Correspondence factor analysis of the publication patterns of 48 countries over the period 1981–1992

Abstract

This study illustrates the application of a descriptive multivariate statistical method, Correspondence Factorial Analysis (CFA), to the analysis of a dataset of over 6 million bibliometric entries (data from ISI). CFA is used to show how the 48 most prolific nations stand in relation to each with regard to their publication interests in 17 specific disciplinary areas and one multidisciplinary field over the period 1981–1992. The output of a CFA is a map displaying proximity among variables (countries and disciplines) and constitutes an impartial working document for experts interested in the evaluation of science. The present study focuses on three aspects of a CFA: (1) The normalized “publication patterns” of countries with a common feature (e.g., that belong to the same geopolitical zone, economic union, etc.) can be pooled in order to highlight the position of the union with respect to individual countries; (2) complex CFA maps can be simplified by selecting reference countries or disciplines and observing how the remaining countries and disciplines relate to these references; (3) data on additional countries (new publication profiles) or on additional variables (e.g., socio-economic data on all the countries under study) can be introduced into the CFA maps used as mathematical models. Our CFA of the ISI dataset reveals the scientific interests of nations in relative terms. The main cleavage (the first factorial axis) is between countries that still concentrate on the disciplines of the industrial revolution such as physics and chemistry (or that have turned toward their offspring, materials sciences) and those that have veered toward more “modern” disciplines such as the life sciences (e.g., clinical medicine), the environment, and computer sciences. The second cleavage, along the second factorial axis, is between countries that focus on the agricultural sciences (the land surface) and those interested in the geosciences (the sea, earth's mantle, and mining). The third and fourth axes discriminate even further between earth, life, and abstract sciences highlighting the ostensible relationship between (organic) chemistry and all life science disciplines and between physics and disciplines related to engineering, materials sciences, etc. The CFA maps disclose the specific behavior of each country with respect to these cleavages. © 1996 John Wiley & Sons, Inc.