Revisiting the exploration-exploitation behavior of scholars' research topic selection: Evidence from a large-scale bibliographic database

• This paper proposes five novel research strategies under the exploration-exploitation behavior, and presents corresponding metrics to quantify and identify them. • The paper discloses the relationship between scientists’ research performance and their preference for research strategies, and analyzes the evolution pattern of the preference. • Results show that eminent scientists tend to follow academic frontiers, study diverse topics, explore emerging topics and innovative topic combinations, but exploit mature topics less often. We also figure out the potential reasons for the phenomenon. • Results show that successful scientists prefer to execute exploratory research strategies from the beginning of their career, and young scientists seem to be more creative. The research on studying exploration-exploitation behavior in topic choice has consistently been the focus of a great deal of attention. In this study, we propose five novel research strategies under exploration and exploitation based on the general but significant features of topics, and present a series of metrics to quantify and identify these strategies. We analyze the relationship between scientists’ research performance (i.e., productivity and impact) and their preference for different strategies, and examine the evolution of their preference in scientific careers through comprehensive statistical analysis. We employ a MAG dataset as our data source, and select about 30 million scientists from the computer science filed and their publications as our analysis objects. Our empirical analysis shows that productive and impactful scientists tend to follow academic frontiers, study diverse topics, explore emerging topics and combinatorial innovation, but exploit mature topics less often. We also figure out the potential reasons for the phenomenon. In addition, we find that successful scientists prefer to execute exploratory research strategies from the beginning of their career, and young scientists seem to be more creative. Our research may help researchers deeply understand topic selection behavior, and therefore provide enlightenment for training scientists and give advice for funding allocation as well as research and development policy formulation.