Examining the Effect of Internet
Use on Television Viewing
Details Make a Difference
The proportion of U.S. households with a computer soared from 8.2% in 1984 to 56.5% in 2001. By September 2001, over 50% of U.S. households also had Internet access (U.S. Department of Commerce, 2002). The dramatic changes now occurring in household computing have the potential to change the lives of average citizens as much as the telephone did in the early 1900s and the television did in the 1950s and 1960s. Social scientists are now attempting to document how computing and the Internet are becoming integrated into the daily lives of users and the effects that this use is having. For example, efforts to document the effect of Internet use are occurring in the domains of psychological use of time and other media (see chapter 5), establishment and maintenance of social relationships (see chapter 19), political participation (Katz & Rice, 2002), psychological functioning, health, education (see chapter 11), and consumer behavior, among other domains.
This chapter examines the methods that many social scientists deploy to examine this effect. We review prior literature to illustrate problems that are widespread in examining the influence of Internet use. We argue that much of this research reaches limited or even erroneous conclusions, both because it uses cross-sectional data to draw causal implications and because it fails to distinguish between varieties of Internet use.
The empirical section of this chapter is based on a national panel survey of 980 individuals. We show that the Internet is used for a wide range of functions. We use confirmatory factor analysis to disaggregate overall Internet use into a set of components, which are only moderately related with each other: interpersonal communication with friends and family, interpersonal communication with strangers, instrumental information seeking, entertainment, and commerce. Just as one would expect that chatting with friends on the phone and watching entertainment on television would have different influences on those who use these earlier technologies, so too would one expect that using the Internet for different purposes will have different effects on its users.
The empirical study examines the effects of Internet use on television viewing, illustrating the time-use perspective on the effect of new media. We use hierarchical linear growth modeling to differentiate cross-sectional from longitudinal relationships. Conclusions from cross-sectional data differ from those based on longitudinal data, and both sets of conclusions depend on how people use the Internet. For example, the cross-sectional analyses