Abstract
Sample size calculations before conducting a health study or clinical trial are important to provide evidence that the proposed study is capable of detecting real associations between study factors. This review aims to clarify statistical issues related to the calculation of sample sizes and is illustrated with an example of a recent study design to improve health outcomes related to water and sewage in NSW Aboriginal communities. The effect of power, significance level and effect size on sample size are discussed. Calculations of sample sizes for individual-based studies are modified for more complex trial designs by multiplying individual-based estimates by an inflationary factor.
