The Survey of Income and Program Participation (SIPP) is a household survey of the noninstitutionalized resident population of the United States, conducted by the U.S. Census Bureau. The survey was designed to improve the measurement of the economic situation of persons, families, and households in the United States and to provide a tool for managing and evaluating government transfer and service programs.

The 1996 SIPP panel consists of 12 interviews, or waves, each of which gathered 4 months of retrospective data. The sample includes approximately 36,000 households, divided into four rotation groups. The SIPP has been matched to the Social Security Administration's (SSA's) administrative records. Receipt of disability benefits and the monthly benefit amount is based on December 1999 SSA records, and all other characteristics are taken from the SIPP. In December 1999, only three of the four SIPP rotation groups were interviewed. However, since rotation groups are assigned randomly, the sample is still nationally representative. SIPP sample weights are used to produce population estimates. The weights have been adjusted to reflect the fact that only three of the four rotation groups were used for the tabulations.

Poverty calculations based on the SIPP use survey-reported family income for 1999, with the survey-reported Social Security disability and SSI benefit amounts replaced by the Social Security disability and SSI benefit amounts from SSA administrative records. The poverty thresholds are provided in the SIPP and are adjusted for family size and composition. In 1999, the poverty threshold for a one-person family under age 65 was $8,667. For a two-person family aged 65 or older, the poverty threshold was $10,070. For a four-person family with two related children under the age of 18, the poverty threshold was $16,895.

In tabulations of Medicaid coverage, the SIPP Medicaid variable has been recoded to reflect the fact that SSI beneficiaries in certain states, sometimes referred to as Section 1634 states, are automatically covered by Medicaid. Thirty-two states plus the District of Columbia provide automatic Medicaid coverage for SSI beneficiaries, accounting for approximately 78 percent of all SSI beneficiaries.

SIPP and SSA counts of disabled beneficiaries differ because the SIPP excludes people living in institutions or outside the continental United States. In addition, the match rate between the SIPP and SSA records is not perfect. Furthermore, the SIPP estimates are based on a sample and therefore are subject to sampling error. Standard errors can be used to measure this sampling variability. Various sources of nonsampling error also may be important.

Because the SIPP uses a complex sample design, it is inappropriate to calculate standard errors assuming a simple random sample. Doing so will result in estimated standard errors that are biased downward. Readers interested in information about SIPP variance estimates are referred to the Source and Accuracy Statement for the 1996 Public Use Files from the Survey of Income and Program Participation, available on the SIPP Internet site at www.sipp.census.gov/sipp/sourceac/s&a96_040501.pdf. The source and accuracy statement provides information on direct calculation of standard errors for SIPP estimates. It also provides detailed instructions for calculating standard errors using generalized variance parameters and simplified tables.