EPIC-MRA Statewide Poll of Likely November 2026 Voters (Aug 22–28, 2026)
Author: EPIC-MRA | Published: 2026-08-29 | Source: report
Summary
A live-interviewer telephone survey (80% cell phones) of 600 likely November 2026 Michigan voters, fielded August 22–28, 2026, margin of error ±4.0 points, published as a full frequency report. It is the only post-primary Michigan Senate poll in the corpus with verbatim question wording.
Senate ballot (Q14): “If the election for U.S. Senate were held today, would you vote for [ROTATE] Mike Rogers the Republican or Abdul El-Sayed the Democrat?” with an undecided push (“if you had to decide right now, which candidate would you lean toward?”). Result: Rogers 39% plus 4% lean = 43% total Rogers; El-Sayed 43% plus 4% lean = 47% total El-Sayed; 10% undecided/refused. EPIC-MRA’s own simulation allocating undecideds by partisan lean: El-Sayed 50, Rogers 45, 5 undecided. Governor (Q13): Benson 45 (40+5) to James 41 (35+6), 14 undecided; simulated 48–44.
Favorability (Q4–9, rotated): El-Sayed is recognized by 94% (6% do not recognize), 39% favorable (23% very), 38% unfavorable (32% very), 17% undecided. Rogers: 10% do not recognize, 29% favorable (14% very), 40% unfavorable (28% very), 21% undecided. Both are net-negative; El-Sayed’s very-unfavorable share (32%) is the highest of the four nominees, but so is his very-favorable (23%). Trump: 34/57, 48% very unfavorable.
For the wiki, this is the methodologically strongest post-primary reading: live interviewers, a likely-voter screen, published wording. It lands between AARP’s tie and MSU’s +5, with El-Sayed +4 with leaners and +5 simulated. It does not ask about socialism, DSA, or Israel.
Key Claims
- Senate: El-Sayed 47 / Rogers 43 (with leaners); 43/39 without; 10% undecided. n=600 LV, ±4.0, Aug 22–28, live phone, 80% cell.
- Simulated with undecideds allocated by partisan lean: El-Sayed 50 / Rogers 45.
- El-Sayed favorability 39/38 (very fav 23, very unfav 32); Rogers 29/40 (14/28); El-Sayed is better known (94% vs 90%).
- Governor: Benson 45 / James 41 (with leaners); simulated 48–44.
- Trump favorability 34/57; 48% very unfavorable.
- No socialism, DSA, or Israel items in the questionnaire.
Direct quotes worth preserving
“If the election for U.S. Senate were held today, would you vote for [ROTATE] Mike Rogers the Republican or Abdul El-Sayed the Democrat? … 43% TOTAL ROGERS … 47% TOTAL EL-SAYED … 10% Undecided/Refused” — EPIC-MRA frequency report, Q14
Why citable: the only verbatim ballot wording for this race in the corpus; note the party labels are read aloud and no ideological label is attached to El-Sayed.
Connections
Entities mentioned: Abdul El-Sayed, Democratic Party Events: 2026 Michigan Democratic Senate Primary Debates: Does socialism win elections in America?
Contradictions / Tensions
- None with other polls beyond ordinary house-effect spread (El-Sayed +1 AARP, +4 EPIC, +5 MSU, +7 Susquehanna). All four post-primary polls have him ahead; none has Rogers ahead.
- The 32% “very unfavorable” for El-Sayed is a datum for the GOP’s polarization frame (Abdul El-Sayed wins Michigan Senate primary — a major victory for progressives) that his topline lead does not show.
Notes
Live-interviewer design is the least exposed of the four to the online-panel self-selection mechanism Morris identifies (I reanalyzed the raw data from Wisconsin's primary polls. Here's what actually went wrong.); it still cannot fix within-cell nonresponse. The Hill’s Aug 29 write-up of this poll (“El-Sayed edges Rogers”) is 403 and unarchived; the frequency report is the primary anyway.