Measurement error in the assessment of finite population parameters

This book deals with Estimation of finite population parameters under super population model when variates are subject to measurement error.Following Royall (1970, 71), and Royall and Herson (1973a, 1973b) approach, estimators of the finite population total under error-in-variables super population model have been developed. This has been extended in stratified sampling under error-in-variable super population model where slopes (regression coefficient) are varying from stratum to stratum and when slope across the strata is same and response variable is subject to measurement error.A limited simulation study has been carried out with Synthetic data generated under the models to examine the effect of measurement error on the precision of the estimate.The overall results show that % R.I. in standard error of the estimator can be reduced to a considerable extent if stratified sampling is resorted to Neyman allocation. It has been found that % R.I. in standard error increases with increase in the value. There would be no % R.I. in standard error.

Autorentext

Dr. Amar Singh is presently working as Assistant Professor (Agricultural Statistics) at CSSS PG College, Machhra, Meerut, U.P., India.Dr. Bhim Singh is working as Associate Professor (Statistics) at SVPUAT, Meerut, U.P., India.Dr. BVS Sisodia is an Ex-Professor & HoD of Agricultural Statistics, ANDUAT, Kumarganj, Ayodhya, U.P., India.