Lies, Damned Lies and Statistics

"There are ways for scientists (and journal editors) to protect themselves against their own biases in order to make sure they’re only publishing valid papers, but perhaps the more important thing is for the general public to realize that the headlines (which trumpet conclusions) don’t always match what’s really going on. If we can teach people to look behind the headlines and look at methodology and data, we’d all be better off."

http://www.patheos.com/blogs/friendlyatheist/2015/08/22/a-statistically-significant-conclusion-isnt-always-an-accurate-one/