Derek Robinson of Scientific Update gave a lecture entitled “Uses and Abuses of Statistical Experimental Design” reviewing a number of papers from the literature showing that people often do not make best use of the possibilities of DoE, often using far more complex designs than necessary.  With the benefit of hindsight one can show that much simper designs requiring far fewer experiments can give just as much information.  On other occasions researchers try to optimise too many variables rather than screening out unimportant variables and then carrying out detailed optimisation work.