Abstract

Although uncommon, some professionals have published research where either the data or the results were fabricated. If there are several authors involved in such research, how can the primary source of the research misconduct be identified from among multiple authors? Use of correlation or regression analysis is one possibility, along with analysis of variance, for predicting research anomalies from level of authorship across multiple authors and articles. The procedures are demonstrated with a simulated set of data. Results indicate that the proposed methodologies could help identify authors primarily responsible for fabrication of data or results, even for medium-to-large samples. Awareness that responsible authors could be identified from among several co-authors of published articles may help discourage research practices involving misconduct. Co-authors, even if they are not directly responsible for such misconduct, should be careful lest they become unexpectedly associated with unethical research behavior.

Keywords: Research misconduct; Fabricated research; Research ethics; Data anomalies; Multiple authorship; Statistics; Research methods