Previous studies have found that nonresponse varies for different demographic characteristics of households. The impact of nonresponse on estimates from Federal household surveys (e.g.; The Current Population Survey) will be examined by comparing estimates from one period depending on response or nonresponse from an adjacent period. This technique is a form of "gross flow", which has been used to study potential bias in survey estimates in other studies. The differences in estimates will be examined relative to household characteristics which have been found to differ in nonresponse. This will provide a measure of the degree of bias due to nonresponse as well as a measure of differential bias for subgroups of the populations. Some of the household characteristics of interest include ownership status, rural/urban environment, and demographic characteristics of the household.