A pair of papers calls for researchers to take greater risks when choosing research problems to study in order to speed up scientific discovery and innovation. The two papers, one from PNAS and another from the American Sociological Review, used computer models to analyze millions of research papers.

The study published by PNAS found that research choices in biomedical journals become more conservative over time, eventually focusing upon a key group of chemicals. The authors believe this method is inefficient for science as a whole. They call for increased risk-taking and the publication of experimental failures to improve the speed of discovery.

Original research paper published in PNAS on November 9, 2015.

Original research paper published in the American Sociological Review on September 1, 2015.

Names and affiliations of selected author for both papers

James A. Evans, Department of Sociology, Computation Institute, University of Chicago, Illinois, U.S.A.