News from the Interns: DA Research Project, Part 3

By Samantha O’Brien and Katherine Stanton
August 25, 2017

As is often the case when one is conducting research, we have in a somewhat roundabout way returned to our original method of calculating blank votes. We surmised that, while participation in other high-profile local elections, such as mayoral races, might be a useful metric to ascertain who is already engaged in local elections, more useful would be to identify who had actually voted in an election at which the DA race was on the ballot, and simply left the DA vote blank. These people comprise a valuable target population, and would perhaps vote in the DA race after a brief education. Thus, we used our Excel Master Spreadsheet to calculate blank DA votes for each of the neighborhoods and consolidated precincts in the city of San Diego.

We purchased a map of the consolidated precincts in the city of San Diego from the San Diego County registrar of voters. It cost slightly over $30 and, fortunately, the Institute covered these fees. Nonetheless, this brought up some concerns for us as to how this sort of project requires the expenditure of both time and money, and might be cost-prohibitive for the communities that we are trying to serve.

While we initially envisioned overlaying our maps onto one another without GIS (geographic information system) training, we finally came to terms with the fact that it was difficult to do this type of analysis without GIS. GIS is designed specifically for this type of analysis, and is by far the most powerful as well as most efficient tool to overlay multiple data sets. Sam and Tolu had, by this point, completed their summer internships, leaving Katherine the sole person in the office to fiddle around with our data and maps.

Using GIS, Nancy McArdle, the consultant who assisted us earlier in our project, made four preliminary maps of the City of San Diego: the first mapping the distribution of blank votes within the city, the second mapping the distribution of registered voters in the city, the third mapping the difference between registered voters and DA votes, and the final showing the distribution of black (non-Hispanic) voting-age population within the city.

Because we had determined that there was a low number of blank votes in San Diego, we decided to concentrate on targeting areas with a large difference between registered voters and DA votes and a large black voting-age population.

Nancy then kindly put together three maps for us: two versions of a map that show block groups with large black voting age populations and precincts with a large difference between registered voters and DA votes, and one map that demonstrates block groups with large black voting age populations and precincts with low numbers of registered voters. The first type of map would allow a community member to focus on increasing voter participation within areas that already have high rates of registration among black voters, and the second type of map would allow a community member to increase registration among these populations, which have historically been barred from registration through insidious forms of legal and social disenfranchisement. However, this map reflects current rates of registration relative to the voting-age population. For a more accurate map, we would need to compare registration rates with data on the numbers of eligible voters in an area. Eligible voters, as opposed to population counts, takes into account those who are disenfranchised as a result of incarceration and other barriers. This data is available, and something that this project will work to incorporate in the future.

Simultaneous to this work, we have been building our toolkit. In addition to outlining our steps in terms of data accumulation, organization, and mapmaking, we have also been developing an educational toolkit for anyone who would like to learn more about the role of District Attorney, its complicity in systemic racism and mass incarceration of black and brown Americans, as well as how to research candidates prior to working on promoting voter engagement. We have included tips and sample questions as to how one might approach attending a District Attorney’s public forum. Finally, we have provided two versions of a step-by-step methodology—one that ought to be generally applicable, and one applied to our specific case study of San Diego.

In the future, we need more research into two issues that we encountered: first, GIS technology access, and second, how to calculate the difference between eligible voting populations and current rates of registered voters.

Regarding the first issue: GIS is a very powerful tool that is necessary for this type of analysis. However, the learning curve is steep, and, while the software is available open-source on the Internet, not everybody has the time or resources to learn GIS. Therefore, an internet-based GIS application that would allow an organizer to easily plug in numbers and churn out a map would be an immensely useful tool. In this case, the organizer would be responsible for inputting blank voter and registration election data for the city or county of choice. The website application would input census data and create maps to fit the organizer’s need. One map could show areas with large black voting age populations and high numbers of blank votes and another map could show an area with a large black voting population as well as a large difference between registered voters and DA votes. In this case, the powerful GIS technology would be harnessed within a nicely-formatted, easy to use interface that would allow an organizer to see and manipulate only the relevant information.

On the second point, we need to conduct further research on data around eligible voting populations. This is important because in many cases, a county will not have high numbers of registered black voters. However, it is impossible to determine why there are low numbers of registered black voters – it could be that the area has a low population in general. Further, it could also be due to factors such as incarceration. Thus, if an organizer wishes to engage in registration efforts in a county with a low number of registered voters, they need to be able to determine the exact numbers of eligible voters within a county (which is different than the 18+ population within a county). Determining this data involves sourcing information from U.S. Census data on voting age populations and state-level disenfranchisement policies. Thus, an organizer wishing to compare current registration numbers vs. eligible registration numbers will need to compile the information of eligible voters taking into account these factors.

Finally, to overlay this information on a map using GIS is more complex. To calculate this information involves subtracting and dividing election data, which uses precinct boundaries, from census data, which uses block group boundaries. This GIS calculation is a little more complex and requires someone who is more highly-skilled in GIS.