An approach based on the geometric mean of basic quantitative and qualitative bibliometric indicators to evaluate and analyse the research performance of countries and institutions

TL;DR

This paper introduces a simple, geometric mean-based bibliometric index called the GRO-index for evaluating and comparing the research performance of countries and institutions, correlating well with established metrics.

Contribution

It proposes a novel, easy-to-understand method combining quantitative and qualitative indicators to assess research output at territorial and institutional levels.

Findings

GRO-index correlates strongly with h-index and other bibliometric indicators.

GROr and RROr provide field-specific and comparative insights.

GRO and RRO indexes are highly correlated with countries' wealth.

Abstract

We present a straightforward procedure to evaluate the scientific contribution of territories and institutions that combines the size-dependent geometric mean, Q, of the number of research documents (N) and citations (C), and a scale-free measure of quality, q=C/N. We introduce a Global Research Output (GRO-index) as the geometric mean of Q and q. We show that the GRO-index correlates with the h-index, but appears to be more strongly correlated with other well known, widely used bibliometric indicators. We also compute relative GRO-indexes (GROr) associated with the scientific production within research fields. We note that although total sums of GROr values are larger than the GRO-index, due to the non-linearity in the computation of the geometric means, both counts are nevertheless highly correlated. That enables us to make useful comparative analyses among territories and institutions. Furthermore, to identify strengths and weaknesses of a given country or institution, we compute a Relative Research Output count (RROr-index) to tackle variations of the C/N ratio across research fields. Moreover, by using a wealth-index also based on quantitative and qualitative variables, we show that the GRO and RRO indexes are highly correlated with the wealth of the countries and the states of the USA. Given the simplicity of the procedures introduced in this paper and the fact that their results are easily understandable by non-specialists, we believe they could become as useful for the assessment of the research output of countries and institutions as the impact factor is for journals or the h-index for individuals.