Ordered Random Selection Without Replacement
A. F. Bissell · Journal of the Royal Statistical Society Series C (Applied Statistics) · 1986
Bebbington (1975) describes a method for drawing a simple random sample without replacement. Briefly, if the population contains N items of which n are required in the sample, dummy variables N' and n' are initially set equal to N and n. The ratio n'/N' is calculated and compared with a random (0, 1) number, say, x1. If x1 n'/N') N' is reduced, n' remaining at its previous value. The ratio n'/N' is recalculated and compared with a new random number x2 to determine whether the second item in the population is to be included in the sample. The process continues, reducing n' and N' each time an item is included, but reducing only N' when an item is passed over. Further selections are automatically excluded as soon as n items have been identified for inclusion in the sample. The method guarantees that