Updated September 23, 2026 · model inputs through September 22, 2026 · latest included poll: September 20, 2026
Michigan Senate Forecast
‹ See all Senate racesThe forecast was recalculated with polling inputs through September 22. The projected margin moved D+3.33 → D+3.32 (R +0.01 points). The rating remains Tilt D. The county map is recalibrated to the September 22 statewide estimate; county baselines and geometry are retained.
Forecast summary
- Model margin
- D+3.32
- Rating
- Tilt D
- Polling average
- D+2.44
- Fundamentals
- D+6.03
- Matchup
- Abdul El-Sayed vs. Mike Rogers
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.
POLLING
Polling average
Included surveys over time.
POLLING INPUTS
Polls behind the average
18 polls included in the workbook average. Newest first; latest included field date: September 20, 2026.
Polling average
D+2.44
| Field dates | Pollster | Sample | Results | Model treatment |
|---|---|---|---|---|
| Sep 16–20, 2026 | Suffolk University | 500 LV | D Abdul El-Sayed 46.8% · R Mike Rogers 39.6% · O Minor candidates / undecided / refused 13.6% | Included in the Senate polling average. |
| 2026-09-17–2026-09-17 | Emerson | — LV | D Abdul El-Sayed 48% · R Mike Rogers 46% | Included in model |
| 2026-09-17–2026-09-17 | Washington Post | — LV | D Abdul El-Sayed 48% · R Mike Rogers 45% | Included in model |
| 2026-09-10–2026-09-10 | Trafalgar Group* | — LV | D Abdul El-Sayed 46% · R Mike Rogers 45% | Included in model |
| 2026-08-31–2026-09-06 | CNN | 843 LV | D Abdul El-Sayed 47% · R Mike Rogers 44% · O Other 9% | Included in model |
| 2026-08-31–2026-09-03 | Detroit News/Glengariff | 600 LV | D Abdul El-Sayed 44% · R Mike Rogers 46% · O Other 10% | Included in model |
| August 29, 2026 | Abacus Data | 300 LV | D Abdul El-Sayed 49% · R Mike Rogers 45% · O Other 6% | Included in the Senate polling average. |
| Fielded August 22–August 28, 2026 | EPIC-MRA | 600 LV | D Abdul El-Sayed 47% · R Mike Rogers 43% | Included in the Senate polling average. |
| Fielded August 10–August 20, 2026 | MSU IPPSR/OSR (YouGov panel) | 779 LV | D Abdul El-Sayed 50% · R Mike Rogers 45% | Included in the Senate polling average. |
| Fielded August 11–August 17, 2026 | Susquehanna | 800 LV | D Abdul El-Sayed 46% · R Mike Rogers 39% | Included in the Senate polling average. |
| Fielded August 9–August 11, 2026 | Fabrizio/Anzalone | 877 LV | D Abdul El-Sayed 48% · R Mike Rogers 47% | Included in the Senate polling average. |
| Fielded August 6–August 10, 2026 | FOX News | 1006 RV | D Abdul El-Sayed 47% · R Mike Rogers 51% | Included in the Senate polling average. |
| Fielded August 6–August 10, 2026 | TIPP | 1528 LV | D Abdul El-Sayed 45% · R Mike Rogers 42% | Included in the Senate polling average. |
| July 31, 2026 | Mitchell Research | 672 LV | D Abdul El-Sayed 45% · R Mike Rogers 45% | Included in the Senate polling average. |
| July 1, 2026 | Quantus Insights | 947 LV | D Abdul El-Sayed 45% · R Mike Rogers 44% | Included in the Senate polling average. |
| June 17, 2026 | Zenith Research | 602 LV | D Abdul El-Sayed 45% · R Mike Rogers 42% | Included in the Senate polling average. |
| Fielded May 20–May 28, 2026 | TIPP | 1154 LV | D Abdul El-Sayed 43% · R Mike Rogers 42% | Included in the Senate polling average. |
| May 14, 2026 | Mitchell Research | 607 LV | D Abdul El-Sayed 41% · R Mike Rogers 42% | Included in the Senate polling average. |
Excluded polls
| Field dates | Pollster | Sample | Results | Reason |
|---|---|---|---|---|
| Fielded July 24–July 31, 2026 | EPIC-MRA | 600 LV | D Abdul El-Sayed 43% · R Mike Rogers 46% | EXCLUDED |
| July 28, 2026 | Glengariff Group | 600 LV | D Abdul El-Sayed 39% · R Mike Rogers 49% | Missing source URL • EXCLUDED |
The published average comes directly from the workbook. Polls are not equally weighted, and the browser does not recompute the forecast.