The Occasional Pamphlet reflects on the new megajournal trend in open access: "A ray of sunshine in the open-access future". PLoS ONE is being joined by SAGE Open and Scientific Reports from Nature Publishing Group.
The mega-journal trend means that strong traditional publishers with name recognition are entering open-access publishing in a big way. They’ll be hard pressed to trot out their hackneyed canards (vanity press, disenfranchisement). And these journals will provide coverage of a huge swathe of academic fields. Between SAGE Open, PLoS ONE, and Scientific Reports, essentially all of the social sciences, humanities, and natural sciences are covered. In addition, the breadth of these journals means that they will be competing for the same pool of articles. Authors will have a choice between submitting papers in genetics, say, to PLoS One or G3, in physics to AIP Advances or Scientific Reports, and so forth. Publishers will have to compete in order to attract authors, either on price or publisher services or both. They’ll have to market these journals to authors, using their intellectual capital to convince authors that OA journals (at least their OA journals) are a Good Thing. As authors and promotion committees get used to using the new article-level metrics (as they already increasingly are, with download counts and h-index), journal brand name — whether of these mega-journals or traditional journals — will become less important, and authors will feel freer to publish in these and other OA journals, again based on publisher services rather than journal brand name.
I find G3 very interesting. A journal devoted to "emphasis on data quality and utility, rather than detailed mechanistic insight or subjective assessment of immediate impact." It seems to me that the theme reflects the reality of a division in the field between data generators and data consumers. The advantage is that it enables recognition and citation for data generation without requiring long-term working relationships between groups. The drawback is obvious -- why the heck are we generating these data, if we're not going to seriously test hypotheses about mechanisms?