Precision of any estimate made from a sample depends both on the method by which the estimate is calculated and on the plan of sampling. In this research work, we examined efficiency of two-phase sampling for regression estimator using data on cassava production (sprout ability and root weight) for 2004 planting season at IITA research site, Mokwa, Niger State, Nigeria. A large first-phase sampling without replacement of size 87 taken from a population of 100 sprout ability, (x) and a second-phase sub-sample of 13 of the initial 87 first-phase sample was then drawn and both y and x were measured. The results show that there is a strong and positive correlation coefficient of 0.934 between the auxiliary variable and the response variable. Efficiency of two-phase sampling was illustrated using different values of the correlation coefficient, ρ and arbitrary cost ratios. We found that efficiency increased with smaller values of λ and higher values of ρ. We also determine when two-phase sampling would be preferred to single phase sampling. When = 0.9340 , = 50, = 5 the inequality gives 0.8724 > 0.3306 in which case, two-phase sampling would be preferred to single stage sampling plan. Hence, two-phase sampling for regression provides a good methodology for estimating the cassava yield with the knowledge of the sprout ability which in-turn leads to better cassava production for sustainable development. This study shows that the method of two-phase sampling would be a powerful and economical technique in survey sampling as it yields a gain in precision

Projects

Currently no objects available

Currently no objects available