50% of studies can't be repeated with the same conclusions

[idoh]

http://marginalrevolution.com/marginalrevolution/2011/09/how-good-is-published-academic-research.html