Abstract
The use of calibration estimation techniques in survey sampling have been found to improve the precision of estimators. This paper adopts the calibration approach with the assumption that the population median of the auxiliary variable is known to obtain a more efficient ratio-type estimator in estimating population median in stratified sampling. Conditions necessary for efficiency comparison have been obtained which show that the proposed estimator will always perform better than the existing asymptotically unbiased separate estimators in stratified random sampling. Numerical evaluations have been carried out through simulation and real-life data to compliment the theoretical claims. Results from the simulation study carried out under three distributional assumptions, namely the chi square, lognormal and Cauchy distributions with different sample settings showed that the new estimator provided better estimate of the median with greater gain in efficiency. In addition, result from the real-life data further supports the superiority of the proposed estimator over the existing ones considered in this study.
Publisher
Granthaalayah Publications and Printers
Reference15 articles.
1. Deville, J. C. and C. E. Sarndal, “Calibration estimators in survey sampling”, Journal of the American Statistical Association, vol. 87, pp. 376-382, 1992.
2. Sarndal, C. E, “The calibration approach in survey theory and practice”, Survey Methodology, vol. 33 pp. 99-119, 2007.
3. Singh, S. and R. Arnab, “On calibration of design weights”, International Journal of Statistics, vol. 69, no. 2, pp. 185-205, 2014.
4. Clement, E. P. and E. I. Enang. “On the efficiency of ratio estimator over the regression estimator”, Communication in Statistics: Theory and Methods vol. 46, no. 11, pp. 5357-5367, 2017.
5. Clement, E. P., “Calibration approach separate ratio estimator for population mean in stratified sampling”, International Journal of Modern Mathematical Sciences, vol. 13, no. 4, pp. 377-384, 2015.
Cited by
2 articles.
订阅此论文施引文献
订阅此论文施引文献，注册后可以免费订阅5篇论文的施引文献，订阅后可以查看论文全部施引文献