"J. Scott Armstrong: Fewer than 1 Percent of Papers in Scientific Journals Follow Scientific Method"
Maybe academic researchers and those who love them shouldn't read this. It's as discouraging as discouraging can be about research.
I would guess that many papers fail to satisfy the third requirement listed by Professor Armstrong: "Full disclosure of methods, data and other relevant information". A small bit of good news is that, at least in economics, that seem to be (slowly) improving.
UPDATE: see Armstrong's "Guidelines for Science: Evidence-based Checklists" for a more general argument. (Thanks to Mike for the comment.)