December 5th 2012 ,
room: FENS L065
High-impact Research and Scientific Misconduct in Biomedical Sciences
Murat Çokol, Sabancı University
An enormous collection of biology related research articles is deposited in several databases, such as Medline. In this work, we combined large-scale data mining tools with statistical analysis to understand several aspects of knowledge acquisition in biomedical communities.
Our study indicates that high-impact research in biomedical sciences results when popular molecular interactions are connected with new interactions, and when popular biological concepts are connected with unpopular chemicals and drugs. We present a quantitative model for scientific misconduct and the related phenomenon of article retractions. Application of real world data to this model finds that scientific misconduct occurs at a similar rate in high or low impact scientific journals, however “wrong” articles in high impact journals are more frequently detected by the scientific community.