Metadata and Citation Network of Retracted Publications in PubMed (1950–2025)

This dataset provides a comprehensive collection of PubMed IDs (PMIDs) for retracted articles and their associated citation networks, covering the period from 1950 to early 2025. The data was compiled to support scientometric research on the persistence and impact of retracted literature in the global scientific record. The dataset consists of three primary components: Retracted PMID List: A curated list of unique identifiers for all articles indexed as "Retracted Publication" in PubMed. Citation Network (Edge List): A structured source-target list representing citations from the scientific literature to these retracted articles. Retraction Metadata Links: Detailed mapping between original articles and their corresponding retraction notices, errata, or expressions of concern. Methodology: Data was harvested using custom scripts (R/PHP) interacting with the NCBI Entrez Programming Utilities (E-utils). The dataset is particularly useful for researchers in scientometrics, research integrity, and network analysis.