Articles dans des revues à comité de lecture
|Absolute and specific measures of research group excellence|
|Mryglod O., Kenna R., Holovatch Yu. , Berche B.|
|Scientometrics 95 (2013) 115-127|
|DOI : 10.1007/s11192-012-0874-7|
|ArXiv : arxiv:1210.0732 [PDF]|
A desirable goal of scientometrics is to introduce, if it exists, a simple and reliable way to measure the scientific quality of publicly-funded research institutions and universities to serve as a basis for their ranking and financing. While citationbased indicators and metrics are easily accessible, they are far from being universally accepted as way to automate or inform evaluation processes or to replace evaluations based on peer review. Here we consider extensive measurements of research strength at an amalgamated, institutional level and intensive measures of average performances per head, such as research quality. Using biology research institutions in the UK as a test case, we examine the correlations between extensive and intensive citation-based indicators as measures of research strength and quality, respectively. We find that citation-based indicators are very highly correlated with peer-evaluated measures of group strength but are poorly correlated with group quality. Thus, and almost paradoxically, our analysis indicates that citation counts could possibly form a basis for deciding on how to fund research institutions but they should not be used as a basis for ranking them in terms of quality.