The study also finds that:

• The trend toward social isolation mirrors other class divides: Non-whites and people with less education tend to have smaller networks than white Americans and the highly educated.

• Racial diversity among people’s networks has increased. The percentage of people who count at least one person of another race in their close network has gone up from about 9 percent to more than 15 percent.

• The percentage of people who talk only to family members about important matters increased from about 57 percent to about 80 percent, while the number of people who depend totally on their spouse has increased from about 5 percent to about 9 percent.

General Social Survey Measures Americans’ ‘Discussion Networks’

The data come from the General Social Survey (GSS), one of the nation’s longest running surveys of social, cultural and political issues. The survey, which has been conducted by the National Opinion Research Center at the University of Chicago since 1972, is composed of face-to-face interviews with Americans over 18 who are not living in institutions.

In addition to the standard set of questions, additional questions are added for specific studies, such as this one. The last survey on confidants was done in 1985. In the 2004 GSS, the questions were repeated to measure how people’s social networks had changed over time. This study, comparing the data, was funded by the National Science Foundation and the CIRCLE Foundation.

In addition to Smith-Lovin, the study was conducted by Miller McPherson, a research professor of sociology at Duke and professor of sociology at the University of Arizona, and Matthew E. Brashears, a graduate student at the University of Arizona.

In the survey, 1,467 people were asked to give the first names or initials of the people with whom they had discussed matters that were important to them in the past six months. Researchers followed up with questions about the gender, race, education and age of their confidants, as well as family ties, the length of their relationship and frequency of contact.

The answers measure what the researchers call “core discussion networks” and provide “a window into an important set of close, routinely contacted people who make up our respondents’ immediate social circle,” the study said.

The dramatic drop in the number of people in these discussion networks was not anticipated by the researchers, who have plans to follow up with more surveys in the future.

“We were surprised to see such a large change. We remain cautious—perhaps even skeptical—of its size. It’s unusual to see very large social changes like this that aren’t tied to some type of demographic shift in the population,” McPherson said. “But even if the change is exaggerated for some reason, given our analyses of the highest quality, nationally representative data available, we are confident there is a trend toward smaller, closer social networks more centered on spouses and partners.”

Other Findings Show Racial Diversity, Disparity in Social Networks

Most sociologists consider these “discussion networks” to be an important social resource, providing counseling and other valuable help in people’s lives.

Hurricane Katrina showed how important these resources are, Smith-Lovin said. “They make up a safety net of people who will help and support us, both in terms of routine tasks and also of extreme emergency. Americans have become much more dependent on a small number of very close family contacts -- usually spouses or partners or parents -- for that kind of help,” she said.

The researchers also found that Americans are stratified according to education and race when it comes to these social networks. African Americans and other non-white Americans have smaller networks of confidants than white Americans. African-American men older than 60 have seen the biggest decline, from 3.6 people in 1985 to 1.8 in 2004.

“People who are disadvantaged in various ways are especially likely to have smaller, more family-based networks,” Smith-Lovin said.

Why such a large change?

While this study did not uncover the reasons behind this social change, the researchers offer some ideas based on other research.

One possibility is that people interpreted the questions differently in 2004 than they did in 1985. What people define as “important” might have changed, or people might not equate emailing or instant messaging with “discussing.”

The researchers also suggest that changes in work and the geographical scattering of families may foster a broader, shallower network of ties, rather than the close bonds measured by this study.

Research also shows a decline in the number of groups that people belong to and the amount of time they spend with these organizations. Members of families spend more time at work and have less time to spend on activities outside the home that might lead to close relationships.

And new technology, while allowing people to connect over larger distances, might diminish the need for face-to-face visits with friends, family or neighbors, the study said.

“Group membership is very important in creating ties to people outside the family,” Smith-Lovin said. “But those ties may be more superficial now. If people spend less time in groups, they may talk to people, but just about matters that involve the club, and they may be less likely to share personal troubles or triumphs with them.”