If you’ve been following election news, you’re likely aware of the controversy surrounding Pennsylvania’s strict Voter ID law.  Part of a trend across the country of imposing new restrictions on voting, the new law requires that citizens show valid photo ID whenever they cast a ballot.  The Committee of Seventy has assembled some good resources explaining the context for this law and exactly what it requires of voters.  In July, shortly after the US Department of Justice began investigating the law under the Voting Rights Act and as hearings were getting underway about the validity of the law under the Pennsylvania Constitution, we were approached about analyzing the impact of the law from a geographical perspective.  As civic-minded data geeks we readily agreed.

The state had recently released two sets of data identifying registered voters would potentially be disenfranchised by the new requirements because they seemed to lack a valid form of ID: their names either did not appear in the Pennsylvania Department of Transportation database, or did appear but whose identification would be expired to long too be valid on election day.  Initial analyses indicated that Philadelphia voters are affected disproportionately relative to the rest of the state — while about 9% of registered voters were impacted statewide, in our city the number was closer to one in three, with more than 280,000 of the city’s 870,000 active voters affected.  Tom Boyer has produced a good analysis of how these numbers break down by age.  Our task was to see how they were distributed spatially, and if there was a pattern in relation to race.

We had previously aggregated block-level demographic information for the population age 18 and older from the Census 2010 Redistricting Data to Philadelphia’s ward divisions in order to support Fix Philly Districts.  While we would have liked to conduct the analysis statewide, we only had access to ward divisions (voting precincts) for the City of Philadelphia.  Using the unique voter ID number we joined the two lists of people without valid ID to the list of registered voters in Philadelphia who had cast a ballot in the past four years.  Each voter record included a ward division number, so we were able to determine the number and percentage of voters in a given ward division who appeared to lack valid ID.  By joining this data to ward division geographies we were able to 1) map the data and identify patterns in the spatial distribution of voters without valid ID and 2) determine if there was a relationship between the proportion of voters without valid ID and the racial composition of a ward division.  You can read the full analysis on our blog, but our key findings were

1. Around the city’s universities the proportion of voters without PennDOT-issued ID is especially high, suggesting that the law may have a particular impact on students.
2. There is a clear relationship between the racial and ethnic makeup of a ward division and the proportion of voters without ID: as the proportion of African-American or Latinos increases so does the proportion of voters without ID; conversely, the greater the percentage of whites, the lower the rate of ID problems.  This relationship is particularly strong for expired ID but persists (more weakly) for those without PennDOT ID.

Based on the Philadelphia data, it would appear that the state’s new voter ID law has a disparate racial impact, most negatively affecting racial minorities as well as college students.

Scatterplot of the percent black versus the percent without valid ID at the ward division level

There have been a number of developments since we released the results of the analysis.  The Commonwealth Court judge hearing the case declined to issue an injunction against the law and the plaintiffs are appealing the ruling to Pennsylvania’s Supreme Court.  Oral arguments are scheduled to take place September 13^th in Philadelphia. A Texas voter ID law similar to Pennsylvania’s was recently blocked by a federal court because it placed an unfair burden on the poor and racial minorities.  This doesn’t necessarily represent a precedent that will apply here, since Texas, unlike Pennsylvania, must submit changes to its election laws for pre-clearance under the Voting Rights Act due to the state’s history of racially discriminatory practices.

While the legal status of Pennsylvania’s law is yet to be decided, community organizations continue to mobilize to ensure that all eligible voters have the ID that will enable them to cast a ballot on November 6^th.  Azavea will be sponsoring Hacks for Democracy September 14^th through 16^th, a hackathon devoted to the development of applications related to elections and politics.  If you’re interested in working on a project related to the voter ID data (or any other civic app) please register and attend!

On another front, we’ve heard that PA Senate Majority Leader Dominic Pileggi is drafting legislation that would require counties to submit voting precinct boundary files in digital format to a centralized database that would be maintained and published by the state, making analyses like the one we conducted easier in the future.  We encourage you to support the open data and open standards movement by contacting legislators and government officials from the federal to the municipal level, and contributing your energy to the work of organizations like the Open Data Foundation, the Sunlight Foundation, Code for America, and Knight-Mozilla OpenNews.