The existence of publication bias and risk factors for its occurrence

Abstract

Publication bias is the tendency on the parts of investigators, reviewers, and editors to submit or accept manuscripts for publication based on the direction or strength of the study findings. Much of what has been learned about publication bias comes from the social sciences, less from the field of medicine. In medicine, three studies have provided direct evidence for this bias. Prevention of publication bias is important both from the scientific perspective (complete dissemination of knowledge) and from the perspective of those who combine results from a number of similar studies (meta-analysis). If treatment decisions are based on the published literature, then the literature must include all available data that is of acceptable quality. Currently, obtaining information regarding all studies undertaken in a given field is difficult, even impossible. Registration of clinical trials, and perhaps other types of studies, is the direction in which the scientific community should move.