Indicadores de input/output de la ciencia iberoamericana: ¿cuán similares son las clasificaciones basadas en los indicadores de RICYT y Scimago?

Abstract

The objective of this study is to evaluate the congruence between the country classifications obtained from the Scimago Journal and Country Rank (SCIJCR) indicators, and those provided by the RICYT, taking as study units 11 Ibero-American countries, during 2006-2017. Thirty-four input/output indicators were taken as variables, 11 from SCIJCR, and 23 from RICYT. The similarity relationships among the countries and the indicators were represented by means of phenograms (Ward's method) and the congruence among the classifications of the countries was represented by strict consensus trees and quantified by means of a consensus index. The main conclusions of the study indicate that: 1) The classification of countries based on the 34 indicators corresponds to their size (e.g., socioeconomic development, population) and to the respective scientific traditions; 2) The indicators show a complex grouping patterns, not observing groupings based on the different typologies (e.g., production, impact, input, context); 3) The vast majority of the SCIJCR indicators show close mutual links, producing redundant information; 4) The percentage of international collaboration is only related to values ​​of moderate similarity with the citations per document, so it does not agree with the idea that the number of citations is directly proportional to international collaboration; 5) Taking into account the results obtained, the most profitable investments in terms of production, impact, and impact and production, are those measured by graduation indicators (graduates), RD 6) The lack of congruence observed when comparing country rankings from only one source (i.e., SCIJCR or RICYT) contradicts the simplistic idea that scientific results can be predicted only from invested resources; 7) Comparing the classification of countries based on all the indicators with those produced based on one or the other (i.e., SCIJCR or RICYT) produces few common groups. This can be explained from issues intrinsic to the analysis, such as the different number of indicators, and the redundancy of the information provided by the vast majority of the SCIJCR. This means that the SCIJCR indicators have a lower weight than those of the RICYT, when it comes to differentiating groups of countries.