Beyond surface correlations: Reference behavior mediates the disruptiveness-citation relationship

Abstract

ABSTRACT Purpose This study examines why papers with high CD indices (measuring research disruptiveness) increasingly show reduced citation impact and investigates whether this represents genuine impact reduction or methodological artifacts. Design/methodology/approach We analyzed 29 million papers (1950-2016) using Poisson regression to examine relationships between the CD index and citation count, with controls for fields, team size, and reference count. Findings Papers with high CD indices showed reduced citation impact over time. However, when controlling for increasing reference counts in papers, this relationship reversed, revealing a positive association. Papers with more references exhibit lower CD indices owing to the index’s sensitivity to the reference count, while achieving higher citation counts. Alternative innovation metrics consistently show positive correlations with citation impact. Research limitations The approach may not adequately capture the reduced citation impact of highly disruptive papers with fewer references. The analysis is limited to journal articles and shows correlation rather than causality. Practical implications The apparent undervaluation of disruptive research stems from methodological artifacts in the CD index calculation driven by evolving reference patterns. Researchers should control for the reference count when using this metric. Originality/value This study reveals that bias against disruptive research is specific to the CD index’s calculation methodology, identifying reference behavior as the key factor affecting the relationship between disruption metrics and citation impact.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 5 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.