The Polling Company Inc.
Graded against the actual result across 26 races (from 31 polls, through 2018).
Head-to-head vs the VotePredictor model
The fair, apples-to-apples test: on the 25 races The Polling Company Inc. 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 |
|---|---|---|
| The Polling Company Inc. | 6.08 | 80% |
| VotePredictor | 5.19 | 84% |
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 (25)
Each race The Polling Company Inc. 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 |
|---|---|---|---|---|
| 2018 IA-1 House | D+2.0 | D+5.1 | 3.1 | ✓ |
| 2018 NY-22 House | R+8.0 | D+2.3 | 10.3 | ✗ |
| 2018 MO Senate | R+3.0 | R+5.8 | 2.8 | ✓ |
| 2018 WV Senate | D+4.0 | D+3.3 | 0.7 | ✓ |
| 2017 VA Governor | D+1.0 | D+8.9 | 7.9 | ✓ |
| 2016 IA-1 House | R+11.0 | R+7.7 | 3.3 | ✓ |
| 2014 AZ Governor | R+7.0 | R+11.8 | 4.8 | ✓ |
| 2014 IA-1 House | R+1.0 | R+2.3 | 1.3 | ✓ |
| 2014 IA-4 House | R+13.0 | R+23.3 | 10.3 | ✓ |
| 2014 VA-10 House | R+16.0 | R+16.1 | 0.1 | ✓ |
| 2014 OR Senate | D+8.0 | D+18.9 | 10.9 | ✓ |
| 2012 IN Governor | R+9.0 | R+2.9 | 6.1 | ✓ |
| 2006 AZ-5 House | R+9.0 | D+4.0 | 13.0 | ✗ |
| 2006 CO-7 House | D+5.0 | D+12.9 | 7.9 | ✓ |
| 2006 CT-4 House | R+10.0 | R+3.4 | 6.6 | ✓ |
| 2006 GA-8 House | R+1.0 | D+1.1 | 2.1 | ✗ |
| 2006 IN-8 House | D+12.0 | D+22.0 | 10.0 | ✓ |
| 2006 KY-4 House | R+4.0 | R+8.3 | 4.3 | ✓ |
| 2006 OH-6 House | D+18.0 | D+24.2 | 6.2 | ✓ |
| 2006 PA-6 House | D+1.0 | R+1.3 | 2.3 | ✗ |
| 2006 TX-17 House | D+12.0 | D+17.8 | 5.8 | ✓ |
| 2006 MO Senate | D+9.0 | D+2.3 | 6.7 | ✓ |
| 2006 MT Senate | D+9.0 | D+0.9 | 8.1 | ✓ |
| 2006 NJ Senate | EVEN | D+16.0 | 16.0 | ✗ |
| 2006 PA Senate | D+16.0 | D+17.4 | 1.4 | ✓ |
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 |
|---|---|---|---|---|
| 1–3 wk | 5 | 4.66 | -0.41 | 80% |
| 3–6 wk | 22 | 6.10 | +0.36 | 77% |
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) |
|---|---|---|---|
| 2006 | 13 | 7.0 | R+3.6 |
| 2014 | 7 | 4.8 | D+1.4 |
| 2017 | 3 | 9.3 | R+9.3 |
| 2018 | 5 | 4.2 | R+2.8 |
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.