Rank-normalized journal impact factor as a predictive tool

Abstract

Citation data accumulated on articles from the top and bottom 25 of impact factor (IF)-ranked international journals are compared using 59 international geoscience journals from 1998 and 378 Polish geological papers from 1989-994. There is a minor risk of being uncited when results are published in high-IF periodicals as the average non-citation rate is 0.88 over a 10-year period in this not very rapidly developing scientific discipline. Similarly, the established error levels in the prognosis of expected citation success versus failure based on the extreme IF quartiles as an evaluation tool is low (at most 12.5). Thus the application of the rank-normalized journal IF as a proxy of real citation frequency and, accordingly, as a predictive tool in the a priori qualification of recently published publications is a rational time- and cost-saving alternative (or at least a significant supplement) to traditional informed peer review. Blanket criticism of using IF for decisions in research funding is therefore at least partly exaggerated.