Detailed demographics of correctional populations now available for nation
Database prepared by PPI for redistricting professionals provides detailed demographics for the correctional population, including race, age and gender.
September 13, 2011
September 13, 2011 — A new database prepared by the Prison Policy Initiative for redistricting professionals provides detailed demographics for the correctional population, including race, age and gender. The database expands on the Census Bureau’s Advance Group Quarters Summary File released in April, which identified the population in each Census block that is incarcerated or in some other kind of group quarters.
Redistricting professionals consider it preferable to use the Census Bureau’s counts of correctional facilities when making adjustments to the redistricting data instead of relying on the correctional system or their own knowledge of where the facilities are located. Because the Census often mixes prison and non-prison populations in the same block, and occasionally does not enumerate prisons at their actual locations, the optimal way to identify correctional populations in the Census data is with actual Census data. Our database makes this easy.
Since the Group Quarters Summary File release in April, the Prison Policy Initiative has been working to make the data easier to use. Shortly after it’s release, we translated it into several more accessible formats, including an ESRI shapefile, a Google map and a table searchable by county or state. Some redistricting professionals have been frustrated that the data doesn’t include the names or types of facilities, or detailed demographic counts. Our new dataset fills these needs.
Since April, we’ve used 2010 vintage correctional system data to label the facilities in more than a third of the census blocks that contain correctional populations. (And more are being labeled each week.) This is particularly useful to redistricting professionals because they may wish to treat different types of correctional facilities differently for redistricting purposes, for example, reallocating state prison populations by not local jail populations.
Now, with the publication of Summary File 1, we’ve been able to pull in more detailed age, gender, race and ethnicity data for these populations for most of the blocks. This database is important because the Census Bureau often mixes prison and non-prison populations within the same block, so knowing the demographics of the block does not necessarily tell you the demographics of the prison population. We drew on different Census Bureau data tables depending on the specific circumstances in each block. For example, if the prison is the entirety of a block, the data was easy to access, and we used a table that reports race/ethnicity for the institutionalized group quarters population where the correctional facilities were the only institutionalized group quarter in the block. For other circumstances, we developed other processes to extract the data, and each data element in our database has a full footnote with the source.
Our data can be accessed via:
- Our interactive facility locator, an online tool powered by Google Maps that shows the correctional facilities as counted by the Census in each county in tabular or map form.
- An ESRI shapefile for the entire Advance Group Quarters Summary File.
The detailed demographic data announced today can be accessed by:
- Clicking “detail” for any block in either the table or map views in our interactive facility locator.
- Clicking the link in the group quarters shapefile for any block that contains a non-zero correctional population.
- Sending a URL in the following form: http://www.prisonersofthecensus.org/data/2010blocks/$geoid/ where $geoid is the 15 digit state-county-tract-code of a block with a correctional facility, ie http://www.prisonersofthecensus.org/data/2010blocks/361219702001024
The below table describes the extent of the data we make available today for the 5,393 Census blocks that contain a correctional facility and what data we expect to add to it shortly:
|Blocks labeled with facility type and name||35%||75% of all blocks, and 95% of facilities larger than 25 people.|
|Age and gender||100%||100%|
|9 categories of race and ethnicity ||98%||99.9%|
|Count question resolution program data and other notes on facilities counted in incorrect location.||–||We’re annotating apparent geo-location errors as we discover them, and we’ll import the Census Bureau’s Count Question Resolution program results as they become available.|
This data will be useful to anyone seeking to identify prison populations in redistricting data in order to avoid prison-based gerrymandering. Prison-based gerrymandering is the result of two related, but distinct, problems that require two separate solutions. First, the Census Bureau counts incarcerated people as residents of the wrong location (at the prison). Second, the Census Bureau fails to count incarcerated people as residents of the correct location (at their homes). Some states like New York and Maryland are addressing both problems by counting incarcerated persons at home for redistricting purposes. Our database helps redistricting professionals addresses the first problem: finding the prisons in the Census data.
The Prison Policy Initiative has, since 2001, been working to identify how the Census Bureau’s prison counts distort the democratic process in state and local governments. For the last decade, the organization has been working to convince the Census Bureau to change how it counts people in prison, and working with state and local governments on interim solutions.
Some technically inclined readers may note a discrepancy between the fact that the Census Bureau publishes 126 combinations of race and ethnicity in the PL94-171 Redistricting Data File, and the fact that our database has 9 combinations. The Census Bureau does not provide the data necessary to complete our calculations for prison populations for all 126 race/ethnicity combinations, but they do for the 9 combinations we make available. As a practical matter, redistricting professionals are unlikely to find the smaller dataset problematic for two reasons: First, they typically aggregate the 126 combinations prior to use, and second, the correctional facilities often do not report people of more than one race as such.