Relative Positions of Countries in the World of Science

A novel picture of the relative positions of countries in the world of science is offered through application of a two-dimensional mapping method which is based on quantity and quality indicators of the scientific production as peer-reviewed articles. To obtain such indicators, different influential effects such as the background global trends, temporal fluctuations, disciplinary characteristics, and mainly, the effect of countries resources have been taken into account. Fifty countries with the highest scientific production are studied in twelve years (1996–2007). A common clustering algorithm is used to detect groups of co-evolving countries in the two-dimensional map, and thereby countries are classified into four major clusters based on their relative positions in the two-dimensional map. The final results are in contrast with common views on relative positions of countries in the world of science, as demonstrated by considering some examples like USA, China or New Zealand. The proposed method and results thereof might influence the concept of ‘scientific advancement’ and the future scientific orientations of countries. Quantitative analysis of the productions of research units (countries, academic institutions, or research groups) has turned out to be one of the most significant issues since the early days of scientometrics. An important part of that analysis is devising methods and indicators to evaluate the multitude of aspects of such a complex production process, especially, its ‘quantity’ and ‘quality’ [e.g. 1, 2]. This quantitative assessment is becoming inevitable 2 today, since such methods and findings are invaluable for policy-makers of research units as well as academic agents directly involved in the process, to evaluate their achievements, and modify their approaches for further progress. As a consequence, a vast amount of research has been devoted to the development of such methods and indicators, and their critiques. One of the essential problems in such analyses is that to provide a holistic perspective from scores of different indicators – each representing a certain aspect of the scientific output – usually, a host of graphs and tables are offered, from which one can hardly grasp a comprehensive picture and an ultimate judgment [e.g. 3, 4, 5, 6]. As a way out of this problem, composite indicators [7] and two-dimensional maps have been introduced [8, 9, 10]. Along the same path, in this article, we have propounded a method to construct a 2-dimensional map from two composite indicators – the first based on publications and the second based on citations thereof – aiming at a simultaneous representation and analysis of ‘quantity’ and ‘quality’ of the scientific productions of countries. This 2-dimensional map is meant to indicate the relative positions of countries compared to each other. However, certain adjustments should be made in advance to take into account influential effects such as the background global or discipline-characteristic trends as well as temporal fluctuations, and mainly, the effect of different amount of resources accessible to countries. Regarding this, we have recognised the effect of human and financial resources upon those relative positions, and have shown that their elimination will provide utterly different positionings which are in sharp contrast with the common views reflected in the national or international reports on scientific production of countries [e.g. 3, 4, 5, 6]. In this study, the topmost 50 countries in scientific production are considered in a period of twelve years (1996– 2007). Since this period will certainly admit temporal fluctuations in positions, we have implemented a common statistical clustering method (k-means algorithm) to analyse the 2-dimentional map (which is unbiased by the resource effects). The finest clustering scheme has been selected by inspection of clustering validity indices. The final clusters (labelled by letters A, B, C, and D) – the objective of this study – indicate the relative positions of country groups compared to each other. We have deliberately avoided the common convention of proposing some clear-cut numbers as ‘ranks’ or ‘position markers’ for countries, mainly because there is actually an insignificant difference between countries belonging to a neighbourhood in the range of values of an indicator; moreover, those numeric ranks prove to be volatile due to inevitable and non-negligible temporal fluctuations.