The impact of fraudulent research on the scientific literature. The Stephen E. Breuning case

Abstract

The goal of this study was to determine the research impact of scientific fraud through citation analysis of 20 Breuning publications, using the 1980 to 1988 Science Citation Index and Social Sciences Citation Index. These publications received 200 citations, of which 80 (40.0%) were self-citations by Breuning or his coauthors. Tracked over time, non--self-citations declined sharply in 1986 and later years, coinciding with disclosure of Breuning's fraud. The data indicated that, in this case, researchers effectively shunned work known to be or even suspected of being falsified. Unique citation contexts (101) were examined to see how citing authors used Breuning's work: 33 were negative (disagreed with findings/methods), 10 positive (agreed), and 58 neutral (no valuation). Also, 63 were inconsequential (no influence on the citing author's analysis/conclusion). Thirty-eight were material, but 21 of these led to negative conclusions. These data diminish the apparent impact of Breuning's work suggested by total citations alone.

KIE: This study is an effort to determine what impact articles that report fraudulent data have on research. The authors evaluated citations to 20 publications by Stephen Breuning, who was convicted of scientific fraud in 1988. Breuning's work had a very high citation rate, but few authors who cited his research agreed with his findings or methods. Garfield and Welljams-Dorof argue that their data indicate that, in Breuning's case at least, fraudulent research is not influential in the field, and is shunned by authors once it is exposed.