1979: Sequential Sample Selection Methods

Sequential sample selection methods are distinguished from conventional methods in the manner in which random numbers are used to determine the sample. Conventionally, a sample of size n is selected from a sampling frame of N sampling units by selecting n random numbers and mapping then into n of the labels in the sampling frame. Sequential methods require each sampling unit in the sampling frame to be considered in order and a probabilistic decision reached concerning its inclusion in the sample. Sequential methods can be implemented efficiently on computers and are particularly adaptable to selecting samples as a file is read or as a process takes place. A thorough discussion of several equal probability sequential selection procedures is given by Fan, Muller, and Rezucha (1962). Their method 1 for selecting a simple random sample of size n from N involves sequentially comparing conditional probabilities of selection for each unit labeled i = 1,2,...,N with a uniform random number