Sample Size Calculations:
Sample size calculations typically arise in significance testing, in the following context: how big a sample size do I need to identify a significant difference of a certain size? The analyst must specify three things: 1) How big a difference is being looked for (also called the effect size; 2) What is the alpha level of the test (how big a Type I error is tolerable), e.g. alpha = .05, and 3) The desired power of the test, or how certain do you want to be of catching a significant difference, if there is one, e.g. power = .8.