Nonparametric regression estimators in dual frame surveys

Abstract

Dual frame surveys, in which independent samples are selected from two frames to decrease survey costs or to improve coverage, can present challenges for regression estimation because of complex designs and unknown degree of overlap. In this research, we developed three nonparametric regression estimators in dual frame surveys and investigated their asymptotic properties. Simulation results show that all the proposed methods work well.

Keywords:

• Cross validation
• Dual frame surveys
• Nonparametric regression estimator
• Prediction error
• Simulations

Acknowledgments

The authors thank the referees for their careful reading of the manuscript and would like to express sincere appreciation for their insightful and helpful comments.