Measuring the indirect journal impact

Résumé

We define a new index of journal impact, Outerfactor, that roughly measures the probability of using a journal by a "random" researcher. Other eigenvector-based measures are usually characterized as stationary probabilities of a Markov chains. We provide a unified framework that allows us to characterize these measures in the same setting. The Outerfactor score obtained by a journal is independent of the citations it makes and of its article share. In the core of Outerfactor is the correction of journal impact by an indirect self-impact coefficient. We illustrate our results with 140 economics journals from JCR 2010. We prove that Outerfactor is robust to rank manipulation and to the inclusion/exclusion of self-citations in the data.