Mason-Dixon Polling & Strategy
Graded against the actual result across 362 races (from 648 polls, through 2022).
Head-to-head vs the VotePredictor model
The fair, apples-to-apples test: on the 222 races Mason-Dixon Polling & Strategy actually polled, how its final poll's margin compared to what the VotePredictor model predicted for those same races.
| Model | Avg miss (pts) | Called right |
|---|---|---|
| Mason-Dixon Polling & Strategy | 4.58 | 84% |
| VotePredictor | 3.51 | 91% |
VotePredictor 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 (361)
Each race Mason-Dixon Polling & Strategy 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.
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 | 65 | 4.31 | +0.03 | 86% |
| 1–3 wk | 294 | 4.66 | -0.41 | 84% |
| 3–6 wk | 183 | 6.04 | +0.30 | 80% |
| 6–9 wk | 106 | 5.87 | -0.28 | 85% |
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 | 109 | 7.9 | R+5.0 |
| 2000 | 86 | 4.8 | R+0.6 |
| 2002 | 58 | 4.7 | D+1.7 |
| 2004 | 99 | 3.5 | R+0.3 |
| 2005 | 3 | 6.4 | R+6.4 |
| 2006 | 91 | 4.6 | R+2.3 |
| 2008 | 60 | 4.6 | R+1.9 |
| 2010 | 47 | 5.8 | D+2.0 |
| 2012 | 37 | 5.5 | R+3.5 |
| 2014 | 8 | 6.4 | D+5.9 |
| 2016 | 14 | 6.2 | D+5.9 |
| 2018 | 14 | 3.2 | R+1.3 |
| 2019 | 3 | 1.0 | D+0.4 |
| 2020 | 6 | 7.5 | D+7.1 |
| 2022 | 3 | 6.5 | D+6.0 |
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.