The Chronicle of Higher Education reports this morning on an ongoing debate as to the validity of a 2006 study which concluded that Americans have become significantly more socially isolated over the last 25 years. 

David Glenn reports, “In the summer of 2006, several major news outletsgave prominent coverage to a sociological study with a grim message: Americans’ social isolation had increased radically since the 1980s. Whereas in 1985 Americans reported that, on average, they had 2.94 friends or family members with whom they could discuss important matters, by 2004 that number had dropped to 2.08. A quarter of Americans had no close confidants at all. Those findings were …[even] startling to the study’s authors, who are sociologists at Cornell University, Duke University, and the University of Arizona, [J. Miller McPherson, Lynn Smith-Lovin, and Matthew E. Brashears].”

UC Berkeley sociologist and social networks scholar Claude Fisher has some concerns:

The [previous] study’s portrait of collapsing social networks, Mr. Fischer writes, is at odds with other recent findings by social scientists. What’s more, he says, some of the 2006 paper’s data seem internally inconsistent or simply implausible. For example, among people who reported belonging to four or more organizations—presumably a highly sociable bunch—14.9 percent reported having no confidants. And what about married people? Surely they discuss important matters with their spouses, if no one else. In 1985 only 6.6 percent of married respondents reported having no confidants, but in 2004, 22.2 percent did so.

Fisher claims that such errors could be due to errors during data collection or coding. Now the original study’s authors have responded…