Ratio and regression estimation use additional information that is known about the population to improve the accuracy of estimators. Unlike stratified sampling, where the information is used in the survey design, ratio and regression estimation use the information to modify the estimators. A common practice is to calibrate a survey to known population quantities, which guarantees that estimates from the survey agree the population statistics for those quantities. Calibration is often used to adjust for survey nonresponse, as will be discussed in Chapter 8.