In the NYTimes, Siddhartha Mukherjee writes a very accessible article discussing experimental designs for experiment, retrospective, and prospective studies. Each approach has advantages and limits.
I find out the examination of personal memory in these studies most useful. Often contemporary records, or better, automated collection show that people are systematically biased in their memories. Our preconceived models influence what and how we recall.
Also of interest are
- the difference between statistical and practical significance.
- the problems encountered when subdividing the groups searching for effects
Read it at the Times,
HT to Alex Tabarok at Marginal Revolution. Alex goes on to note that cell phone use by drivers of cars is known to lead increased accident rates. The most likely way to damage your (or someone else’s) brain is the concussion after using your cell phone while driving.