Patterns of Information Search and Access on the World Wide Web: Democratizing Expertise or Creating New Hierarchies? By Alexandre Caldas, Ralph Schroeder, Gustavo S. Mesch, William H. Dutton, Journal of Computer-Mediated Communication 13 (2008) 769–793
The patterns of information search and access are found to have no power law distribution on the Web scale, but this result is not consisted in the case of small clusters of expertise within a narrower scale. There are three propositions of arguments on what patterns of access to information over the Web: Power Law (winner-take-all); Non Power Law (egalitarian effect of search engines); Benkler’ complex interpretation (The Wealth of Network).
To investigate the extent that the use of alternative search technologies decentralizes access to scientific knowledge, Caldas et al (2008) focus on quantitative methods for analysing the Web in terms of the structure of hyperlinks among Web resources. Together with the qualitative interviewing approach, they test two hypothesis: (1) different search engines result in different outcomes; and (2) centrality, connectivity and subgroup structure can be used to identify regularities and patterns of the Web.
Two hypothesis have been tested positively on the Web scale through comparing six main search engines (Google, Yahoo, MSNSearch, AskJeeves, Gigablast, and ScholarGoogle) with a set of three keywords for each six global topics (climate change, poverty, HIV/AIDS, terrorism, trade reform, and Internet and society). Regularities of results in Web Networks show that the average distance among reachable pairs is small, and the density within link blocks is relatively low. Therefore, they view this fragmented nature of these Web network as the evidence of “democratization” of the Web. Moreover, the significant clustering process, which defines a clique as a subset of the network with at least 3 nodes interlinked, is the main reasoning for the “reinforcement (clique) effects and not the power law effect.