The calibration approach is a very general approach to reduce the sampling error and the nonresponse error in sample surveys. Nonresponse in sample surveys can be considered as a kind of sample selection problem. In the econometric literature, the Heckman two-stage procedure is suggested for handling sample selection problem in regression analysis. In this study, we mainly describe the calibration approach, the GREG estimator and the Heckman two-stage model. On the basis of the calibration techniques, we make use of the GREG estimator and the Heckman two-stage procedure to deal with the nonresponse problem, respectively. Via a simulation study, the two kinds of estimators are compared with each other, regarding variance and bias, and ratio of MSE.