The shortage for post-graduate researchers, practitioners, and policy makers is in state-of-the-art reviews that are comprehensive and authoritative. For instance, the standards of many medical journals for reporting expensive clinical trials (CONSORT) recommend that authors "state general interpretation of the data in light of the totality of the available evidence." JAMA editors emphasized this commitment to quality by asking authors to use a checklist that includes this recommendation. Unfortunately, as Fytton Rowland pointed out last week, it is the research sponsor - for example, the NIH - not the journal, that calls the tune. A study reported at the International Congress on Peer Review held at Prague in 1997 showed little evidence that authors complied or that journal editors were able to insist on it. (I can supply cites for anyone interested.)
This shortage undermines authorship and credibility of grant proposals. It also casts a shadow of bias and insularity on research results. Insularity, of course, comes with the burden of too-many-to-count unevaluated, undistilled reports of primary research -- including journal articles as well as unreviewed preprints. One of my engineer friends calls this a "signal to noise problem." The greater the noise, the greater the energy must be to obtain a clear signal. At the risk of repeating the obvious, author-"archiving" preprints contributes more to the problem than to the solution. Newt Gingrich, speaking as a politician rather than a scientist, emphasized the policy implications of incoherence. However, the problem that he perceived has been recognized for decades as impairing the producitivity of research. It continues to fester as many scientists, like bureaucrats, prefer to work harder rather than smarter. Albert Henderson <70244.1...@compuserve.com>