Efficiency evaluation of scientific data sharing platforms in serving research considering heterogeneity based on DEA: a case study of national science data centers in China

Scientific research is increasingly highlighting the significance of scientific data. This trend has led to the emergence of scientific data sharing platforms, which are tasked with managing scientific data and promoting the open sharing of data resources. The objective of this paper is to assess the efficiency of these platforms in supporting research, extending the analytical lens of research evaluation from conventional actor-centric evaluations to a platform-oriented perspective. We construct a two-stage efficiency evaluation model based on the Data Envelopment Analysis (DEA) method and propose a methodology for mitigating the impact of heterogeneity on evaluation outcomes through adjustment coefficient. A case study of China’s national science data centers is conducted. According to the calculation results and findings of this study, it is recommended that the national science data centers establish a differentiated resource optimization mechanism that prioritizes stage efficiency, which aligns with the institutional demand for a platform-based scientific research paradigm in the digital economy era. The national science data centers should establish subject-differentiated data service strategies. It is suggested to integrate the efficiency results of dynamic evaluation with the experience of benchmarking centers to conduct classification optimization for data centers, addressing technical deficiencies and management bottlenecks.