Research on Imputation Estimation Methods for Items Non-response Error in Dual Sampling Frame
Non-response error is inevitable in sample surveys,even in dual sampling frame.In this paper,the bias of the estimator caused by the items non-response error in sample surveys in dual sampling frame has been corrected.First,the information of auxiliary variables is obtained through two-phase sampling,and the combined estimator of the constructed ratio estimator and the exponential ratio estimator is used to impute the items non-response data in the sampling survey in dual sampling frames,and the mean estimates of the corresponding sub-populations are obtained.The population value is estimated using the form of the Hartley estimator.By calculating estimator bias,mean square error and optimal weight coefficient,and comparing with the relative accuracy loss of the mean square error of the same type of combined estimator and the relative accuracy loss of the single ratio-type exponential estimator under the same condition,the results of random simulation show that the loss rate is low and the imputation method is effective.Since the appropriate auxiliary variables are selected to construct the combined estimator of the ratio estimator and the exponential ratio estimator as the imputation values,the information of the auxiliary variables and the answered research variables is more fully utilized,the proposed combined estimator has certain application value in sampling survey work.
NETL
NSTL
万方数据
