• Statistical data analysis
• reproducibility of results
• epidemiology

A recent article in this journal1 elicited strong response from readers, primarily due to the conclusions drawn that the groups were not different. In the letters that were sent in response, readers were critical of the lack of sample size and general methodology. Since the goal of medical research is to uncover findings that are biologically plausible, reliable and reproducible, it is important that the conclusions drawn are based on sound clinical and statistical methodology. Proper use of statistical methods goes beyond the execution of statistical tests and models, but also relies on their appropriateness given the study design and outcomes, as well as a sample size that can provide meaningful and reproducible results.

Many software packages and online calculators make it easy for power or sample size calculations to be performed and the topic has been covered in detail elsewhere.2 Depending on the situation, the effect size or other criteria about the expected results may be all that …