible exception of teenagers since 2010. (For a couple of decades before 2010, teenagers had been reporting declining loneliness.) Other serious reviews of the loneliness claims concluded much the same. 

Many of these loneliness studies are based on measures taken at only one time and lack trend data. In 2018, for example, the Cigna Group issued a well-publicized report on a large survey full of loneliness-related questions. Statistical analysis of the answers led the authors to conclude that about half of Americans were lonely. The report presented this rate as high and worrisome but provided no historical benchmark for comparison. The report also set a low diagnostic bar for loneliness. Cigna labeled half of the survey respondents as lonely, but only by grouping the many respondents who answered “sometimes” to various questions about their loneliness together with the few respondents who answered “always.” Both groups became “Lonely Americans.”

Another source of confusion is blindness to the fact that reported loneliness depends mainly on whether someone has a spouse or a romantic partner, not on friends or other ties. ²³ Still yet another confusion is conflating living alone with feeling lonely. Yes, those “going solo” ²⁴ are a bit likelier to report days of loneliness than are other people, but living alone is much less associated with expressed feelings of loneliness than is a survey respondent’s claimed number of friends, and much, much less than the respondent’s lack of a romantic tie. ²⁵

In my second post on loneliness, I briefly strolled down the memory lane of loneliness panics: in the 1990s with worries about computer games, in the 1960s and 70s when capitalism was often described as creating “alienation” (as in the collection Man Alone), ²⁶ in the 1950s when the “silent generation” of the “organization man” was presumably a robotically impersonal generation, back to the early 1900s when rural isolation was the concern of federal investigations. Sometimes there was good reason to worry about loneliness — for farm women, in particular; or newly landed immigrants in the 1880s — but for the most part, talk about loneliness probably did not reflect the reality of isolation so much as rising standards for relationships. As I discussed above, the historical changes in friendship — its greater role, its infusion with sentimentality and therapy, and the like — have probably raised Americans’ expectations for at least “close” relationships. Historical divorce records show this inflation of expectations with regard to marriage. From the nineteenth into the twentieth century the courtroom standards for an acceptable spouse rose from simply being a good provider or home-keeper to being an intimate confidant and a constant companion. ²⁷ Along these lines, one historical study argued that “modern technologies have made friendship, romance, and social connection look easier than ever, and therefore the absence of such relationships has become all the harder to bear.” ²⁸ If there has been increasing chatter about loneliness, then, it has been more the result of higher expectations and greater self-reflection — especially among the chattering classes — rather than greater isolation.

And then came COVID. And then came a flood of research on COVID and loneliness. The short story is that there was a surge, although not immediately, in Americans confessing loneliness. But by 2023, that surge had subsided almost to the pre-COVID level. ²⁹

Americans have been repeatedly warned about loneliness and about the loss of friendship. The alarms, of course, may not always be false. History might prove the 2010s to be one of those times where there was some reason for concern — at least about young men. (The 2020s are still unfolding and have been radically affected by COVID.) The 2010s were also a difficult period in several ways, such as the long, slow climb out of the 2008-09 economic crisis and the spillover of boiling politics into other aspects of life. It is difficult to step back, think hard, accurately analyze, and draw solid conclusions about something as personal, subjective, and spongy as friendship. It is important to do so, however — in particular for the public health establishment that has turned its attention to loneliness — because false alarms come with a price, particularly in diverting our attention from other, less amorphous matters, such as economic dislocations, violence, and real epidemics.