This report uses data from the Survey of Income and Program Participation (SIPP) to examine issues related to health insurance coverage. It focuses primarily on the extent to which people are covered by health insurance over a 32-month period beginning in early 1991. The source of this information is the 1991 panel of the SIPP, which contains records for each survey person for whom a reasonably complete set of data for a 32-month period was obtained. Efforts were made during the life of the panel to follow persons that moved to ensure that the sample remained representative of the noninstitutional population of the United States.
It should be noted that longitudinal estimates presented here are based on persons who were interviewed in all waves1 of the reference period or for whom imputed wave information exists.2 Insofar as persons who left the panel were differentially covered by health insurance, estimates on months covered across the panel may be biased. Additionally, there may be a time-in-sample bias present in the longitudinal estimates.
All demographic surveys, including Current Population Survey (CPS) and SIPP, suffer from undercoverage of the population. This undercoverage results from missed housing units and missed persons within sample households. Compared to the level of the 1980 decennial census, overall CPS and SIPP undercoverage is about 7 percent. Undercoverage varies with age, sex, and race. For some groups, such as 20 to 24 year old Black males, the undercoverage is as high as about 27 percent. The weighting procedures used by the Census Bureau partially correct for the bias due to undercoverage. However, its final impact on estimates is unknown.
During each SIPP interview, information is collected on health insurance coverage (along with other information on income, labor force, and program participation) for each month in the 4-month reference period. It is, therefore, possible to classify persons by the number of months over the 32-month period that the person was covered by one or more types of health insurance. It is also possible to measure the number of months continuously spent without insurance coverage; that is, spells of noncoverage.
Health insurance in this report refers to the following types of coverage:
1 Sample households within a given panel are divided into four subsamples of nearly equal size. These subsamples are called rotation groups and one rotation group is interviewed each month. In general, one cycle of four interviews covering the entire sample, using the same questionnaire, is called a ware.
2 A ‘‘missing wave imputation’’ procedure was used for persons who missed an interview but had completed interviews before and after the missing wave. See appendix B.