Publication Type

Journal Article

Version

submittedVersion

Publication Date

12-2018

Abstract

We model the nonresponse probabilities as logistic functions ofthe outcome variable and other covariates in the survey sampling study withcallback. The identification aspect of this callback model is investigated. Semiparametricmaximum likelihood estimators of the parameters in the responseprobabilities are proposed and studied. As a result, an efficient estimator ofthe mean of the outcome variable is constructed using the estimated responseprobabilities. Moreover, if a regression model for conditional mean of the outcomevariable given some covariate is available, then we can obtain an evenmore efficient estimate of the mean of the outcome variable by fitting the regressionmodel using an adjusted least squares method based on the estimatedunderlying distributions of the observed values. Simulation results show theproposed method is more efficient compared with some existing competitors.The method is applied to data from a survey of health spending in a populationof individuals aged 50-70 years, where non-response can may be related tohealth.

Keywords

Auxiliary information, Calibration estimation, Followup, Logistic regression, Nonignorable nonresponse, Paradata

Discipline

Econometrics | Economic Theory

Research Areas

Econometrics

Publication

Scandinavian Journal of Statistics

Volume

45

Issue

4

First Page

962

Last Page

984

ISSN

0303-6898

Identifier

10.1111/sjos.12330

Publisher

Wiley

Copyright Owner and License

Authors

Additional URL

https://doi.org/10.1111/sjos.12330

Share

COinS