Congressional Forecast - MI, MN, WI

📅 Preview Mode: This forecast reflects structural baselines only. Live polling data and active forecast updates begin Late August 2026. Model methodology and historical context are fully functional.

Michigan (13 Districts)

MI-1

93% R

R+11.7

Likely R

R inc

MI-2

98% R

R+15.6

Safe R

R inc

MI-3

83% D

D+7.6

Likely D

D inc

MI-4

86% R

R+8.4

Likely R

R inc

MI-5

98% R

R+16.0

Safe R

R inc

MI-6

97% D

D+14.4

Safe D

D inc

MI-7

62% D

D+2.4

Tilts D

D inc

MI-8

82% D

D+7.2

Likely D

D inc

MI-9

99% R

R+18.2

Safe R

R inc

MI-10

82% R

R+7.1

Likely R

R inc

MI-11

93% D

D+11.5

Likely D

D inc

MI-12

99% D

D+21.1

Safe D

D inc

MI-13

99% D

D+20.6

Safe D

D inc

Minnesota (8 Districts)

MN-1

93% R

R+11.6

Likely R

R inc

MN-2

92% D

D+10.8

Likely D

D inc

MN-3

91% D

D+10.7

Likely D

D inc

MN-4

98% D

D+16.8

Safe D

D inc

MN-5

99% D

D+22.1

Safe D

D inc

MN-6

96% R

R+13.2

Safe R

R inc

MN-7

99% R

R+18.2

Safe R

R inc

MN-8

91% R

R+10.5

Likely R

R inc

Wisconsin (8 Districts)

WI-1

84% R

R+7.6

Likely R

R inc

WI-2

99% D

D+19.0

Safe D

D inc

WI-3

68% R

R+3.6

Leans R

R inc

WI-4

99% D

D+24.0

Safe D

D inc

WI-5

96% R

R+13.3

Safe R

R inc

WI-6

91% R

R+10.3

Likely R

R inc

WI-7

96% R

R+13.3

Safe R

R inc

WI-8

84% R

R+7.7

Likely R

R inc

Recent Polls

Date	District	Pollster	Sample	Dem	Rep	Margin
2026-01-28 00:00:00	MI-7	EPIC-MRA	600 LV	49%	46%	D+3
2026-01-25 00:00:00	MI-3	Mitchell Research	500 LV	52%	44%	D+6
2026-01-22 00:00:00	MN-2	SurveyUSA	550 LV	53%	43%	D+10
2026-01-20 00:00:00	WI-3	Marquette Law School	800 LV	51%	45%	D+6

Methodology

Structural Model (Elastic Net): dem_share = 48.1 + 0.41 × partisan_lean + 7.1 × incumbency
Partisan Lean: Based on 2022-2024 US House results relative to state average
Incumbency Bonus: +7.1 points for incumbent party (trained on 2022-2024 data)
Open Seats: Higher uncertainty, no incumbency advantage
Poll Blending: Structural 90% at 180+ days → 20% at 7 days
Monte Carlo: 10,000 simulations with within-state (0.85) and cross-state (0.60) correlations
Time-Varying Uncertainty: Error SD increases from ~4.7 near election to ~8.9 one year out

Rating Scale: Safe (>95%), Likely (80-95%), Leans (65-80%), Tilts (55-65%), Tossup (<55%)

Model Validation: Backtest on 2024 races: RMSE 6.1 pts, 89.7% correct winners