A new analysis indicates that rushing to avoid being scooped affects the research methods scientists adopt, Science reports.
Researchers from Eindhoven University of Technology and Arizona State University applied an evolutionary agent-based model to examine the effects of rewarding scientists who publish first on the overall scientific research enterprise. The simulated scientists, Science notes, had different strategies: some collected large datasets and some stopped pursuing a line of inquiry if scooped, but others would soldier on. This then carried on for generations of scientists varying levels of rewards given for publishing first being tested.
As reported in Nature Human Behaviour, the model suggests high rewards for publishing first leads to populations of scientists who publish on smaller datasets, which, Eindhoven's Leonid Tiokhin and Arizona State's Minhua Yan and Thomas Morgan note, reduces research reliability. But the researchers also found possible approaches to combat this effect such as increased startup costs as well as increased rewards for secondary studies or publishing negative results.
"Effective interventions to improve scientific practice require a causal understanding of the forces that shape scientists' behaviors," the trio writes in the paper. "Our model takes one step towards this goal."