Skip to Main Content

Science is hard. Anyone who disagrees should take themselves to a desert island and try to build a gene sequencer from scratch.

But is there a difference between inherent difficulty and superfluous complexity?

A new analysis in the journal PLOS ONE has found that research papers in the life sciences have become increasingly dense over the past quarter-century. They have higher page counts, more — and more detailed — figures and tables, longer lists of authors, and richer appendices of supplementary data. The authors of the paper gathered all of those elements into an “average publishing unit,” which they found has doubled in the past two-plus decades.


Although all that additional information packed into an individual study might be a good thing for science, the authors suggest that it could be too much strain for the system to bear. Peer reviewers are, after all, practicing scientists themselves with their own deadlines to meet; scrutinizing a paper always takes a while, but scrutinizing a longer one, well, takes longer. “Increasing reviewer workload could translate into information overload for reviewers and a reduction in the quality of the peer-review process,” the authors of the new study wrote.

And there are other reasons it might not be so positive, too. Because the paper is the chief unit of output in science today, the rising density and complexity of articles likely indicates that the costs of production of that unit also are rising. Which, in turn, means that researchers are devoting more resources to publishing papers even as the availability of grant funding is diminishing: Moore’s law turned on its ear.


Meanwhile, the inclusion of more authors could reflect greater collaboration — but it also opens the door to unethical researchers who see a chance to disguise their misconduct amid the valid data of others. (This fear isn’t hypothetical; the authors of the new study cited the case of Joachim Boldt, a German anesthesiologist who had more than 90 papers retracted for ethical concerns and data chicanery.)

The new paper, by a group from Brazil and the United States, is more observation than prescription. And the authors acknowledged that what they’ve seen is consistent with impressive gains in technology, clinical medicine, and other areas that have made science so advanced, as well as in computer software and hardware that enable the increasingly detailed generation and display of data. (In case anyone’s wondering, the authors said they found no evidence that journals are driving these trends by demanding more from manuscripts.)

One refreshing reaction to the data bloat has been a movement in the opposite direction, toward the scientific equivalent of short stories — or even just scenes — rather than epic novels.

ScienceMatters is a new initiative launched by a pair of researchers who want to give investigators an outlet for publishing single observations quickly. The project is a rebellion against one part of the problem for which journals and their editors are certainly responsible: The thirst for compelling and sexy narratives on which data is hung.

“Storytelling has thus become the prevailing paradigm of scientific publishing,” the site’s founders, Lawrence Rajendran and Mirko Bischofberger, told us recently. “But this emphasis on storytelling is accompanied by further problems such as delayed reporting of major observations, non-reporting of ‘inconvenient’ facts and orphan observations, and a strong publication bias.”

If scientists don’t have to create a major work of literature every time they wanted to publish a paper, there’s no reason for delay, nor for omitting that inconvenient data, said Rajendran and Bischofberger. “It also allows scientists to publish those intriguing observations that might otherwise lie forever unpublished, whether for lack of resources or because they cannot currently be explained,” they said.

And researchers can still publish papers in big journals; they just need to gather various findings in ScienceMatters and cite them.

Slow down. Savor the data. If artisanal cheeses have a market, why not artisanal science? Small plates, of course.