Pew Research Center
Graded against the actual result across 8 races (from 28 polls, through 2020).
Every race (8)
Each race Pew Research Center 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 |
|---|---|---|---|---|
| 2020 US President | D+10.0 | D+4.4 | 5.6 | ✓ |
| 2016 US President | D+6.0 | D+2.2 | 3.8 | ✓ |
| 2012 US President | D+3.0 | D+3.9 | 0.9 | ✓ |
| 2008 US President | D+6.0 | D+7.4 | 1.4 | ✓ |
| 2006 PA Governor | D+29.0 | D+20.7 | 8.3 | ✓ |
| 2006 PA Senate | D+21.0 | D+17.4 | 3.6 | ✓ |
| 2004 US President | R+3.0 | R+2.4 | 0.6 | ✓ |
| 2000 US President | R+2.0 | D+0.5 | 2.5 | ✗ |
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 | 5 | 1.77 | -2.51 | 60% |
| 1–3 wk | 7 | 4.20 | -0.87 | 43% |
| 3–6 wk | 9 | 3.37 | -2.37 | 78% |
| 6–9 wk | 7 | 6.27 | +0.12 | 86% |
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) |
|---|---|---|---|
| 2000 | 6 | 3.0 | R+1.2 |
| 2004 | 7 | 3.8 | R+2.7 |
| 2008 | 6 | 4.1 | D+0.6 |
| 2012 | 4 | 4.2 | R+2.2 |
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.