Full summary over here: http://blogs.discovermagazine.com/notrocketscience/2012/03/10/failed-replication-bargh-psychology-study-doyen/
Basically, John Bargh, a researcher at Yale, conducted a psychological experiment to see if by talking about old age he could prime people to feel the effects of age (aka walking slower). This was back in 1996.
Just recently, Stephane Doyen attempted to replicate this study, but with blinding in some units, something that the original study did not have. Which ended up showing that the individuals walked slower "only when they were tested by experimenters who expected them to move slowly". A basic case of blinding revealing unconscious bias.
Bargh then wrote a blog post where he proceeded to throw a fit.
Mind you, it's the sort of gentlemanly like fit that you see in academic circles, but most obviously a case of a grown man loosing it because he wasn't right. He attacks the journal, the authors, and Ed Yong as well. The last was a particularly dumb move, since Yong is probably the best damn science journalist out there, and wasn't exactly going to let it go at that.
This all illustrates that scientists aren't really different than other people when it comes to territoriality, and that replication, /published/ replication studies, are needed now more than ever. Until a study is replicated, it's a sample size one.