BU Announces New Data Source: National Voter File

Boston University (BU) has acquired a three-year subscription to a national voter file for the United States. This data, which is from the commercial data vendor L2, includes records on over 200 million voters plus 2.3 billion voter history records. 

Acquired through a grant to the Center for Antiracist Research (CAR), this data has been made available to the entire BU academic community, including faculty, graduate students, and undergraduates. 

We spoke to Maxwell Palmer, Associate Professor of Political Science and CDS Civic Tech Fellow, to learn more about this fantastic new resource for research, teaching, and policy analysis.


Tell us about this database. Who is it for?

Voter files are administrative records of every person registered to vote in the United States. While each voters’ ballot is private (who we vote for is not public), the act of voting is public and recorded on the voter file. This database is a national voter file; the vendor collects the voter files from each state, cleans them, and assembles them into a uniform format for analysis. The vendor also merges in additional geographic, electoral, and demographic data.

How can this database be accessed?

The data is provided as large .zip files for each state. They are all available on the BU CAS server, at "smb://casfs.bu.edu/CAR L2 Voter File”. These files are updated by the vendor a few times each year and available to anyone at BU.

Why is this data set so important and what are the broader implications of having access to this data?

Voter files are important administrative records. They provide a list of every registered voter and their history of voting in elections. These files can be used for a wide range of purposes, including studying voting rights, voter mobilization and turnout, and campaign strategies.

What does having access to this database mean for the data science community in particular?

Many data scientists are interested in studying questions about elections, voting rights, representation, and democracy. This database provides current, individual-level data that can be leveraged to help answer these questions. For example,  I used voter files as part of a study on the effects of car ownership on voter turnout. They have also been used to study partisan sorting, voter mobilization, participation in local government, and many other important topics.

How can it be leveraged to solve or analyze research questions or policy issues?

Voter files are a powerful way to study elections, voting rights, and other questions because they include every voter in the United States at the time of the election, and including a variety of demographic and geographic data. In particular, every voter’s address is geocoded and matched to census geography, allowing for a wide variety of analyses.

Who should people reach out to with any questions?

People can contact me (mbpalmer@bu.edu) with questions. There is an extensive codebook available with the data. Over the next few months I will be adding additional documentation and some files, including county-level extracts, to make the data easier to use.


Interested in accessing this new database? All of the data is available on the CAS server, at "smb://casfs.bu.edu/CAR L2 Voter File" (you will need to connect to the network using the BU VPN if you are not on the BU network).

Additionally, if you do use the data, please fill out the Google form here, detailing how you are using the national voter file data. This will help justify future funding to renew the data subscription