Skip navigation to content
eriu: Economic Research Initiative on the Uninsured Initiating and disemminating research to spark new policy discussion on health coverage issues.
Research Findings  
   
Fast Facts Home
 
Facts to Consider
 
 
Counting the Uninsured by Reference Period
All Year
Ever in Year
Point in Time
By Data Source
CPS
MEPS
SIPP
By Population
Total Non-Elderly
Adults
Children
Workers
Immigrants
By Additional Characteristics

Income
Education Level
Family Composition

Comparison Tables
CPS-SIPP
CPS-MEPS
Medical Utilization and
Expenditures Tables
 
Notes on Data Sources
and Variables
CPS
MEPS
SIPP
Data Dictionary
 

Notes on SIPP Data Source and Variables

Source data: 2001 Panel, Survey of Income and Program Participation. Data for first twelve survey months (waves 1, 2, and 3.) All data are weighted by the final person weight for the first year of the survey, WPFINWGT_12, except for data relating to immigrants. Immigrant status is from the topical module administered in wave 2, and the weight is that on the topical module file.

Gender is determined using the variable ESEX

Race is determined using the variable ERACE

Hispanic origin is determined using the variable EORIGIN. All individuals coded as 19 – 29 (inclusive) were recoded as Hispanic.

Immigrant status is determined using the variable TCITIZNT (from the wave 2 topical module). Individuals coded as "naturalized citizen" or "not a naturalized citizen" (responses 2 and 3) are viewed as immigrants. Some individuals interviewed about immigrant status were not interviewed for each of the three interviews that provide twelve months of data and thus health insurance status, which requires 12 months of data, is unavailable; these individuals have a weighted total of 3,295,506.

Age is determined using the variables TAGE_01, TAGE_02, TAGE_03,…, TAGE_12, using the highest age over the year.

Education is determined using the variables EEDUCATE_01, EEDUCATE_02, EECUCATE_03 … EEDUCATE_12, using the highest level of education attained over the year.

Workers, the universe for some tables, includes those who worked at anytime (EPDJBTHN) over the year (i.e., in any of the three waves.) Those who report only self-employment are excluded, as are those who report both employment and self-employment and for whom self-employment makes up a larger share of their labor force participation over the year.

Individual work status is determined by the usual hours worked per week recode (RMHRSWK) for month 4 of wave 3, with "full time" coded for those with RMHRSWK values of 1 (all weeks 35+ hours) and 5 (at least 1, but not all weeks 35+ hours; all other weeks 0 hours); and "part time" for values 2 (all weeks 1-34 hours), 3 (some weeks 35+ and some weeks less than 35, all weeks equal to or greater than 1), and 4 (some weeks 35+, some 1-34 hours, some 0 hours.)

Family work status is determined from wave 3 responses. Families are identified by sample unit identifier (SSUID), household address ID (SHHADID) which differentiates households in the sample unit, and family ID number (RFID.) Part-time v. full-time is determined in the same manner as described for individual work status, above.

Wage/salary for primary earners. Wages and salaries are calculated based on wave 3 interview responses. For those who report an hourly wage (TPYRATE1 or TPYRATE2), that wage is used (with the higher or TPYRATE1 and TPYRATE2 used where those who report two jobs.) For those who do not report an hourly wage, monthly earnings (TPMSUM1 and TPMSUM2) are divided by four, and the result is divided by usual hours worked (EJBHRS1 and EJBHRS2.) For those who report two jobs, the job with the highest calculated wage is used. For those who report self-employment, an implicit wage is calculated in a similar manner as for the employed, and for those who report two businesses, the wage variable reflects the higher of the two implicit wages. Business earnings are the higher of TBMSUM1 or TPRFTB1 (or TBMSUM2 or TBMSUM2 for those who report two businesses.)

Family composition. Martial status is determined by the variable EMS. Families are identified by sample unit identifier (SSUID), household address ID (SHHADID) which differentiates households in the sample unit, and family ID number (RFID.)

Income as a percentage of the poverty level is determined using the variables RFPOV_01, RFPOV_02, RFPOV_03,…, RFPOV_12 and TFTOTINC_01, TFTOTINC_02, TFTOTINC_03,…, TFTOTINC_12. The sum of TFTOTINC_01 – TFTOTINC_12 was divided by the sum of RFPOV_01 – RFPOV_12.

Uninsured is determined using ECRMTH_01 – ECRMTH_12, ECDMTH_01 – ECDMTH_12, and EHIMTH_01 – EHIMTH_12.

  • Uninsured all year – individuals who did not have Medicare (ECRMTH), Medicaid (ECDMTH) or private insurance (EHIMTH) for all 12 months of the year.
  • Uninsured point in time: individuals who did not have Medicare (ECRMTH), Medicaid (ECDMTH) or private insurance (EHIMTH) during just the 12th month of the year.
  • Uninsured ever during the year: individuals who did not have Medicare (ECRMTH), Medicaid (ECDMTH) or private insurance (EHIMTH) during any month of the year.