Working Paper: CEPR ID: DP7887
Authors: Pedro Albarrán; Ignacio Ortuno-Ortín; Javier Ruiz-Castillo
Abstract: Albarrán et al. (2009a ) introduced a novel methodology for the evaluation of citation distributions consisting of a pair of high- and a low-impact measures defined over the set of articles with citations below or above a critical citation level CCL. Albarrán et al. (2009b ) presented the first empirical applications to a situation in which the world citation distribution in 22 scientific fields is partitioned into three geographical areas: the U.S., the European Union, and the rest of the world. In this paper, we compare our results with those obtained with average-based indicators. For reasonable CCLs, such as the 80th percentile of the world citation distribution in each field, the cardinal differences between the results obtained with our high-impact index and the mean citation rate are of a large order of magnitude. When, in addition, the percentage in the top 5% of most cited articles or the percentage of uncited articles are used, there are still important quantitative differences with respect to the high- and low-impact indicators advocated in our approach when the CCL is fixed at the 80th or the 95th percentile.
Keywords: citation distributions; measuring scientific performance
JEL Codes: I23; I32
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| high/low impact indicators (E01) | better representation of citation distributions (C46) |
| high-impact index (C43) | improved understanding of scientific performance (O36) |
| high-impact indicators (C80) | more accurate understanding of scientific performance (D29) |
| mean citation rates (A14) | misrepresentation of scientific performance (C90) |