Voter data on 217M+ US residents

What is L2?

L2 is a nonpartisan voter data provider that consolidates official state and county voter registration files into a single national dataset covering more than 217 million U.S. voters. L2 standardizes records across jurisdictions and enriches them with demographic, behavioral, and modeled attributes, which makes it a longstanding backbone for political research, election analytics, and academic work on the U.S. electorate.

‍

What academic researchers should know about L2 voter data

On Dewey, L2's National Voter File covers approximately 217 million currently registered voters across all 50 states, with each row anchored by L2's persistent voter identifier (LALVOTERID) so individuals can be tracked across snapshots and elections. Records include vote history, age, gender, location, party affiliation, and hundreds of additional demographic and modeled attributes, sourced directly from state and county election authorities and standardized through deduplication, address standardization, geocoding, and household linkage. A separate Historic National Voter Files product preserves processed snapshots from 2014 through 2024 with 400 to 600 fields per record, allowing longitudinal designs that follow the same voters across many years. The Early Voting Return dataset adds near-daily updates on requested, received, and returned absentee and early ballots for federal, statewide, and local elections going back three years, and the National Consumer File extends the picture to roughly 264 million U.S. adults with income, home ownership, donor activity, and other commercial attributes linked at the individual and household level.

‍

Why academic researchers choose L2 on Dewey

Voter file research lives in a difficult place: state and county registration files come in 50 different formats, change definitions over time, and require expensive engineering to harmonize. L2 has been doing that harmonization work for over three decades, which is why its data has been used in published research on turnout, polarization, mail voting, and partisan sorting at Stanford, Harvard, Yale, MIT, Princeton, Cornell, the University of Michigan, UCLA, and dozens of other universities. The persistent voter identifier across snapshots makes longitudinal study designs tractable in a way that working from raw state files almost never allows. Pairing L2 with SafeGraph foot traffic, ATTOM property records, BrightQuery firm data, or LobbyingData contracts opens up multidisciplinary work on how voting behavior connects to where people live, work, and move.

‍

L2 academic research ideas and use cases

Voter turnout and election policy. L2's vote history fields make it possible to study how policy choices like mail voting rules, ID requirements, registration deadlines, and polling place changes affect actual turnout across different populations. A 2024 Brennan Center analysis used L2 data to show that restrictive mail voting rules in Texas placed disproportionate burdens on minority voters. Recent work from the Yale Cowles Foundation drew on L2 records to analyze how voting policies and selective turnout produce partisan bias in election outcomes.

Voter registration and access. L2's nationwide coverage and longitudinal snapshots support research on who registers, who gets purged, and how registration churn varies across states. A study in Social Science Quarterly used L2 data to clarify what is actually known about eligible versus registered voters in the United States, a useful corrective for work that relies on self-reported turnout from surveys.

Demographic and consumer behavior linkage. The persistent voter identifier connects the National Voter File to L2's National Consumer File, which covers roughly 264 million U.S. adults with attributes like income, home ownership, donor activity, and education. Researchers can study how consumer characteristics relate to voting behavior, donation patterns, or response to political mobilization, and pair the file with SafeGraph or Advan Research foot traffic for richer behavioral context.

Field experiments and causal inference. With records at the individual level, persistent IDs, and detailed vote history, L2 has become a standard backbone for randomized field experiments on voter mobilization, ballot information, and civic engagement. Researchers running mailers, digital outreach, or canvassing audits can measure turnout after treatment by linking assignments back to L2 records, scaling experimental designs across states and election cycles.

‍

Dive deeper with Dewey documentation

Detailed information on onboarding with Dewey, data partner details, and technical documentation on data access.

View documentation

‍