Wednesday, December 30, 2009

American Community Survey Contrast Charts

We've just uploaded over a thousand documents with tables and charts from the 2008 American Community Survey (ACS) released this fall by the Census Bureau. The charts reveal socio-economic disparities experienced by African-Americans, Latinos, and Native Americans vis-a-vis non-Hispanic whites.

Click through these directories to download PDF files by nation, state, and selected counties and places (updated February 2012).

The national and state documents are from the 2008 one-year survey and the county documents are based on the 2006-2008 three-year survey.

We've included all counties with estimated populations of at least 20,000 persons, of whom 10,000 or more are African-American or Latino. ACS estimates for counties with fewer than 20,000 persons will not be available until the fall of 2010.

We use charts similar to these for testimony in Section 2 voting rights cases (redistricting). But the charts may also prove useful for local and state advocacy.

The documents are up to 60 pages in length -- but may be smaller depending on sample size and data suppression issues. We've included most of the key variables broken out by race or ethnicity in the American Community Survey. Persons familiar with the the SF 3 long form sample data from the 2000 census will notice differences in the data presented. For example, median home value, median rent, access to a vehicle, and telephone service are not available by race in the 2008 ACS.

The FairData website has several thousand contrast chart documents for most places in the country with significant minority populations, based on the 2000 census.

Thanks to M-SLICE  in St. Louis,  MO for suggesting these contrast chart updates.

Thursday, December 24, 2009

Potential New Majority-Black Congressional District in South Carolina

According to the latest population estimates from the Census Bureau, South Carolina is set to gain a new congressional seat in 2011. This means that a new majority African-American congressional district is in the cards.

The map below shows one possible configuration -- without getting into messy little details such as where incumbents live. The map is based on 2006 census tract population estimates (the latest we have available).

District 6 (yellow) has a 56% overall African-American population (about the same as the current District 6) and District 7 (light orange) is 53% African-American. We're cutting it close on the minority voting age percentage in District 7, but the district can be fine-tuned and probably strengthened when precinct-level data is released following the 2010 census

District 6 encompasses predominantly African American neighborhoods in Charleston and surrounding rural areas. The new upstate District 7 would combine parts of Columbia with the Greenville-Spartanburg area.

View Larger Map

There is no guarantee that South Carolina will get a new congressional seat when the actual count for the 2010 census is released. As explained in this post by the Swing State Project, it will all boil down to a few thousand persons. Community census advocates can use our Hard-to-Count Mapper or our Google Map overlays to help focus outreach efforts.

Monday, November 2, 2009

¡Hágase Contar!

Probably the single-most important factor which may contribute to an undercount in the 2010 Census is citizenship status. The Census Bureau estimates that there are 21.6 million non-citizen residents -- an increase of 3 million non-citizens over the 18.6 million reported in the 2000 Census.

The interactive table below shows the percentage of non-citizens for all places in the United States with a current population of 20,000 or more -- based on estimates from the 2006 to 2008 American Community Survey (ACS) released last week by the Census Bureau.

At 41.6%, Hialeah, Florida has the highest non-citizen percentage, with about 86,000 non-citizen residents. New York City has the largest non-citizen population -- about 1.48 million persons (17.8%).

You can search, sort, and filter the data in the table.

View expanded table

Source: U.S. Census Bureau, 2006-2008 American Community Survey, Table C05001.
Link to ACS tables
Data are based on a sample and are subject to sampling variability. Data are not available for some places because the number of sample cases is too small.

The interactive table below shows the same dataset for all counties with a population of 20,000 or more.

View expanded table

Hendry County, Florida and Miami-Dade County, Florida lead the pack with populations that are over 25 percent non-citizen.

The Census Bureau will not release ACS citizenship estimates for small jurisdictions (under 20,000 pop.) and census tracts until late 2010.

If you need street-level detail, the interactive map described in our July 23, 2009 blog post includes a block-group level thematic map showing the percentage of non-citizens, according to the 2000 census. At a scale of about 5 miles, the block group count of non-citizens (as of April 1, 2000) is displayed in purple boxes.(A related map in this application shows the percentage of the voting age population in 2000 that did not speak English "very well".)

