[stiff]

Peter Norvig wrote a good article on this:

"Warning Signs in Experimental Design and Interpretation" http://norvig.com/experiment-design.html