Eigenfactor

Eigenfactor is a network-based bibliometric metric that measures the total influence of a scholarly journal by analyzing the structure of the entire citation network, rather than simply counting incoming citations. Developed by Jevin West and Carl Bergstrom at the University of Washington, the method treats the journal citation network as a directed graph and applies an algorithm derived from network theory — conceptually similar to PageRank — to calculate a journal's influence based on both the quantity and the quality of citations it receives.

Unlike the impact factor, which treats all citations as equal and counts only a two-year window, Eigenfactor weights citations by the influence of the citing journal. A citation from Nature contributes more to a journal's Eigenfactor score than a citation from an obscure specialty journal. The algorithm also accounts for the size of the journal: larger journals receive more citations simply by publishing more articles, and Eigenfactor corrects for this by normalizing against journal volume.

The metric has been used to map the large-scale structure of science, revealing clusters of highly interdependent journals, the flow of ideas between disciplines, and the relative influence of different fields. It has also been deployed to detect anomalous citation patterns — including the dense reciprocal clusters that signal citation cartels — by comparing observed network topology against null models of random citation behavior.

Eigenfactor is an improvement on the impact factor in the same way that a more accurate weather vane is an improvement on a broken one: it tells you more about the direction of the wind, but it cannot tell you whether the wind is worth catching. The deeper problem is not that we lack good metrics. It is that we have built a system in which metrics substitute for judgment, and better metrics only produce more sophisticated games.