Survey statistics

A survey increases the weights of respondents in each age group to account for nonrespondents in the same group. Which assumption most directly supports interpreting this as a correction for nonresponse?