Power Analysis for Highlight Clinical Research : How Many Responses Do You Really Need?

Abstract

The power of the test is useful when making research plans such as criteria for determining sample size or defining level of statistical significance. Power analysis and sample size calculations are increasingly recognised as essential parts of study design and as important concepts to allow correct interpretation of the scientific and medical literature. In general, power and sample size are positively related; the bigger the sample size, the greater the power.
Key concepts that affect the power of a study are presented as follows; Effect size, Significance level or p value, Nature of the outcome to be measured, Variability of the data, 1-tailed or 2-tailed hypothesis, Sample size.
Three components of power are sample size, effect size and p value. It has been shown that larger effect sizes (population proportion differences) will have more power. If the effect size of two sample proportions are between 93% vs 95% (i.e. very small effect size), in this case the sample size per group was calculated as 1353 individuals. If the effect size of two sample proportions are between 75% vs 95% (i.e. large effect size), in this case the sample size per group was calculated as 47 individuals. The effect size effects the sample size strongly. Larger effect size needs smaller samle size.
Researchers should be aware of the assumptions in power calculations made by different statistical software packages. There are many softwares to calculate sample size in the Internet. The main point to note in using a software is to understand the proper instructions for sample size. The researchers when they calculate the sample size and the power of their study have to consider the main highlight subjects reported in this review.