Raking

Raking (also called raking ratio estimation) is a post-stratification procedure for adjusting the sample weights in a survey so that the adjusted weights add up to known population totals for the post-stratified classifications when only the marginal population totals are known. The resulting adjusted sample weights provide a closer match between the sample and the population [Page 672]across these post-strata than the original sample. Raking, however, assumes that nonrespondents in each post-stratum are like respondents. Nonetheless, when implemented with care, raking improves the mean squared error of sample estimates.

The term raking is used to describe this statistical technique because the raking ratio—the ratio of the population total (or control) total for a given post-stratum to the marginal row (or column) total from the sample for ...