this article by Jonah Lehrer in the New Yorker about the decline effect, from December 2010. (For those of you who prefer to listen to your news, Radiolab did a short podcast called "Cosmic Habituation" summarizing the article here.)
Of course, I sent the New Yorker article along to my lab group, and my advisor (Aaron Clauset) responded with a link to an even better summary of p-values and the resulting bias in the scientific literature by Carnegie Mellon professor Cosma Shalizi, entitled "The Neutral Model of Inquiry (or, What is the Scientific Literature, Chopped Liver?)."
These collection of articles and posts summarize the related issues of false positives, stochastic effects resulting in positive results that may or may not be significant, and the file-drawer problem.
But really, I like my educational experiences to be thoroughly hilarious, as well. So, ladies and gentlemen, I refer you to the second link in Prof. Shalizi's post, The Story Behind the Atlantic Salmon. Because, really, if you can't derive significant scientific discoveries from dead ichthyological specimens purchased at a local grocery chain, how are any of us to make sense of this curious existence? Enjoy.