If 2026 votes like the generic ballot, where does it land?

The Times' latest generic-ballot poll shows Democrats up 8 among likely voters, against an even presidential race in 2024. That 8-point national move is not spread evenly: Hispanic voters move 33 points, voters under 30 move 23, voters over 65 move 2. This applies each group's swing to each state's electorate and compares the result to a plain uniform swing.

D+8national generic ballot, 2026
D+33largest group swing: Hispanic voters
D+2smallest: voters 65 and older, and the Northeast
Weights are rescaled to sum to 100%. The model is recentred so the national result always equals the environment above.

Arizona

Click any state. Region: West

How each dimension pushes this state away from the national swing, in points

State 2024 Swing 2026 model Baseline Diff White Hispanic Black B.A.+ 65+

What the model does

Every state starts at its 2024 presidential margin. Four demographic dimensions from the Times table — race and ethnicity, age, education, and census region — are each applied to the state's estimated 2024 electorate, producing four implied swings. Those are combined using the weights above, then the whole map is shifted so the turnout-weighted national figure matches the generic ballot.

The dimensions overlap, so they are averaged rather than stacked. Stacking them would count the same Hispanic non-college voter three times and produce swings twice as large as the poll supports.

The baseline

The baseline view adds the same national swing to every state. It is the null hypothesis: if the composition of the electorate carried no information, the two maps would be identical. The fourth view shows the gap between them, which is where the model is actually making a claim.

Limits worth keeping in mind