King of the hill: which journal is best?

A paper on the BioArxiv describes a new metric to measure scientific journals by based on their efficiency of information distribution (citations) within the network of journals. It provides a providing together a complex picture of the intricate relations between scientific journals; but basically Science, Nature and PNAS are the top 3 journals!

In the paper they use data from WoS's more than 35 million papers to show that several different hierarchies exist between scientific journals. Their networks of interconnected journals are not really connected; rather they are linked by citations of each other’s papers. They try to avoid a method that overtly avoids favouring “high-quality” journals with few higher quality papers, or “low-quality” journals with many low-quality papers.

The authors present two hierarchies; a flow hierarchy based on m-reaching centrality (flow), and a nested hierarchy based on tag hierarchy extraction (tag). These come to two slightly different conclusions about who’s top-dog (see figure above) with Science top of the flow hierarchy and Nature top of the nested hierarchy (I wonder how English or US news agencies will report this story). Many journals remain in similar positions in both rankings, although some are very hierarchy dependent e.g. Lancet and NEJM appear high in flow compared to tag, while Geophysical Research Letters appears at the top of tag but is not present at all in flow.

The authors state in the discussion that “providing an objective ranking of scientific journals…is a…complex problem of high relevance”; I’d agree this is a complex problem, but the relevance is very subjective, and looks to be changing.