http://ctx.sagepub.com/Contexts http://ctx.sagepub.com/content/7/3/32The online version of this article can be found at: DOI: 10.1525/ctx.2008.7.3.32 2008 7: 32ContextsLynn Smith-Lovin, Matthew Brashears and Miller McPhersonThe Ties that Bind are Fraying Published by: http://www.sagepublications.comOn behalf of: American Sociological Association can be found at:ContextsAdditional services and information for http://ctx.sagepub.com/cgi/alertsEmail Alerts: http://ctx.sagepub.com/subscriptionsSubscriptions: http://www.sagepub.com/journalsReprints.navReprints: http://www.sagepub.com/journalsPermissions.navPermissions: http://ctx.sagepub.com/content/7/3/32.refs.htmlCitations: What is This? - Aug 1, 2008Version of Record >> at Texas A&M University - Medical Sciences Library on January 14, 2014ctx.sagepub.comDownloaded from at Texas A&M University - Medical Sciences Library on January 14, 2014ctx.sagepub.comDownloaded fromThink back over the last six months and the people with whom youdiscussed the things most important to you. How many were there?Sociologists asked that question in the 1985 General SocialSurvey (GSS), a national survey of nearly 1,500 adults, and createdthe first representative picture of Americans’ networks of confi-dants. Answers to the same question in the 2004 GSS uncoveredsomething remarkable: Americans had one-third fewer confidantsthan two decades earlier. the ties that bind are frayingby miller mcpherson, lynn smith-lovin, and matthew brashears32contexts.orgIllustration courtesy xkcd.com at Texas A&M University - Medical Sciences Library on January 14, 2014ctx.sagepub.comDownloaded fromIn 2004, many more people said they don’t discuss matters ofimportance with anyone. And they’ve shifted away from tiesformed in the community—at places like church choirs, neigh-borhood associations, social clubs, and sports teams—andtoward conversations with family, especially spouses. This question and its disturbing answer are significantbecause the closer and stronger the tie we have with some-one, the more support they offer us. Close ties help us withroutine things, like picking up a child at day care, and offermajor aid in a crisis, like providing a place to live after a disas-ter. They influence us directly through our interactions and indi-rectly by shaping the people we become. Having at least oneclose connection can be vital for both physical and mentalhealth.Strong social ties can be important for society, too. Harvardpolitical scientist Robert Putnam argues that neighborhoodrelationships and voluntary group membership are importantfor community well-being and democratic participation. Hefollows a rich tradition, dating back to Alexis de Tocqueville,suggesting that Americans’ ties to other members of their com-munities enhance our democratic institutions. measures and missesWhen researchers study connections between people, akey issue is choosing which type of relationship to measure. In1985 and 2004 the GSS focused on one type of relationship—who people use as their confidants. From earlier research weknow these “discuss important matters” relationships are sim-ilar to questions asking about best friends and other very closeemotional ties. These aren’t the kinds of friends we have on Internet siteslike Facebook, MySpace, LinkedIn, or Friendster. While we mayknow more about people connected to us through those sitesthan people we talk to every day, what we know about thesefriends is equally available to hundreds of other people. Theserelationships are what sociologists call “weak ties”—friendsof friends or acquaintances we know only through one type ofconnection. It may be interesting to collect hundreds of suchcontacts, but we aren’t likely to call on them for help on a dailybasis. When answering the GSS question, people tend to men-tion family members and friends who are as close as family—people they’ve known a long time and who they interact withseveral times a week. The “important matters” question does not measurewhat people talk about in their relationships. Important mat-ters vary dramatically from respondent to respondent, encom-passing relevant personal matters like intimate relationships,finances, health, hobbies, and work problems, as well as moregeneral topics like politics. For example, a 70-year-old manmight talk about his health while a 21-year-old woman mighttalk about which job to accept after college. We also tend to talk about different things with differentpeople. Respondents talk about different things with theirspouses (children, education, finances) than with their co-work-ers and neighbors (community, politics, work). In fact, onestudy found that some people interpreted the question in termsof the frequency or intimacy of relationships, rather than thespecific discussion of an important matter during the last sixmonths.In other words, the GSS network question about dis-cussing important matters leads people to tell us about a closeset of confidants who make up an important interpersonalenvironment for the transmission of information, influence,and support. We would be unwise to interpret the answers tothis question too literally. However, they do give us a windowinto an important set of close, routinely contacted people whomake up our respondents’ immediate, close social circle. the numbersGiven the close, densely interconnected nature of thesocial ties measured by the GSS question, we didn’t expectmuch change in the core confidants of the typical Americanbetween 1985 and 2004. We were clearly wrong. Most notable was the large drop in the size of core discus-sion networks. The number of discussion partners decreased bynearly one person (from a mean of 2.9 to a mean of 2.1) inthe past two decades. The most common answer to our ques-tion about confidants dropped from three to none. Almost halfthe population (43.6 percent) now reports they discuss impor-tant matters with only one other person or with no one at all—a level of connection Claude Fischer called “inadequate socialsupport” in his classic book To Dwell Among Friends. Havingonly one close person was inadequate because it made onevery vulnerable to losing that lone relationship, he argued. Thechange is especially dramatic among those who reported fouror more discussion partners. These well-connected respondentshave decreased from more than one-third of the population(38.4 percent) to only one-fifth (20.2 percent).This