Potential Pitfalls of Experimental Design
Good experimental design begins with the end in mind. An early conversation with a statistician will both increase the chances of an experimental study contributing to the literature and minimize the risks to participating human subjects. Sir R.A. Fisher felt that “to consult the statistician after an experiment is finished is often merely to ask him to conduct a post mortem examination: he can perhaps say what the experiment died of.” To this end, some questions from a statistician are presented along with the associated experimental study pitfalls to avoid during the study planning phase. Several concrete examples are provided to give some practical knowledge on how to improve an experimental study at the onset. Hypothesis formulation, sample size determination, randomization, and double-blinding are all explained from the viewpoint of a statistician’s final analysis. Confounders, sampling, and missing data are also briefly covered through this hypothetical question and answer session.
Dawson B., Trapp R.G. Basic & Clinical Biostatistics, 4th ed. New York, NY: Lange Medical Books/McGraw-Hill; 2001.332-343
Kang M, Ragan BG, Park JH. Issues in outcomes research: an overview of randomization techniques for clinical trials. J Athl Train. 2008; 43(2):215-21.
Nebe J, Heier M, Diener HC. Low-dose ibuprofen in self-medication of mild to moderate headache: a comparison with acetylsalicylic acid and placebo. Cephalalgia. 1995; 15(6):531-5
Moffet HH. Sham acupuncture may be as efficacious as true acupuncture: a systematic review of clinical trials. J Altern Complement Med. 2009; 15(3):213-6.
Dziura JD, Post LA, Zhao Q, Fu Z, Peduzzi P. Strategies for dealing with missing data in clinical trials: from design to analysis. Yale J Biol Med. 2013;86(3):343-58