Abstract:
|
Wages once were the main, and typically, the sole source of income for most people, and they have always been one of the best-measured income sources in household surveys. However, in the last decades of the 20th century, more people received self-employment income and dividend and interest payments, potentially challenging the ability of statistical organizations to maintain the accuracy and quality of income information collected in household surveys. Item response rates provide information about the quality of specific elements within a survey, yet few item response rates are routinely calculated or published. One reason is that the appropriate response concepts are not obvious and must be developed. In this paper, we define and calculate conceptually appropriate income item response rates for the various components of income using the March Current Population Surveys for 1990 and 2000. We find that response rates for all income components fell and that the proportion of income that was imputed rose over the decade.
|