A sociologist argues on page 395 of this week’s issue of Science that making scholarly articles available online has narrowed citations to more recent and less diverse articles than before–the opposite of what most people expected. (…) Oddly, “our studies show the opposite,” says Carol Tenopir,(…). She (…), have surveyed thousands of scientists over the years for their scholarly reading habits. They found that scientists are reading older articles and reading more broadly–at least one article a year from 23 different journals, compared with 13 journals in the late 1970s.
I turned to the article on “Citation Communities” in Berkshire’s Encyclopedia of Community (Christensen et al., editors, Sage 2004) to get the background, and also talked to scholars who understand better than I exactly how important citations are in the academic world. My personal concern is that people should know about similar work being done – because we learn from one another, and because the idea of vast amounts of research money being spent on duplicate efforts appalls me. But scholars care abaout citations for practical reasons, too: being cited can help them when they have job reviews, and raises their status in their academic department and knowledge community.
I have been thinking about the need for updates to the Encyclopedia of Community, given the many changes that have taken place since 2004. This is certainly one article that needs to take account of the changes of the last six years, and to assess the research that’s been done so far to determine the effects of online publishing and online academic social networking services (which bubble up now and then: I take a look, am awed by the grant money they’ve managed to get hold of, and never hear about them again). I’m posting the article “Citation Communities” from the Encyclopedia of Community here, for your reading pleasure. Some cool subheadings, by the way: “Intercitation and Cocitation” and “Intellectual Versus Social Ties.” Comments are most welcome.
“Citation Communities” from the Encyclopedia of Community (Christensen et al., editors, Sage 2004)
Citation communities are abstract networks that are built up as authors cite other authors in the footnotes, endnotes, and bibliographies of learned literatures, especially when citations are made repeatedly. These communities contain not only acquaintances but persons the citer has never met, the dead as well as the living. Their chief significance is that they are reasonably objective manifestations of key evolutionary units of science and scholarship, whether these are called invisible colleges, discourse communities, schools of thought, disciplinary specialties, theory groups, or cultural networks (Randall Collins’s phrase, “coalitions of the mind,” may capture them best). As such, they are of interest to historians of ideas, to sociologists who study communication patterns among scholarly elites, and to information scientists who exploit bibliographic ties among authors to improve document retrieval.
Citation occurs when the author of one document refers explicitly to another document. The intent is generally to relate some new claim to a context set by one or more precedent texts. By convention, citations identify what is being cited by means of a few standardized details. Author and title identify a work; the other details serve to identify editions of works, copies of which are the actual units of document retrieval.
Networks of Names
Citation relations can be represented as graphs in which the works are nodes and the citations are links that connect some nodes but not others. Citing and cited documents can thus be rendered as networks, making graph-theoretical research on networks, especially social networks, applicable to citation data. Human communities are often studied as networks of relations between pairs of persons, and citation communities can similarly be studied by examining relations between pairs of documents. Documents, after all, are in some sense surrogates for the persons who wrote them, and, just as the attributes of persons can be used in social analysis, the attributes of both citing and cited documents can be used in analyzing citation communities.
Usually, however, the term “community,” however, usually refers not to a set of documents, but to a group of persons with shared attributes. The term “citation community” thus implies that the linked authors are to be understood as persons, whether they appear as citers or citees. Hence, for studying community, the key attribute of documents is authors’ names, with their interesting duality of meaning. Names can be shorthand for documents, but they also evoke people. An analyst can write “E. O. Wilson” or “S. J. Gould” and mean two books or two papers in a network, but she or he can also mean two famous contemporary biologists. Using the latter interpretation, the analyst can add to the study of documents any relations that might hold between their authors as persons, such as whether they knew each other or worked in the same place or exchanged e-mail. The analyst can ask such questions as What other marks of community do these authors share? Does their relatedness through citation correlate with particular social ties or communication behaviors? Do any of these other variables cause citation? Does citation cause any of them? Is the community formed by intellectual ties or by social ties? Does it exhibit lines of conflict? The answers depend on the other variables used to interpret the citation ties.
Intercitation and Cocitation
One way to study citation communities is to start with authors as the unit of analysis. In practice, this means analyzing not individual documents but collections of authors’ works (for example, any writing by Wilson or Gould). “Relatedness through citation” can then be defined in two basic ways. In a citation-network grapha numeric matrix with identical lists of authors on the rows and in the columns the cells can be filled with intercitation counts or cocitation counts (both are available through online searches of the Institute for Scientific Information’s citation databases in academic or special libraries).
