Independent review

For the title of several magazines, journals and newspapers, see Independent Review.

Independent review of scientific results is the cornerstone of scientific accountability. By giving other scientists access to one's own data and methods, and giving them an opportunity to gauge the reproducibility of one's results, one ensures that errors due to incompetence, unconscious bias, or other causes can be found by others. It's the scientific version of proofreading.

An important part of the independent review process is anonymous peer review of scientific papers before they are published in scientific journals.

Sometimes researchers will bypass the pre-publication review process (see "science by press conference") or will refuse to share their data and methods with other scientists. In general, the scientific community frowns on this, as it makes it difficult or even impossible for other scientists to verify the data and interpretations of the research.

Another possible reason for bypassing traditional peer review is when reporting results which are radically at odds with mainstream scientific views. In several historical cases, discoveries announced in this way have eventually reached the mainstream (see Semmelweiss and his theory of an "invisible substance" infecting women after childbirth; see also Continental drift).

Quotes

• The best means for assessing risks and benefits is through independent review of the proposed research by individuals who have no direct vested interest in its outcome.
• A central tenet in the protection of research participants is the independent review of research protocols to assess their scientific merit and ethical acceptability.
• All protocols involving human participants should undergo an independent and rigorous scientific review to assess scientific quality, the importance of the research to increase knowledge, and the appropriateness of the study methodology to answer a precisely articulated scientific and, in some cases, clinical question. For example, the design of clinical trials should be based on sound statistical principles and methodologies, including sample size, use of controls, randomization, population stratification, stopping rules, and the feasibility of relating endpoints to objectives.

See also

• Independent medical review