ABC/Washington Post comes out ahead on 4 of 6 measures in the Crosstab Trust Index, which grades pollsters on how close their final polls came to actual results from 1998 to 2022.
| Measure | ABC/Washington Post | Suffolk |
|---|---|---|
| Trust Score | 80/100 ★ | 68/100 |
| Average error how far final polls missed the result | 3.1 pts ★ | 4.1 pts |
| Versus expected error compared with what its races predicted | 1.1 better ★ | 0.7 better |
| Consistency random miss once its tilt is removed | ±3 pts ★ | ±4 pts |
| Historical bias closer to even is better | D+1.5 | D+1.5 |
| Winners called | 67% | 81% ★ |
| Track record final polls analyzed, 1998–2022 | 69 polls | 60 polls |
| 2026 polls | 2 | 14 |
| 2026 lean versus other polls of the same races | — | D+1.3 |
| Race | ABC/Washington Post (latest) | Suffolk (latest) |
|---|---|---|
| Michigan Senate | El-Sayed 48, Rogers 45 (D+0.4 vs. avg) | El-Sayed 47, Rogers 40 (D+4.5 vs. avg) |
| Michigan Governor | Benson 53, James 41 (D+5.1 vs. avg) | Benson 51, James 34 (D+10.6 vs. avg) |
Compare them in the interactive Trust Index All pollster ratings
More comparisons: Suffolk University vs. YouGov · ABC News/The Washington Post vs. YouGov · Emerson College vs. Suffolk University · ABC News/The Washington Post vs. Emerson College · Suffolk University vs. University of New Hampshire Survey Center · ABC News/The Washington Post vs. University of New Hampshire Survey Center · New York Times/Siena College vs. Suffolk University · ABC News/The Washington Post vs. New York Times/Siena College
ABC/Washington Post has the smaller average error (3.1 pts vs. 4.1 pts). Overall, aBC/Washington Post comes out ahead on 4 of 6 measures in the Crosstab Trust Index (ABC/Washington Post: A, Trust Score 80; Suffolk: A-, Trust Score 68).
Yes: 2 races in Crosstab's data, including Michigan Senate, Michigan Governor.
Grades: Crosstab Trust Index, computed from FiveThirtyEight's pollster-ratings data (CC BY 4.0), 1998–2022; 2026 measures from Crosstab's polling data.
Updated 2026-11-11 · Crosstab, nonpartisan polling analysis.