Intercitation reflects who cites whom in a closed setfor instance, does Wilson anywhere cite anything by Gould? Does Gould likewise cite anything by Wilson? Since their choices need not be the same, intercitation is an asymmetric measure. The researcher can create a binary matrix of ones and zeroes (for “cites/doesn’t cite”) or a valued matrix, so called because it contains the actual frequencies of intercitation in the cells (the higher the count, the more important the tie).
The same can be done with cocitation counts, except that here the valued matrix reflects the number of documents in which any two authors have been jointly cited by citers in generalfor example, in how many documents have they cited anything by Wilson with anything by Gould? This matrix is symmetricthe count for Wilson-Gould is the same as that for Gould-Wilson. Matrices containing either intercitation or cocitation counts yield standard network statistics (e.g., measures of which authors are most highly connected) and can be tested for correlation with other measures of community.
Intercitation is a good focal variable if the analyst has other measures of the citers as actors and wants to explore their community in terms of why they cite (or fail to cite) each other. Cocitation is good if the analyst simply wants an empirical snapshot of who goes with whom in a field, to look for explanations such as sameness of subject matter, methods, or nationality. Author cocitation has been extensively used to map authors in “intellectual space,” but relatively few studies have used cocitation or intercitation counts to probe the social space of research communities. This area seems ripe for development, especially given the software now available (e.g., UCINET). Analyzing and visualizing change over time in these communities will probably be highly valued in the future.
Intellectual versus Social Ties
Studies to date leave little doubt that citation communities are based primarily on intellectual ties, such as the perceived relevance and authority of the citees. Social ties, such as acquaintanceship or personal interaction, are secondary. They co-occur with citation often enough, but they are neither necessary nor sufficient to explain it. For example, intellectual ties such as perceived relevance explain the citation of authors whom the citer has never met and in many cases could never have met (as when someone today cites Aristotle).
Some scholars even so take a sardonic view of “authority” in citation relations, implying that citers annex whatever citees they think make their own works most persuasive, with little regard for their real relevance. Others think that citation communities are rife with self-promotion and cronyism at the expense of true merit. When hard data are examined, these darker interpretations seem unwarranted, but considerable new research is necessary to settle the matter.
Howard D. White
Baldi, S. (1998). Normative versus social constructivist processes in the allocation of citations: A network-analytic model. American Sociological Review, 63(6), 829846.
Borgatti, S. P., Everett, M. G., & Freeman, L. C. (2002). UCINET for Windows: Software for social network analysis. Harvard, MA: Analytic Technologies.
Chen, C. (2003). Mapping scientific frontiers: The quest for knowledge visualization. Heidelberg, Germany: Springer-Verlag.
Collins, R. (1998). The sociology of philosophies: A global theory of intellectual change. Cambridge, MA: Belknap Press of Harvard University Press.
Crane, D. (1972). Invisible colleges: Diffusion of knowledge in scientific communities. Chicago: University of Chicago Press.
Hargens, L. L. (2000). Using the literature: Reference networks, reference contexts, and the social structure of scholarship. American Sociological Review, 65(6), 846865.
Latour, B. (1987). Science in action; How to follow scientists and engineers through society. Cambridge, MA: Harvard University Press.
Mullins, N. C. (1973). Theories and theory groups in American sociology. New York: Harper & Row.
Mullins, N. C., Hargens, L. L., Hecht, P. K., & Kick, E. L. (1977). Group structure of co-citation clusters: Comparative study. American Sociological Review, 42(4), 552562.
Perry, C. A., & Rice, R. E. (1999). Network influences on involvement in the hybrid problem area of developmental dyslexia. Science Communication, 21(1), 3874.
Stewart, J. A. (1990). Drifting continents and colliding paradigms: Perspectives on the geoscience revolution. Bloomington: Indiana University Press.
Van Dalen, H. P., & Henkens, K. (2001). What makes a scientific article influential? The case of demographers. Scientometrics, 50(3), 455482.
White, H. D. (1990). Author cocitation analysis: Overview and defense. In C. L. Borgman (Ed.), Scholarly communication and bibliometrics (pp. 84106). Newbury Park, CA: Sage.
White, H. D. (2000). Toward ego-centered citation analysis. In B. Cronin & H. B. Atkins (Eds.), The web of knowledge: A festschrift in honor of Eugene Garfield (pp. 475496). Medford, NJ: Information Today (ASIS Monograph Series).
White, H. D. (2001). Authors as citers over time. Journal of the American Society for Information Science and Technology, 52(2), 87108.
White, H. D., & McCain, K. W. (1997). Visualization of literatures. Annual Review of Information Science and Technology, 32, 99168.