USDA Economic Research Service Data Sets
" "  
Search ERS

 
Data Sets

Print this page Print | E-mail this link E-mail | Bookmark & Share Bookmark/share | Translate this page Translate | Text only Text only | resize text smallresize text mediumresize text large

2010 County-Level Poverty Rates: Model Based Estimates

Decennial and Model-based Intercensal Poverty Estimates

The 1979, 1989 and 1999 poverty estimates are from the Summary File 3 (SF3) release of U.S. Census data from the last three decades. Each poverty estimate is for the year preceding the Census date because the survey questions ask about income from the previous year. The SF3 files are compiled from the "long form" Census questionnaire, which was collected from about 1 in every six households (or about 19 million housing units in 2000). These data are often referred to as the SF3 Census data, but it should be noted that they are sample data with sampling error, and not a Census of the population. Poverty status is determined for all people except institutionalized people (e.g. prisoners or residents of nursing homes), people in military group quarters, people in college dormitories, and unrelated individuals under 15 years old. These individuals are excluded from both the numerator and denominator when calculating poverty rates. They are considered neither poor nor nonpoor. For more details on the source and accuracy of the SF3 data, see Chapter 8 of the Census Technical Documentation.

The 2010 poverty estimates, the most recent available, are based on different data sources and are generally considered less accurate. The long form Census data provide sufficiently large sample sizes to produce reliable poverty estimates for small geographic units, such as the county. Census data though are only available every ten years and therefore are unable to provide poverty estimates for intercensal years. The U.S. Census Bureau created the Small Area Income and Poverty Estimates program to provide more current estimates of income and poverty. Census Bureau combines several different sources of annual data to derive county-level poverty estimates from statistical regression models. These models relate income and poverty to indicators based on summary data from federal income tax returns, data on participation in the Food Stamp program, data about Supplemental Security Income program recipiency rate, economic data from the Bureau of Economic Analysis (BEA) and the most recent census. These estimates are then combined with direct estimates based on the March Supplement to the American Community Survey (ACS) to provide figures which are more precise than either set alone. The final combined estimates are referred to as model-based. For an overview of the estimation procedure used for the small area poverty estimates, see the Description of Estimation Procedure, and for cautions about comparing these estimates with the SF3-based estimates, see General Cautions about Comparisons of Estimates.

 

For more information, contact: Tracey Farrigan

Web administration: webadmin@ers.usda.gov

Updated date: December 2, 2011