young adults, males, minority groups, never-married people, poor people, and people with lower educational attainment.
In addition, more limited evidence suggests that the current noninterview weighting adjustments do not fully compensate for differential attrition across groups. One evaluation of the procedures to adjust for household nonresponse at each wave developed two sets of weights for Wave 2 households in the 1984 panel—one set based on all Wave 2 households and one set based just on those Wave 2 households that provided interviews at Wave 6. Comparing Wave 2 estimates from these two samples showed that the latter set produced higher estimates of median income and fewer households with low monthly income than those produced with the former set, evidence that the weights do not adequately adjust for higher attrition rates among low-income households (Petroni and King, 1988). A subsequent study that compared samples from the 1985 panel of all Wave 2 households and those that provided interviews at Wave 6 obtained similar findings (King et al., 1990).
With regard to annual estimates of poverty from SIPP, one study (Lamas, Tin, and Eargle, 1994) found that the inclusion of people with missing waves, using an imputation process, produced somewhat higher poverty rates than the use of complete reporters. Approximately one-sixth of the difference between annual poverty rates in SIPP and the March CPS is apparently due to attrition bias.
It is important to note that the current cross-sectional nonresponse adjustments in SIPP make only minimal use of the information that is available from previous waves for many current nonrespondents. Also, in constructing longitudinal files from SIPP panels, the Census Bureau assigns zero weights to original sample members who missed only one or a few waves in addition to those who missed all or most waves. The Census Bureau has recently committed itself to an intensive program of research to improve the weighting adjustments for attrition as part of the decision to move to 4-year panels for SIPP with no overlap (Weinberg and Petroni, 1992).
In addition to household and person nonresponse, there is substantial item nonresponse in the March CPS. The Census Bureau imputes as much as 20 percent of the total income in the CPS. For some income sources, imputation rates are even higher—as much as one-third of nonfarm self-employment income, interest, and dividend payments are imputed (Bureau of the Census, 1989a: Table A-2; Bureau of the Census, 1992b: Table C-1).
SIPP compares favorably with the March CPS on item nonresponse rates: overall, only 11 percent of total regular money income for 1984 was imputed in SIPP, compared with 20 percent in the March CPS. The SIPP and March CPS imputation rates for earnings were 10 percent and 19 percent, respec-