Norvig's "Warning Signs in Experimental Design and Interpretation" is also worth reading and covers the higher level problem of bad research and results. Including mentioning bad statistics.

http://norvig.com/experiment-design.html