What model and assumptions are used in the Sample size calculator?
We use the T-Normal distribution model simplified for use with central means (that is the 50% of the population extends in either direction from the mean) and with two-sided confidence intervals. This model is generally used in business to produce sample sizes for populations with unknown distributions and standard deviations (essentially a rough but generally suitable approximation useful for common business applications). The elements within the population are expected to behave in a consistent and “smooth” pattern without excessive clustering following the Normal bell curve distribution. The distribution pattern is expected to follow the shape in the graph as shown in Figure 1 below. Figure 1. Bell Curve Normal distribution shape. The mean value is depicted at the origin z = 0 with the central density inside of the two-sided confidence interval segments excluded. This model requires two primary inputs (apart from the population size) that define how large the sample should be to mee
