New York Times Details 2024 Voter Model Using 5 Data Sources
Updated
Updated · The New York Times · Oct 25
New York Times Details 2024 Voter Model Using 5 Data Sources
3 articles · Updated · The New York Times · Oct 25
Summary
The New York Times said its Voter Trends reports estimate each registered voter's likelihood of backing the Democratic, Republican or minor-party presidential candidate in 2024 and 2020.
Five main inputs drive the model: Times/Siena polls, L2 voter-file data, precinct results collected by The Times and the Voting and Elections Science Team, AP county results, and Census surveys.
L2 data anchors a synthetic population of registered voters, and vote-choice probabilities are modeled from self-reported demographics, voter-file records and geography before being adjusted to match precinct outcomes.
The paper does not provide a sampling margin of error because the estimates are model-based rather than sample-based, and it warns that small subgroups and states lacking Times/Siena polls or reliable partisanship data require extra caution.