This is another type of restricted random sampling in which the initial unit of the sample is selected at random from the initial stratum of the universe, and the other units are selected at a certain space interval from the universe arranged in a systematic order like, numerical, alphabetical, or geographical order. For this, the size of the sample is determined first on some basis, and then the space interval of the universe is determined by the following model:
K = N/n, where K is the space interval of the universe of the universe form which samples are to be taken, N the size of the universe and n the size of the sample.
Having determined the two important factors, n and K as cited above, the initial unit or the sample will be taken at random from the initial space interval (stratum), and the other units will be selected at every Kth point of the universe counting from the point of the initial unit thus taken out for the sample. Thus, if there are 100 students from whom 10 are to be selected for the sample, all the 100 students will be arranged first, in serial order form 1 to 100, and then the space interval will be determined by K = N/n as 100/10 = 10. After this, one student will be selected first at random from the first 10 students. If his serial numbers comes out to be 9, then the serial number of the other 9 students to be selected for the sample would be 19,29, 39, 49, 59, 69,79,89, and 99. The sample would then consist of the 10 students bearing the serial numbers; 9, 19, 29, 39, 49, 59, 69, 79, 89, and 99.
Note. If the values of K comes out in faction it should be approximated to the nearest minimum integer shown as under:
K = 100/9 = 11.11 approximated to 11
K = 100/15 = 6.67 approximated to 6
K = 110/20 = 5.5 approximated to 5