Abstract
This paper discusses unresolved key problems in social sciences research and publishing. Researcher bias can lead to the publication of studies with highly questionable conclusions, which can impact adversely on important social topics (e.g., the influence of heredity on intelligence, criminality). Outright fraud occurs in science, and social sciences in particular [30]. To help counter the prevailing publication bias in research, four editorial suggestions for scientific reporting are proposed: (1) Before the work starts, investigators would be required to submit a written description of their research methodology. An international social sciences evaluation registry would be created to monitor research activity. (2) The Editor's decision about whether to publish a research study would be based on a paper that omitted the results and conclusions sections. These sections would be included in the published paper after an editorial decision was made. (3) With these appropriate scientific safeguards, researchers would also be required to provide raw data to other interested scientists, especially when the study is socially important, or is difficult to replicate. (4) Journal Editors who greenlight socially relevant studies would provide extra publishing space to people whose views differ significantly from the principal investigators. Bias and distortion cannot be completely eliminated from the scientific publication process. Adoption of these four proposals, however, would reduce scientific reporting problems such as data dredging and the dissemination of unsupported conclusions, and would also help to reduce fraud. Such adoption would eventually lead to the publication of improved scientific research in the social sciences.