Mitchell Research & Communications
Graded against the actual result across 24 races (from 76 polls, through 2024).
Head-to-head vs VotePredictor Elections
The fair, apples-to-apples test: on the 19 races Mitchell Research & Communications actually polled, how its final poll's margin compared to what VotePredictor Elections predicted for those same races.
| Model | Avg miss (pts) | Called right |
|---|---|---|
| Mitchell Research & Communications | 3.95 | 74% |
| VotePredictor Elections | 2.58 | 95% |
VotePredictor Elections aggregates all the pollsters, so it's expected to beat any single one on margin — that's the value of averaging. The honest comparison among forecasters is on the combined board.
Every race (24)
Each race Mitchell Research & Communications polled, scored on its final poll — the call right before the vote — against the actual Dem−Rep result. Click a race for its full detail.
| Race | Their call | Result | Miss | Call |
|---|---|---|---|---|
| 2024 AZ President | R+2.0 | R+5.5 | 3.5 | ✓ |
| 2024 MI President | D+1.5 | R+1.4 | 2.9 | ✗ |
| 2024 AZ Senate | D+3.5 | D+2.4 | 1.1 | ✓ |
| 2024 MI Senate | D+2.0 | D+0.3 | 1.7 | ✓ |
| 2022 MI Governor | D+2.0 | D+10.5 | 8.5 | ✓ |
| 2022 MI-7 House | R+0.9 | D+5.4 | 6.3 | ✗ |
| 2020 MI President | D+7.1 | D+2.8 | 4.3 | ✓ |
| 2020 MI Senate | D+5.4 | D+1.7 | 3.7 | ✓ |
| 2018 MI Governor | D+7.8 | D+9.6 | 1.8 | ✓ |
| 2018 MI Senate | D+3.0 | D+6.5 | 3.5 | ✓ |
| 2016 MI President | D+6.0 | R+0.2 | 6.2 | ✗ |
| 2014 MI Governor | EVEN | R+4.1 | 4.1 | ✗ |
| 2014 MI-11 House | R+12.0 | R+15.4 | 3.4 | ✓ |
| 2014 MI Senate | D+12.0 | D+13.3 | 1.3 | ✓ |
| 2012 MI President | D+5.0 | D+9.5 | 4.5 | ✓ |
| 2012 MI Senate | D+14.0 | D+20.8 | 6.8 | ✓ |
| 2008 MI-9 House | EVEN | D+9.5 | 9.5 | ✗ |
| 2008 MI President | D+16.0 | D+16.5 | 0.5 | ✓ |
| 2004 MI President | D+2.0 | D+3.4 | 1.4 | ✓ |
| 2002 MI Governor | D+13.0 | D+4.0 | 9.0 | ✓ |
| 2002 MI Senate | D+30.0 | D+22.7 | 7.3 | ✓ |
| 2000 MI President | D+5.0 | D+5.1 | 0.1 | ✓ |
| 2000 MI Senate | R+3.0 | D+1.6 | 4.6 | ✗ |
| 1998 MI Governor | R+32.0 | R+24.4 | 7.6 | ✓ |
Accuracy by time to election
Lower is better. Time to election runs right (election week) to left (~2 months out).
By the numbers
| Time to election | Polls | Avg miss | vs field | Called right |
|---|---|---|---|---|
| ≤1 wk | 21 | 3.60 | -0.64 | 67% |
| 1–3 wk | 32 | 4.72 | -0.34 | 69% |
| 3–6 wk | 17 | 6.01 | +0.31 | 71% |
| 6–9 wk | 6 | 7.20 | +1.06 | 83% |
vs field is this pollster's average miss minus all pollsters' at the same lead time — green beats the field, redtrails it. Our historical polls reach ~2 months out; earlier polling isn't in the record.
Track record by cycle — getting better?
| Year | Polls | Avg miss | Lean (house effect) |
|---|---|---|---|
| 1998 | 3 | 11.6 | R+11.6 |
| 2000 | 7 | 6.1 | R+6.1 |
| 2002 | 6 | 7.1 | D+7.1 |
| 2004 | 3 | 3.8 | R+3.8 |
| 2008 | 3 | 4.1 | R+4.1 |
| 2014 | 15 | 2.6 | R+0.7 |
| 2016 | 11 | 6.8 | D+6.8 |
| 2018 | 10 | 2.7 | D+0.2 |
| 2020 | 6 | 5.8 | D+5.8 |
| 2022 | 4 | 7.2 | R+4.5 |
| 2024 | 6 | 2.0 | D+1.2 |
Do we credit a pollster for fixing its bias? Each cycle, the model re-estimates every pollster's lean from all its earlier polls (walk-forward) and subtracts it before using the poll. We tested weighting recent cycles more — it doesn't help: a pollster's lean in one cycle barely predicts the next (correlation 0.28), so the swings above are mostly noise, and averaging over more history beats chasing the latest cycle. The all-time estimate we use came out within ~0.5% of the best option.