You can also overlay tract-level citizenship info from Stanford's gCensus onto our Hard-to-Count census tract Google Maps using the technique described in this Google Earth Forum post (with map links).

Thursday, July 23, 2009

Get out the Count -- Census 2010

The decennial census count has major implications for reapportionment, redistricting, and funding. Groups across America are conducting outreach campaigns to ensure that everyone is counted on April 1, 2010.

The Census Bureau has developed a list of "Hard-to-Count" (HTC) census tracts based on underlying demographic and socio-economic indicators. These are the areas where community-based "get out the count" efforts should be focused. Nearly one-fifth of the population lives in hard-to-count census tracts.

For details on the HTC methodology employed by the Census Bureau, see this PDF document or this one-page summary. Source data is available from the Census Bureau in an Excel spreadsheet.

Some state agencies have web pages with maps depicting HTC tracts -- see, for example, Maryland, Pennsylvania, Michigan, and Indiana. You can also download PDF maps by county from the Census Bureau showing census tract boundaries via this directory.

To help in a small way, we've developed an interactive map that highlights the Hard-to-Count tracts.

To access the interactive map, click the arrow icon or text link below.

Hard-to-Count 2010 Mapper

(added recently released block group HTC scores -- see Map 2 --  11/25/09)
Source: 2010 Enhanced Planning Database

(3.5 mb --nationwide -- sorted by state and county -- 7/27/09)

(view and print from browser -- 10/17/09)

(organized by state and county for easy export -- 11/30/09)


Census tracts are displayed on the interactive map beginning at a scale of about 200 miles. Zoom the map by clicking the buttons at the top and then on the map. Click the INFO tool and then on the area of interest for demographic data from the 2000 census. HTC scores for counties and tracts are displayed near the top of the pop-up INFO window. At a scale of about five miles, the census tracts are labeled with identifying numbers.

Using the Hard-to-Count 2010 Mapper, you can zoom in for street-level detail on tracts where an accurate census count will be problematic (shaded in red). The Census Bureau defines HTC census tracts as those with a profile score of 60 or higher. The national mean score is about 33. Tracts above the mean (with a score between 33 and 60) are shaded green. Tracts shaded yellow have HTC scores below the mean average and are therefore less susceptible to a census undercount.

The Hard-to-Count 2010 Mapper initially displays a nationwide county-level map -- thematically shaded to reflect the mean average HTC score across all tracts in a county. This county average is weighted by the underlying tract populations. Census tracts generally have a population of about 4,000 persons, but the tract-level population varies significantly within and between counties.

Counties in the South, Appalachia, Rio Grande Valley, and Mountain West have the highest HTC mean average scores. Central cities everywhere have numerous HTC tracts.

Additional maps available via the drop-down menu to the left of the map image window provide block group-level detail to further refine census outreach efforts (e.g. citizenship status and Spanish speakers). For details, refer to the notes on the SocioEcon Mapper on the FairData website. All maps are based on data from the 2000 census, except for the undercount map which relies on statistically adjusted data from ACE I released by the Census Bureau in 2002.

The tri-color thematic shading for the HTC interactive map is consistent with HTC maps prepared by Southern Echo. Southern Echo has also developed an excellent Census and Redistricting 2010 Tool Kit for several states.

The Southern Coalition for Social Justice (SCSJ) has prepared a series of reports on the undercount in the 2000 census, with a special focus on the South. These reports are especially useful for identifying cities and counties where community outreach is needed to avoid an undercount in the 2010 census.

The materials available from Southern Echo and the Southern Coalition include county-level analyses of the potential negative impact on federal grants due to a census undercount.

Also, the Nonprofit Voter Engagement Network has developed a website -- Nonprofits Count! -- with resource materials and maps to help communities mobilize for the 2010 census. The Funders' Census Initiaitve is another useful Census 2010 resource website, with links to key organizations working on census outreach.

Persons interested in joining or organizing a complete count committee for their community should call the regional census center or visit (See, for example, the Denver Complete Count Committee web page.)