On allocation procedures using systematic sampling

The properties of stratified sampling can differ considerably by using systematic sampling in each stratum, as opposed to the commonly used simple random sampling within strata. Several authors have worked on stratified sampling using simple random sampling; this is known as stratified simple random sampling with relatively little attention given to systematic sampling. This research is concerned with the development, analysis and implementation of a different class of stratified sampling called stratified systematic sampling. We investigated proposed estimation method under two methods of allocation of units within stratum and tested its performance with the existing method in three populations using the standard error, coefficient of variation and design effect. We also examined the gain and relative gain in precision of the design technique considered coupled with the efficiency of the proposed estimation methods with the existing methods in terms of variance ratio. The results of the analysis revealed that the proposed estimation methods irrespective of any methods of allocations is more precise and efficient than the existing methods and the Neyman stratifi ed systematic sampling performed the best in gaining variance reduction in all three sets of data.