david silver wrote:
ooh, interesting sounding thread!
like frank, i'm not sure if the study was an example of unethical research rather than methodologically flawed research. along those lines, in addition to the notorious rimm report (for a great summary of the controversy see http://www.fair.org/media-beat/950719.html) i'll add two more:
1. Carnegie Mellon's HomeNet study
the study: http://www-2.cs.cmu.edu/afs/cs.cmu.edu/user/copetas/www/public/pr/aug31-98.h...
reaction: http://www.columbia.edu/cu/21stC/issue-3.4/featherstone.html
I have read on a number of occasions attacks on the HomeNet studies for poor methodology. I am yet to read in these same messages details as to why considering the claims that were REALLY made by the researchers as to why the research was so "Flawed". The depression claim was not made as the journalists wrote. Nearly all research is flawed, the question then is relative. Perhaps David would like to explain why HomeNet is particularly flawed research. I should point out that I have NO CONNECTION to HomeNet although I very much liked many of the papers that have come out this study. For example M. Christ, R. Krishnan, D. Nagin, R. Kraut, O. Günther, Trajectories of individual WWW usage: implications for electronic commerce, Proc. 34th Hawaii International Conference on System Science (HICSS-34), 2001. The growing presence of the Web in everyday life is inextricably connected to the exponential growth in number and variety of Web sites offering information, commerce and services. While the number of users making use of the Internet and the Web has also grown tremendously, little is known about the extent to which individuals utilize (say, in terms of the number of visits to web sites) the Web and the trajectory of the change over time of such utilization. For example, we do not know whether the overall growth in Web usage is attributable to the increased numbers of users or to increased intensity of use of established users or both. This research is aimed to report the results of an analysis of three years of longitudinal data on residential Web usage. This data was assembled as part of the HomeNet project at Carnegie Mellon University. Drawing upon recent advances in semi-parametric, group-based statistical modeling, I examine whether there are distinctive clusters of trajectories of Web usage. I find that Web users can be clustered into four groups with distinct trajectories of use. Each of these groups achieve saturation in their extent of Web usage as measured in the number of distinct Web sites they visit over time. Demographic profiles of these different user groups will also be developed. The results are considered to have important implications for Internet marketing strategy, and public policy pertaining to the digital divide.