Americans tend to have fewer close confidants today than they did two decades ago -- but that isn't because they're all huddled over their computers playing World of Warcraft or reading the Volokh Conspiracy.

A report released Wednesday by the Pew Internet & American Life Project suggests that the Internet and other new communication technologies have, if anything, a modestly positive effect on the size and diversity of people's friendship networks.

The study found that using the Internet is associated with having more, not fewer, intimate friends. And Internet users are generally no less likely than nonusers to maintain face-to-face ties with their neighbors. Bloggers, for example, are 72 percent more likely than the general population to belong to a local voluntary organization.

So the common fear that old-fashioned kinds of social capital will evaporate as people spend more time online doesn't seem to be warranted.

But not all the news in the Pew report is sunny. The authors, who include three scholars at the University of Pennsylvania's Annenberg School of Communication, found evidence that Americans' friendship networks have shrunk significantly in the last two decades. The Internet isn't to blame for that trend, the Pew authors say, but the trend seems to be real.

The Pew report is based on a 2008 telephone survey of roughly 2,500 American adults. The survey included the same questions about friendship networks that were asked by the General Social Survey -- a longstanding study based at the University of Chicago -- in 1985 and 2004.

The 2004 round of the General Social Survey appeared to discover Americans' intimate-friendship networks had drastically shrunk since 1985. Among other things, the proportion of Americans reporting that they have zero intimate friends rose from 10 percent to 24.6 percent.

But that finding has been called into dispute. Claude S. Fischer of the University of California at Berkeley believes it is highly implausible that friendship networks have declined so badly, and he has argued that something must have gone wrong in the collection or coding of the 2004 survey.

The Pew researchers tried to shed light on that argument by employing essentially the same questions that the General Social Survey had used. (The core question is: "Looking back over the last six months -- who are the people with whom you have discussed matters that are important to you?")

Pew's 2008 survey found that only 12 percent of respondents reported having zero confidants -- less than half the level in the 2004 General Social Survey. Score one for Mr. Fischer.

But the Pew study also found that people named an average of 1.93 intimate friends, which is close to the 2.08 level that was found in the 2004 General Social Survey. In 1985 that figure had been much higher, at 2.98. So if these surveys are to believed, Americans have, on average, one fewer intimate friend than they did during the Reagan administration.

For more musings on the relationships between technology and social capital, see this new post by the University of Arizona's Lane Kenworthy.