The Disruption Index (DI) is a significant metric for identifying research that expands scientific frontiers and pioneers new fields. Unlike readily accessible metrics like citation counts, DI requires complex analysis of citation patterns, specifically, how subsequent research that cites a given work also references its foundational predecessors, representing a paradigm shift in assessing scientific impact. Current DI studies remain largely confined to single indicators, disciplines, or databases, lacking comprehensive benchmarks to evaluate the intrinsic properties and temporal dynamics of diverse DI metrics. To bridge this critical gap, we introduce Cross-source Disruption Indexes (CrossDI) dataset, a comprehensive benchmark resource that integrates multiple established DI metrics for key articles across four fields. Curated from three major bibliographic databases (WoS, Dimensions, and OpenCitations), this dataset is designed as a reusable benchmark for the systematic evaluation and comparison of disruption indexes.
Xu et al. (Wed,) studied this question.