Although important steps have been taken to prevent publication of a disproportionate number of non-reproducible chance findings, null findings are usually still considered disappointing. There is every reason to change this perception, because lack of associations can teach us just as much as significant ones.