Bibliographic coupling networks reveal the advantage of diversification in scientific projects

Scientific research project serves as a fundamental part of supporting research progress, the data of which can reflect the research behavior of scientists. We use data from the acknowledgements of papers published by the American Physical Society (APS) and focus on the patterns of the successfully established scientific research projects. We find that the number of articles published by each project and the number of projects acknowledged by each article both obey the power-law distribution. After constructing each project’s bibliographic coupling network, it can be found that there is sublinear relationship between project’s core research fields and project scale. We explore factors affecting projects’ outcome and find the advantage of diversification in scientific projects, i.e., the dispersion of research fields can lead to higher ratio of papers in top journals and higher originality of papers. Meanwhile, evolution of research fields indicates that a typical project tend to study the core topics first, and then explore other possible research directions.