A new election-analysis project will offer estimates of how demographic groups vote in U.S. House and Senate races across every state, adding a broader view of election results beyond the candidates who win or lose. The project, called Times Voter Trends, is designed to examine which groups shifted their support, stayed home or helped determine the outcome of individual races.
The project will organize estimates by age, race, education and gender. Preliminary findings are planned for late on election night, when vote counts and turnout patterns begin to clarify the makeup of the electorate. The effort is intended to provide a state-by-state portrait of demographic voting behavior rather than a single national snapshot.
The project is also releasing retrospective findings from the 2020 and 2024 presidential elections. Those earlier analyses are meant to provide a baseline for comparing demographic changes in the midterm electorate. The released examples include estimates involving Donald Trump’s support among Hispanic voters and Kamala Harris’s performance among white voters in the Sun Belt.
The estimates combine several types of information: national polling, precinct-level election returns, voter-file data and a live election-night model. Precinct-level returns can provide geographic detail, while voter files help connect turnout patterns with demographic characteristics. The project’s underlying election-data work includes standardized precinct results and geographic files, with caveats where jurisdictions report absentee, provisional or other ballots at broader geographic levels.
Similar voter-file analyses have used verified state records to compare who voted in 2020 and 2024, who participated for the first time and where each party gained or lost voters.
The analytical estimates have been used internally for about a decade and have informed polling and election-night modeling. Their public release represents a change in timing and scope: readers will be able to examine demographic estimates for congressional elections soon after polls close, alongside comparable presidential-election findings from the two previous cycles.