Updated September 29, 2026 · model inputs through September 27, 2026
Massachusetts Senate Forecast
‹ See all Senate racesSenator Ed Markey, a fixture of Massachusetts politics for decades, faces Republican John Deaton, a trial lawyer and political newcomer. Massachusetts is among the country’s bluest states, but Markey’s path still runs through a Democratic primary before the general election.
Forecast summary
- Model margin
- D+31.67
- Rating
- Safe D
- Polling average
- D+24.36
- Fundamentals
- D+36.96
- Matchup
- Ed Markey vs. John Deaton
COUNTY MAP
County forecast
How the statewide forecast translates across counties.
County estimates apply the statewide forecast to 2024 presidential results. They are not separate county forecasts.
Baseline: 2024 presidential two-party results. Boundaries: U.S. Census Bureau, 2024. Methodology and limitations.
FORECAST HISTORY
Model margin over time
Forecast history will appear after the first recorded model release.
View the recorded forecast history table
| Model date | Model margin | Polling weight | Fundamentals weight | Polling average | Fundamentals |
|---|
POLLING
Polling average
Included surveys over time.
MODEL BREAKDOWN
How the forecast is built
The polling average and fundamentals are blended at their published weights.
Other model inputs and adjustments—
Margins are D-minus-R points. Candidate strength is included in fundamentals; its separate score is not published. Values are rounded to two decimals.
POLLING INPUTS
Polls behind the average
Polling average
D+24.36
| Field dates | Pollster | Sample | Results |
|---|---|---|---|
| Fielded Aug 20–Aug 24, 2026 | UNH | 881 LV | D Ed Markey 49% · R John Deaton 30% · O Joe Tache 7% |
| Fielded Aug 5–Aug 12, 2026 | UMass Amherst | 445 LV | D Ed Markey 51% · R John Deaton 30% |
| Fielded Jun 18–Jun 23, 2026 | UNH | 633 LV | D Ed Markey 50% · R John Deaton 34% |
| Fielded May 26–May 29, 2026 | Boston Globe/Suffolk | 500 LV | D Ed Markey 55% · R John Deaton 30% |
The published average comes directly from the workbook. Polls are not equally weighted, and the browser does not recompute the forecast.
The published average comes directly from the workbook. Polls are not equally weighted, and the browser does not recompute the forecast.