Also known as statistical power, procedures that test the likelihood that differences in data are due to true difference, rather than chance or random error, and that they can thus be inferred from the sample studied to the total population.