Distortions in Scientific Literature -- A Replication Analysis of Greenberg’s Citation Network of 302 Alzheimer’s Science Research Papers

Scientists use ideas and facts from earlier research, which they cite to show how their paper builds on prior knowledge. However, errors and overstatements can spread through citation. Greenberg (2009)’s study describes how misleading information spread in Alzheimer's science research by analyzing a citation network of 302 papers. This fall, we replicated what Greenberg did with the same network, and tried to make improvements. We used Python, Pajek, and Excel, to investigate three main citation distortions: citation bias, amplification, and invention. Citation bias is caused when authors greatly prefer citing supportive papers instead of citing both supportive and critical papers at the same time, which generates an unbalanced network. Amplifications are caused by not citing primary data papers directly. Invention is created by altering the meaning when authors cite papers, such as citing a “fact” from a paper which only stated a “hypothesis”. This spring we are working on automating parts of the analysis, for instance to better visualize the change in the citation network over time. This research was completed in collaboration with the Department of Statistics, College of Liberal Arts and Sciences and the School of Information Sciences at the University of Illinois at Urbana-Champaign.Ope