Because of the widespread use of citations in evaluation, we tend to think of them primarily as a form of colleague recognition. This interpretation neglects rhetorical factors that shape patterns of citations. After reviewing sociological theories of citation, this paper argues that we should think of citations first as rhetoric and second as reward. Some implications of this view for quantitative modeling of the citation process are drawn.
Derek Price's theory of variation among the sciences stressed that the essential differences lay in the process through which scientists use each other's results. He maintained that the critical processes were those which took place within small groups of scientists who shared an intellectual focus, and proposed that an indication of those processes could be found in referencing patterns. Later research, reviewed in this paper, has corroboratedPrice in these observations. Several bodies of evidence point to the desirability of further application of the basic conceptsPrice introduced for the purpose he proposed: as diagnostic tools to describe and compare processes of knowledge growth in the sciences.
In order to attribute journals to specialties in a dynamic journal set by using aggregated journal-journal citations derived from theScience Citation Index, it is necessary to complement the multi-variate analysis of this data with a time-serices perspective. This calls for a more analytical approach to the problem of choice among the many possible parameters for clustering. Changes in the disciplinary structure of science are tracked by using thedifferences among the multi-variate analyses for the various years. It is impossible to attribute change systematically to structure, noise, or deviance if these uncertainties are not clearly definedex ante. The study discusses the various choices which have to be made, in both conceptual and methodological terms. In addition to hierarchies among journals, one has to assume heterarchy among journal groups (and their centroids). For comprehensive mapping, a concept of macro-journals as representations of a density of points in the multi-dimensinoal space is defined. Empirical reslts indicate the feasibility of dynamic journal-journal mapping by using these methods.
Growing income and wage inequality in a range of countries has raised concern. High-technology development may be contributing to this inequality, by encouraging higher wages at the upper end of the income distribution. Most studies of the possibility of this effect have used generic, aggregated data. In this paper, we introduce the possibility of linking wage inequality directly to specific industrial strategies using the Theil Index of inequality. This measure portrays the portion of wage inequality that is attributable to wages in specific industries. We illustrate this concept with data from U.S. states.