Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
EFFECTS OF THE PROPOSED POVERTY MEASURE 285 We are in full agreement with the recommendation that SIPP become the basis for the nation's official poverty and related statistics. The March CPS does not collect all of the information needed for poverty measurement, has problems with the quality of the information that it collects, and does not have much room for further improvement. In contrast, SIPP collects most of the needed information, has achieved quality improvements, and, because of its focus on income, has ample opportunity for further improvements in both the scope and the quality of income-related data. The best time to put this recommendation into effect would be in 1996, when other changes to the survey are made. Orienting SIPP to Poverty Measurement A decision to use SIPP to produce the official poverty data means that all aspects of the survey should be reviewed to determine their suitability for providing the most accurate statistics possible under the proposed measure. A key aspect for review is the proposed redesign of the survey. Although the Census Bureau has accepted many of the recommendations of the CNSTAT Panel to Evaluate SIPP, it has decided against the recommendation for a design that would have two panels of about 27,000 households each in the field each year, with new panels introduced every 2 years. Instead, the Census Bureau has proposed a design that would have one large panel of 50,000 households in the field each year, with new panels introduced every 4 years. The Census Bureau's design has the advantage of maximum sample size in a single panel for purposes of longitudinal analysis. For cross-sectional analysis, the two designs are equivalent: the two panels in the field each year under the CNSTAT SIPP panel's design can readily be combined to produce the same sample size as the single, larger panel of the Census Bureau's design. Longitudinal estimates are important, but we believe that the time series of annual poverty rates and other statistics is paramount and that the design must support the production of reliable annual estimates. In this regard, the Census Bureau's proposed design provides no overlap between panels. Hence, every 4 years, it will be hard to determine if changes in the poverty rate are real or due to the introduction of a new panel in place of an old panel that may have uncorrected attrition bias or other problems.25 Since most attrition of sample cases from SIPP occurs by the end of the first year of a panel, there may be problems of attrition bias with the CNSTAT SIPP panel's design as well as the Census Bureau's, as the former does not 25 Attrition bias can occur when attrition rates differ between groups: for example, higher rates of attrition for low-income people could produce a downward bias in the poverty rates. Adjustments to the survey weights are usually made to compensate for attrition bias, but the adjustments may not be adequate.
EFFECTS OF THE PROPOSED POVERTY MEASURE 286 refresh the sample for cross-sectional estimates more frequently than every 2 years. Research on attrition and the most appropriate corrective actions is obviously needed, whichever design is used, and the Census Bureau has stated its commitment to such research for SIPP. However, it is still the case that attrition bias or other problems with a panel that may affect the poverty estimates cannot be fully assessed with a nonoverlapping design. Indeed, a nonoverlapping design also limits the possibility of using SIPP for longitudinal analysis of important policy changes, such as changes in the welfare or health care systems. Ideally for such analysis, one wants information for a sufficient length of time before a change in order to accurately characterize people's behavior under the old policy regime. One then wants information for as long as possible after the policy change to assess the effects on behavior. However, if policy changes take effect near the beginning or end of a 4-year panel under the Census Bureau's design, information either before or after the change will be limited, reducing the ability to adequately evaluate the effects. In contrast, under the design of the CNSTAT Panel to Evaluate SIPP, there will likely always be a panel in the field that is suitable for analysis of before-and- after effects, albeit with a smaller sample size. In addition to considering the best survey design for purposes of poverty measurement, the SIPP questionnaire should be reviewed to determine what changes may be required. Thus, some questions may need to be added at least occasionally (e.g., work expenses) or asked more frequently (e.g., child care expenses or child support payments), while others may need to be modified. In some cases, such as the estimation of tax liabilities, it may make sense to collect a limited set of variables that will enhance the Census Bureau's simulation model rather than to try to collect detailed information directly.26 Finally, from the perspective of improved poverty measurement, we urge that high priority be given to several areas of methodological research for SIPP. First, questionnaire research should be pursued to develop ways to improve the quality of reporting of wage and salary income in SIPP, which falls short of independent estimates (very likely because many people report net rather than gross pay). Second, research should be conducted to improve the weighting process so that the weights adequately account for the higher rates of attrition evidenced by low-income population groups (see Appendix B on both these points). Third, and very important, research should be conducted to improve population coverage in SIPP. A problem that affects all household surveys, including SIPP and the March CPS, is that not all people who are associated with sample households are in fact listed as household residents. Particularly subject to undercoverage are low-income minority groups. For example, it is 26 See Citro and Kalton (1993:Chap. 3) for suggestions of content changes to SIPP that generally comport with the proposed resource definition for the poverty measure.