© BRYAN SATALINOLooking for patterns of potential bias in scientific studies, a Stanford University–based research team found a number of risk factors. Among more than 3,000 meta-analyses, small studies that were highly cited were more likely to contain bias, as were studies authored by scientists with a history of misconduct or by small but global research teams. On the other hand, the study, published in PNAS yesterday (March 20), found no association between bias and the authors’ country of origin giving incentives to individuals for performance, refuting the idea of a “publish or perish” environment.
“To the best of my knowledge, all the evidence that we have about pressures to publish comes from surveys, i.e. what scientists say. Now, there is no question that these pressures exist, we all feel them, and it is reasonable to suspect that they might have negative effects on the literature,” coauthor Daniele Fanelli told Retraction Watch. “However, as scientists we should verify empirically whether these concerns are justified or not. And, to the best of my current understanding, as explained above, evidence suggests that they are partially misguided.”
Fanelli and colleagues collected data from more than 3,000 meta-analyses, and looked for correlations among various characteristics, such as authors’ retraction history, citations, career level, and gender, as well as studies’ effect sizes. Scientists who had at least one retraction were more likely to overestimate effect sizes, as were small studies and those published in the “gray literature,” such as PhD theses and conference proceedings.
“Our results should reassure scientists that the scientific enterprise is not in jeopardy, that our understanding of bias in science is improving, and that efforts to improve scientific reliability are addressing the right priorities,” the authors wrote in their study.