Sample size and power calculations made simple （样本容量和功率计算简单）.pdf

ResearchResearch methodologymethodology seriesseries Sample size and power calculations made simple Evie McCrum-Gardner Background: Sample size estimation is an important concern for researchers as guidelines must be adhered to for ethics committees, grant applications and publications. Studies may be underpowered (too few participants) or overpowered (too many participants) so it is important to achieve the correct balance. Content: The process of sample size calculations, including relevant definitions, is explained and clear examples for different study designs are provided for illustration. A range of software packages and websites are discussed and evaluated. Conclusions: The information regarding sample sizes and statistical programs should be useful for researchers in order to perform sample size calculations for research projects. Key words: n minimal clinically important difference n power n sample size n significance level n type I error n type II error Submitted 10 July 2009, sent back for revisions 23 September 2009; accepted for publication following double-blind peer review 2 November 2009 ample size estimation is an important con- ReseaRch ethics committees cern for researchers undertaking research Guidelines projects, but is