23 May 2003

If you are like most scientists, chances are you frequently obtain negative results to a study. Do you discuss them with friends? Let them languish in a drawer? If you would like to publish data that refute a hypothesis, that go against your assumptions, or that simply, maddeningly, suspiciously did not turn out as expected even though you had all the right controls, then you may want to turn to the Journal of Negative Results in Biomedicine. Launched by BioMed Central last fall, peer-reviewed, and indexed by PubMed, this free online publication seeks to fill a gap well-known to any researcher who has tried to publish such data in conventional scientific journals.

Perhaps you would like to show colleagues that an antibody the field widely assumed to recognize a specific epitope actually binds to something else, or perhaps you want to help prevent wasteful duplication by demonstrating that a particular method is ill-suited to address a given question. You may want to report results of a trial that showed no advantage over standard therapy. These sorts of negative data are an integral part of science, which, according to Karl Popper, advances through a process of "conjectures and refutations."

In an inaugural editorial, the journal’s editor-in-chief, Bjorn Olsen and editor Christian Pfeffer, both at Harvard Medical School, describe now-famous examples of scientific discoveries that did not see the light of day in their time because they went against prevailing hypotheses. "The purpose of the journal is to have all types of scientific results published and discussed so that we can most optimally move science forward. Sometimes results may be negative only in the context of current thinking and experience. Those so-called negative results have not gotten the publicity they deserve," said Olsen.

But how does he distinguish valid negative results from poor science? "Poor science is not based on reproducible experiments and careful analysis of all possibile interpretations-the trivial as well as the nontrivial. Its experiments frequently lack proper controls. Our peer review makes absolutely certain that weak science is not published in the journal," Olsen said.

A negative result that comes out of a rigorous experiment could mean that the field is going in the wrong direction. How can the field consider this if this result is not published? "The field needs valid negative results," Olsen added. Olsen himself studies bone, cartilage, and blood vessel development, and genetic disease thereof. Does his journal welcome submissions from neurodegenerative disease research? "Absolutely!"—Gabrielle Strobel