[ Executive Summary | Methodology | Results | Conclusions | Critiques ]
A longitudinal study by Carnegie-Mellon University's HomeNet Project found that increased Internet use correlated (and likely caused) a decrease in social involvement. While we lack some of the data necessary to validate this claim, the study makes perhaps the strongest case for a link between Internet use and isolation.
Families were surveyed in the home both before getting Internet access and also one year afterwards. Also, their Internet usage (both amount and type of use) was monitored electronically.
Participants were recruited from adults who were on the board of directors at community development organizations, and teenagers who worked on their school newspapers. Each didn't have Internet access before the study. Carnegie Mellon University provided each with a computer, software, phone line, Internet service, and basic support; in almost all cases, this was "their first experience with a powerful home computer."
Email was extremely popular, being used both more frequently and more consistently than the World Wide Web. This mirrors the SQISS's finding that email is the most popular use for the Internet.
Greater use of the Internet correlated with less communication within the family, smaller local social networks, greater loneliness, and greater depression. This correlation is statistically significant and remains even if one corrects for prior conditions (e.g. how depressed people were before receiving Internet access).
The Carnegie Mellon researchers were initially quite surprised at finding social isolation, despite a great deal of anecdotal evidence about people becoming more connected and enjoying the Internet. Robert Kraut, the lead psychologist on the study, said that they "were surprised to find that what is a social technology, unlike television, has kind of antisocial consequences" (qtd. in Sleek). Their paper was thus titled "Internet Paradox," highlighting the Internet's paradoxical role in both facilitating social interaction as well as making one feel more isolated.
While the researchers caution that there might be outside variables, such as the difficulty of adolescence, which might possibly lead to both high Internet use as well as social withdrawal, they point to the longitudinal nature of the study to suggest causal relationships. "Measuring at multiple times…allow[s] us to make a causal claim from correlational data," they write in a follow-up paper. "Social and psychological well-being measured before they gained home access to the Internet did not predict how much they subsequently used it. In contrast, their hours of Internet use did predict subsequent declines in their psychological and social well-being." That is, they can just about rule out the possibility that lower social involvement causes higher Internet use, thus suggesting the converse.
It is unlikely that the Internet, like television, completely displaces social activities. This is because most Internet use was social, such as email—so it is merely replacing one kind of social interaction with another. The investigators suggest that perhaps social interactions over the Internet encourage weaker ties than with other technologies (such as telephones) or face-to-face interactions. They point to several anecdotes about people who made friends online, yet who do not feel particularly close to them. Additionally, they note, many relationships online are limited by physical distance, making it all but impossible to offer tangible support to one's friends.
Before picking apart the study, it should be noted that its methodology is far greater than any of the other survey-only studies. By designing a longitudinal study, the researchers were able to accurately assess changes in behavior over time due to the Internet—something that a simple survey cannot do.
Despite the longitudinal nature of the study, because there was no control group that did not receive Internet access, one cannot irrefutably show that Internet use causes lower social involvement. Indeed, if one examines the data, one finds that the participants communicated with family members and increased the size of their social networks. This suggests that, while high Internet use may indeed have a causal relationship with isolation, Internet use itself actually has the reverse effect. Likewise, as the researchers admit, there could be outside variables that cause both high Internet use as well as decreased social involvement (such as a higher stress period of one's job). Without a control group, one cannot tell for certain.
The researchers admit that, while their results were paradoxical, the "effects were not large." Indeed, others question whether the small effects would generalize over the population at large, seeing as the 169 people were selected from certain age groups in a small area around Philadelphia. The investigators believe that their results "will generalize to groups similar to those studied," however, that is, young and middle-aged adults.
The Carnegie Mellon study report merely shows correlations between increased Internet use and certain other variables, but it does not elaborate as to what levels or types of use correspond to social isolation. For example, it may be the case that under 10 hours of use a week, the Internet is a positive social force, but with increasing amounts of Internet use over 10 hours, it becomes increasingly negative. Similarly, the investigators admit that certain measures are only significant if one considers time using the World Wide Web, as opposed to email, but they do not provide a breakdown of time on the Internet. In order to test their "displacement of strong ties" theory and—most importantly—in order to draw conclusions about Internet policy, it is vital that one understand how the Internet was used by the participants, and how this affects the various variables.
