Maximizing Regularity, Minimizing Predictability

systematic sampling is a method for selecting cases in such a way that they are distributed evenly through a series. The two stochastic systematic methods described in this article allow selection of cases that are separated by regular but not necessarily identical intervals. A method with linear probability thresholds defines a range of cases that are candidates for selection and chooses one, whereas a more sophisticated method uses logistic cumulative probability thresholds to select cases; the probability of selection increases near the center of a segment, but is nonzero at the edges. The present article will describe variations on a method for systematically sampling with a stochastic process. SAS code has been included to illustrate the process: SAS should not be hard to translate into any higher or lower level computer language.