Researchers are finding that previously indisputable findings are losing their statistical significance. Read about it in the Dec. 13 issue of the New Yorker: http://www.newyorker.com/reporting/2010/12/13/101213fa_fact_lehrer
The implications are incredibly disturbing–an extreme interpretation would indicate that the scientific method is meaningless, and all of nature is a version of the “butterfly flapping its wings in China and causing a snow storm in New England” trope. I suspect this article is itself somewhat of an overstatement of experimental findings, though. And any experiment involving a person’s reaction to something is going to be subject to all kinds of complex, hidden factors–that most antipsychotic meds fail to replicate success over time does not surprise me.