DNC Announces New National Record Linkage System

Algorithm developed by DNC expert in the field of record linkage will increase organizing efficiency by 9 percent and provide campaigns with more comprehensive view of the overall electorate

Today the Democratic National Committee is launching a new, proprietary national record linkage system that will be integrated into the DNC’s voter file — deduplicating millions of voter records, providing campaigns with a more comprehensive view of the electorate and increasing the effectiveness of Democratic organizers across the country.

“This record linkage system is an example of how the DNC is developing and implementing cutting-edge data science solutions to increase the effectiveness of Democratic organizers, improve our view of every possible voter and make sure Trump is a one-term president,” said Democratic National Committee Chief Technology Officer Nellwynn Thomas. “The DNC made early, historic investments in our data and technology infrastructure and is working with top data scientists to develop these kinds of innovative products that will help Democrats at every level of the ballot secure victory in November.”

“At the DNC, we’re taking the most advanced data science ideas and applying them to help Democrats win,” said DNC Data Science Director Jody Heck Wortman, who developed this record linkage model. “Having a high-quality record linkage system in place will help Democrats talk to more voters, more efficiently — and we are proud to support the work of our organizers with the best data possible.”

The national record linkage system will:

  • Provide a 9 percent increase in efficiency across the entire voter file by enhancing and deduplicating the information in 30 million records;

  • Help campaign staffers waste less time by trying to contact a voter who has moved (and especially increase the effectiveness of voter registration efforts by matching voters who have moved and not yet registered in their new home state); and

  • Create more accurate projections of the entire electorate for a variety of campaign purposes by integrating additional information about a voter from their matching records.

The National Record Linkage algorithm was derived from work pioneered by DNC Data Science Director Jody Heck Wortman during her PhD at Duke University. Jody is a leading expert on Bayesian matching algorithms, and her expertise in the science of record linkage is providing a unique resource for the DNC and Democratic campaigns.

Starting today, data matched by the National Record Linkage algorithm will be uploaded onto the voter file and can be used by the Biden campaign, sister committees, Democratic state parties and others who have access to the DNC’s voter file.

The DNC’s Data and Technology team maintains a state-of-the art, cloud-based data warehouse and national voter file, which includes every registered voter in America. Under Chair Perez, the DNC has made massive investments to  improve the quality and quantity of data available on the file so that our view of every possible voter is accurate as possible — integrating key data sets, proprietary modeling and augmenting the file in a variety of ways, like purchasing tens of millions of high-quality cell phone numbers. The DNC also provides key data tools to Democratic campaigns to make utilizing the data as convenient as possible, like the DNC’s recently announced Blueprint tool set.