User Satisfaction with Information Seeking on the Internet

The research described in this article focused on how just how satisfied users are when they look for informasatisfied Australian academics are when they use the tion using the network is well overdue. Internet to search for information. The significant methUnfortunately, this is not a simple matter. Satisfaction odological outcome of the research was its validation of is an extremely complex variable. This is particularly true magnitude estimates of user satisfaction with informawhere satisfaction is characterized as a summary assesstion seeking on the Internet. Testing the validity and reliability of magnitude estimation as a technique for gathment of information seeking by an information user. Satering and analyzing interval data on satisfaction with inisfaction with information seeking is a state of mind which formation seeking was key to the investigation. Data for represents the composite of a user’s material and emouser satisfaction were then associated with end-user tional responses to the information seeking context. An characteristics like training, frequency of use, and expectation of success. The study found that Australian information user will experience material satisfaction as academics generally have a high expectation of success a result of factors associated with various features of an when they engage in information seeking on the Internet, information system’s performance. Emotional satisfacand are satisfied with the process regardless of how tion, on the other hand, deals with feelings of satisfaction frequently they use the network or whether or not they based on various things like the user’s requirements, exhave received formal training. pectations (Applegate, 1993; Tessier, Crouch, & Atherton, 1977), goal determination, and task orientation Introduction (Hiltz & Johnson, 1989; Waern, 1989). Clearly, if satisfaction with information seeking is difIn recent years, the Internet has been described as an ficult to define, it is equally difficult to measure—at least information superhighway or information infrastructure in a reliable and valid way. There have been numerous to emphasize the widely held view that this global netattempts to do so, however (see, for example, Auster & work will inevitably transform the way we create, manipLawton, 1984; Bailey & Pearson, 1983; Baroudi & Orliulate, store, retrieve, transfer, and utilize information kowski, 1988; Hiltz & Johnson, 1989; Ives, Olson, & (Arms, 1990; Cerf, 1994, 1995; Dempsey, 1993; Kahin, Baroudi, 1983; Murfin and Gugelchuk, 1987; Nath, 1989; 1993, 1995). This sort of rhetoric is particularly useful Sandore, 1990). In general, studies like these have used a for popularizing and promoting the network because it multivariate approach when measuring satisfaction. They implies that the Internet somehow provides a supportive have tended to operationalize satisfaction from a list of context for effective information seeking by information indicators, and then inferred a level of satisfaction from users. Indeed, this untested imputation and the benefits that are assumed to derive from Internet use have, to date, the sum of responses to these indicators. In the end, these provided sufficient justification for a huge investment in multivariable measures are generally validated in a similar internetworking across the academic, government, miliway. The technique being used to measure satisfaction is tary, and commercial sectors (McClure & Lopata, 1996). often purported to verify the factors that comprise a conOf course, there a many issues associated with the Internet struct for the variable, if these judgments are consistent which warrant empirical study, but an attempt to find out with a user’s response to a direct question about overall satisfaction measured on a category rating scale. In other words, at the end of a series of observations, each subject Received February 28, 1997; revised May 2, 1997; accepted June 6, is asked about their overall satisfaction. The response to 1997. this direct univariate measure is then used to validate the multivariate measure being proposed. q 1998 John Wiley & Sons, Inc.