In a new debate published in BMC Medicine, Peter Bachetti talks about the concept of statistical power in determining sample size, the undue weight that is given to power calculations in clinical research, and the way that power calculations can actually contribute to less transparent reporting.
Randomised controlled trials are one of the highest forms of evidence-based clinical research, and can heavily influence treatment options in the clinic. However, as with most research, they are only as good as their basic design. Sample size is considered to be a key parameter in deciding whether a study is credible, and this decision is most often based on a statistical calculation of power. However, is this really “the best” way of determining sample size?
Bachetti advocates looking to other parameters to influence the determination of adequate sample size, including information already available and choices based on cost and feasibility.
Visit BMC Medicine to read the full debate and gain additional insight into this interesting and controversial topic. While you’re there, why not contribute to the debate or sign up for article alerts, which will help keep you informed of the latest publications in general medical and clinical research?