Toward an Accurate Voter File: Reversing the Redlining of Voters of Color

Hi everyone.

Welcome to 2024!

Over the next several months, as we get closer to the November election, the political landscape will continue to evolve and change, but one thing will remain the same: the importance of having accurate data to inform organizing and civic engagement strategies. Data -- how we collect it, what it means, and how we use it -- is fundamental to modern political life. Data can allow us to “know” a lot about whose doors we are knocking on. Data can help organizers and others develop “smart” targeting for outreach and communications programs.  

The promise of data is that it can mean that we are not flying blind. But, unfortunately, the reality of our current data collection and analysis systems is not so rosy. We think that we see clearly, but we are, in fact, operating with big blind spots that distort the landscape and undermine our work to support organizations in building the multiracial power they need to win change.

For years, organizers have known that the models, data systems, and voter files developed and promoted by the traditional progressive political industry are deeply flawed. Their practice has taught them that these tools are not helpful for determining outreach targets and strategies because a high percentage of Black, Latino, AAPI, and Native people are missing or misidentified in the voter file. And the algorithms that determine “propensity” to vote undercount, undervalue, and underestimate people of color.

Given this reality, organizers in communities of color have had a choice: continue to use the existing, flawed systems that leave so much power on the table or develop new methods for reaching and engaging these missing voters. Groups like ISAIAH, Ohio Organizing Collaborative, LUCHA, Florida Rising, Voces de la Frontera, and others chose the latter. Each has developed innovative methods for identifying and reaching “missing” voters, adding them to their outreach lists, and bringing them into political engagement.

In order to understand and document the problem of “missing voters” and the solutions being developed and deployed by organizers, DPI data analysts Miriam McKinney Gray and Tianyi Hu worked with several DPI partner organizations to document and share their learnings and reflections about the problems with the existing data systems, strategies for filling in the gaps, and ideas for wholesale reform. You can read the report here: Surfacing Missing Voters: Addressing Data Systems, Tools and Engagement Models that Invisibilize Black and Brown Communities.

2024 is a big year for all of us and our communities. As we look toward November, there is much more work to do to improve the voter files, models, outreach lists, and data systems to make them accurate and reliable tools for organizers to use to build community engagement and increase participation. But our organizing leaders have shown us that they are not waiting. They are innovating now and promoting ideas for lasting change for the future.

Onward!

Julie

Read the Full Report

Reflections from The Strategists

As an organizer, what is the impact of your members missing from the file?

Historically, data has been used against Black, Brown, Indigenous, and other communities of color, which has led to leaving us behind. That is why we decided to reclaim data and use it to inform our strategy to win in our communities.  For the last decade, we have been focusing our outreach and organizing in places where people are (home, grocery stores, parks, etc.) to build an intentional relationship, discuss important issues, and provide information on how they can be part of fixing those issues. For most of this time, all we had to keep track of this work was pen and paper which resulted in us missing opportunities.

In 2019 our organization made a decision to invest in our data infrastructure to try to understand and keep track of our organizing and outreach program. One of the first things we did was create an inventory of our list. What we found was hundreds of wrong email addresses, duplicates, wrong information, and an astronomical amount of missing data. For the last four years, we have worked across departments to develop more robust and uniform systems for data collection to improve the accuracy and quality of our outreach and canvassing lists.

Although we have been doing this for a few years now, we are still missing data, which makes it harder for us to do our work effectively. For example, we want to send Spanish-only content but we are missing a lot of data that would allow us to target by language, so we will be leaving folks behind.

Aside from not being able to get a full understanding of our base, having inaccurate and incomplete lists can create a false representation of who our supporters are and what their priorities are. The reliance on the voter file and existing outreach tools also makes our work harder when we want to target unregistered voters as they are often left out by these tools. These are community members, often the ones very close to the issues we are trying to solve, so it is important for us to find them and connect with them.‍

How are you going to use this research to build power in your community?‍

We have already started to implement better tracking guidelines that will help us better match people to voter files. There are still a lot of issues with matching people who have names that do not follow the standard single first and last name, people who use the same phone number or email, but we are being more intentional in capturing other data points that will help with the matching process. We also have our Voteria campaign designed to identify and reach out to members of our community that are not in the voter file because they are not registered to vote. We are developing strategies to find these people, register them to vote, and introduce them to our organization to continue to build that relationship with them. These are only some of the things we are doing at LUCHA to help find and activate missing voters, but also to get those who can't register, to get involved and give them a voice of their own through other means.

‍Why are you researching missing voters?

Black, Latino, AAPI and Native people are disproportionately missing from seemingly inclusive data systems for engagement. I wanted to know why.

According to an analysis by researchers Jackman and Spahn in 2021 as well as 2021 ACS 1-Year Estimates Subject Tables for Citizens 18 years and over, approximately 24.76 million Black and Latino eligible voters are inaccurately listed or missing from voter files nationwide. And given that the Census routinely undercounts Black and Latino communities, I suspect that even this is a low estimate.

Throughout all of my work and research, I have observed that Black and Latino individuals are often ignored by widely-accepted, data-driven voter engagement systems in the United States. And yet, the ability to access wholly representative data (meaning data that reflects a community in its entirety) is necessary to support multiracial democratic practice. Having missing or incorrect data in voter databases also distorts the algorithms assigning vote propensity scores to individual voters.

Our traditional processes in the power-building movement rely on inaccurate and incomplete data, voter file matching systems that fail or are unable to draw accurate conclusions about certain people, and data tools that are not inclusive or without error. If not resolved, the effective erasure of millions of Black and Latino individuals from virtualized voter engagement systems will continue.

What do you hope this community will take away from this research?

This research is a call to action for funders and an opportunity for organizers to learn from each other about what is missing in their data and what techniques they can use to overcome the existing biases. In the short-term, organizing groups should feel equipped to ask deeper questions of their programs (for example: how many people are missing from our voter files, what vote propensity score buckets do most people fall into, and how can relational organizing be leveraged in light of what we now know?), recognize that even the smartest-sounding algorithms will have their shortcomings, and be prepared to investigate their datasets further.

‍

How organizers and funders can begin to use this research today?

1. We recommend that organizations move away from using propensity scoring as a threshold for including or excluding voters for contact. Instead, we suggest that organizations prioritize engagement with their active membership through a peer-to-peer engagement strategy, then branch out from there.
2. We recommend that organizations begin or continue to leverage available community-driven databases and tools to analyze pooled data. One example of such data is the TMC_Activist_Pool schema produced by The Movement Cooperative, currently utilized by several groups.
3. We recommend that organizations marketing people-focused data tools invest more time into quality assurance and increased functionality that can better support users’ needs for impending data analysis.
4. We suggest that funders move away from traditional input-based metrics for analyzing organizational effectiveness toward metrics focused on impact.
5. Finally, we recommend that both funders and organizations take time to invest deeply in analytics, critical thinking, and cultural competency skills within organizations. These skills support the practice of healthy skepticism and dissection of data and data systems, as well as the ability to “think” through data to find solutions.

‍