That's what Michael Shermer put out in the October Scientific American:
Over the past three decades I have noted two disturbing tendencies in both science and society: first, to rank the sciences from "hard" (physical sciences) to "medium" (biological sciences) to "soft" (social sciences); second, to divide science writing into two forms, technical and popular. And, as such rankings and divisions are wont to do, they include an assessment of worth, with the hard sciences and technical writing respected the most, and the soft sciences and popular writing esteemed the least. Both these prejudices are so far off the mark that they are not even wrong.
If you cannot tell a good story about your data and theory -- that is, if you cannot explain your observations, what view they are for or against and what service your efforts provide -- then your science is incomplete. The view of science as primary research published in the peer-reviewed sections of journals only, with everything else relegated to "mere popularization," is breathtakingly narrow and naive.
I like his argument. It is really hard to resist the pressures to make peer-reviewed articles dry and lifeless. Writing for review is too much like lawyering before a hostile jury.
As a reviewer, I find the really good research articles to be the ones with the really good literature reviews. If an author really knows the subject, it will show in the citations. Of course, this particular information may be irrelevant for an "integrative" article, depending on the level. A well-written, well-referenced article with a clear test of a hypothesis is like the golden fleece -- it's widely considered to be impossible, so few people even try!