Small Area Health Insurance Estimates

SAHIE Methodology

SAHIE continually evaluates its methods and investigates other potential data sources for each release. For these reasons, we revise and improve our methodology with each successive release, as time and resources allow. For a description of the methodology used to create each year of data, refer to the information below. Note that caution should be used when making comparisons across geographies or from year to year. See this FAQ for more guidance.

Data Inputs

"Data Inputs" describes the data that are used in the models to produce SAHIE estimates. It discusses the sources of each dataset, key concepts, and any adjustments made to the data.

Demographic and Income Model Methodology (2008 - 2010)

The 2008-2010 SAHIE estimates are adjusted so that for key estimates, before rounding, the county numbers sum to their respective state totals and similarly the states sum to the national ACS poverty universe for the numbers insured and uninsured.

• Ages 0-64, 18-64, 40-64;
• Ages 50-64 for states and, starting in 2010 for counties;
• Female, male, and both sexes;
• All incomes and income-to-poverty ratios (IPR) less than or equal to 138, 200, 250, and 400 percent of the poverty threshold;
• Ages 0-18, both sexes, and same 5 income categories as above; and
• For states only:  White not Hispanic, Black not Hispanic, and Hispanic (any race).

Measures of uncertainty are also provided.

Demographic and Income Model Methodology (2005 - 2007)

Our demographic and income model produces county estimates of the number of people with and without health insurance coverage by age, sex, and income, as well as state estimates by age, sex, race, ethnicity, and income. For 2005-2007, SAHIE estimates utilize the Current Population Survey's Annual Social and Economic Supplement (CPS ASEC) as an input.  Also each year's SAHIE estimates are benchmarked to the national CPS ASEC corresponding to the income year of the estimates.  For example, the 2007 SAHIE estimates are adjusted so that for key estimates, before rounding, the county numbers sum to their respective state totals and similarly the states sum to the national 2008 CPS ASEC (which contains questions about income during calendar year 2007) poverty universe for the numbers insured and uninsured.

• Ages 0-64, 18-64, 40-64;
• Female, male, and both sexes;
• All incomes and income-to-poverty ratios (IPR) less than or equal to 200 and 250 percent of the poverty threshold;
• Ages 0-18, all incomes, and at or below 200 percent of poverty (2006 and 2007 only);
• For states only:
  • Ages 50-64; and
  • White not Hispanic, Black not Hispanic, and Hispanic (any race).

Measures of uncertainty are also provided.

Demographic and Income Model Methodology (2001)

Our demographic and income model creates state estimates of the number of people with and without health insurance coverage by age, sex, race, ethnicity, and income. The estimates are adjusted so that, before rounding, state numbers sum to the national 2002 CPS ASEC (which contains questions about income during calendar year 2001) poverty universe. We have developed 2001 experimental state estimates of women with and without health insurance coverage for:

• Ages 18-64, 40-64, and 50-64;
• All Races, White not Hispanic, Black not Hispanic, and Hispanic; and
• Women at or below 200 percent and 250 percent of poverty.

Measures of uncertainty are also provided.

Age Model Methodology (2000)

Our age model creates county and state estimates of people with and without health insurance coverage by age. We have a county-level model and state-level estimates, which are aggregated from the county estimates. The estimates are adjusted so that, before rounding, county numbers sum to their states and similarly the states sum to the national 2001 CPS ASEC (which contains questions about income during calendar year 2000) civilian, non-institutional estimates. We have developed experimental county and state estimates of people with and without health insurance coverage by age in 2000 for:

• Total population; and
• Children under age 18.

Measures of uncertainty are also provided.

SOURCE: Estimates obtained from models are applied to CPS data. Model inputs also include Census 2000, intercensal population estimates, and administrative records data. For information on the data sources, measures of uncertainty, confidentiality protection, and sources of error, see SAHIE source and accuracy.

Source and Accuracy

The data provided are indirect estimates produced by statistical model-based methods using sample survey, decennial census, and administrative data sources. To maintain confidentiality, the Census Bureau uses procedures to assure that the estimates and related information that are released cannot be used to disclose individual data or violate other confidentiality restrictions applicable to the source data.

Methodological Review

The Small Area Health Insurance Estimates (SAHIE) program's methodology and experimental estimates have undergone internal U.S. Census Bureau review as well as external review.