In the early 1990s, researcher Kevin Dunbar embarked on a particularly fortean project: investigating what scientists do with all that 'damned data' they produce. Wired reports on his findings.

Thanks to David Pecotic for the link.