Fig. 4From: Estimating the prevalence of text overlap in biomedical conference abstractsThe distribution of similarity scores assigned by eTBLAST across all abstract pairs. Text similarity (eTBlast) identifies 20,857 abstracts similar to at least one other conference or published abstract. We used random samples of all such comparisons to estimate rates of text overlap – same meeting (2%); text overlap – same conference (3%); and text overlap - MEDLINEBack to article page