Here’s an “important new paper”:http://www.asanet.org/galleries/default-file/June06ASRFeature.pdf by two former colleagues of mine (just departed for Duke) and one of our grad students here at Arizona. The paper compares the social network module of the 2004 General Social Survey (GSS) to the 1985 GSS, the last to include network questions. The key question of interest is this:
bq. From time to time, most people discuss important matters with other people. Looking back over the last six months—who are the people with whom you discussed matters important to you? Just tell me their first names or initials.
The survey went on to probe the respondents about their relationship to the people they mentioned, and the relationship of these people to one another. The new findings are striking: since 1985, the number of people saying there is _no-one_ with whom they discuss important matters nearly tripled, rising to about a quarter of the respondents. As McPherson et al say,
bq. The modal respondent now reports having no confidant; the modal respondent in 1985 had three confidants. Both kin and non-kin confidants were lost in the past two decades, but the greater decrease of non-kin ties leads to more confidant networks centered on spouses and parents, with fewer contacts through voluntary associations and neighborhoods. … Educational heterogeneity of social ties has decrease, racial heterogeneity has increased.
The predicted probability of social isolation is much higher the fewer years of education one has. Also. “Young (ages 18–39), white, educated (high school degree or more) men seem to have lost more discussion partners than other groups.”
The observed differences are pretty big, as these things go. Are they real? It may be that the 2004 respondents differed from the 1985 respondents in their interpretation the words “discuss” and “important.” (People might interpret “discuss” as face-to-face discussion, when they may also be pouring out their hearts on a blog somewhere, for instance.) Because of these issues, the authors spend a lot of time investigating the validity of the measure. More interestingly, it may be that we really are observing a shift in patterns of network affiliation. Feel free to speculate in the comments, but also take a look at the paper — it discusses several of the most plausible interpretations of the shift, in addition to documenting the findings.