Moore Information Group
Graded against the actual result across 39 races (from 49 polls, through 2022).
Head-to-head vs the VotePredictor model
The fair, apples-to-apples test: on the 30 races Moore Information Group 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 |
|---|---|---|
| Moore Information Group | 7.53 | 60% |
| VotePredictor | 4.10 | 83% |
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 (39)
Each race Moore Information Group 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 |
|---|---|---|---|---|
| 2022 IA-3 House | R+2.0 | R+0.7 | 1.3 | ✓ |
| 2022 GA Senate | R+5.0 | D+1.0 | 6.0 | ✗ |
| 2022 WA Senate | R+0.4 | D+14.5 | 14.9 | ✗ |
| 2020 AL President | R+17.0 | R+25.5 | 8.5 | ✓ |
| 2020 AL Senate | R+15.0 | R+20.4 | 5.4 | ✓ |
| 2020 ME Senate | EVEN | R+8.6 | 8.6 | ✗ |
| 2018 NV-4 House | R+4.0 | D+8.2 | 12.2 | ✗ |
| 2016 AZ-1 House | D+3.0 | D+7.3 | 4.3 | ✓ |
| 2016 MT-1 House | R+11.0 | R+15.6 | 4.6 | ✓ |
| 2016 AK President | R+3.0 | R+14.7 | 11.7 | ✓ |
| 2016 AK Senate | R+40.0 | R+32.7 | 7.3 | ✓ |
| 2014 AZ Governor | D+3.0 | R+11.8 | 14.8 | ✗ |
| 2014 CA-3 House | D+6.0 | D+5.4 | 0.6 | ✓ |
| 2014 WA-1 House | R+21.0 | D+10.1 | 31.1 | ✗ |
| 2014 AK Senate | R+4.0 | R+2.1 | 1.9 | ✓ |
| 2012 CA-21 House | R+20.0 | R+15.5 | 4.5 | ✓ |
| 2012 OR-1 House | D+4.0 | D+14.2 | 10.2 | ✓ |
| 2012 AZ President | R+4.0 | R+9.1 | 5.1 | ✓ |
| 2012 AZ Senate | R+3.0 | R+3.0 | 0.0 | ✓ |
| 2010 AZ Governor | R+17.0 | R+11.9 | 5.1 | ✓ |
| 2010 ID-1 House | D+6.0 | R+9.7 | 15.7 | ✗ |
| 2010 OR-5 House | R+4.0 | D+5.3 | 9.3 | ✗ |
| 2008 OR President | D+14.0 | D+16.3 | 2.3 | ✓ |
| 2008 OR Senate | R+4.0 | D+3.4 | 7.4 | ✗ |
| 2008 SD Senate | D+18.0 | D+25.0 | 7.0 | ✓ |
| 2006 OR Governor | EVEN | D+8.0 | 8.0 | ✗ |
| 2006 CA-50 House | D+4.0 | R+4.6 | 8.6 | ✗ |
| 2004 OR President | D+2.0 | D+4.2 | 2.2 | ✓ |
| 2004 WA President | D+2.0 | D+7.2 | 5.2 | ✓ |
| 2004 WI President | R+2.0 | D+0.4 | 2.4 | ✗ |
| 2002 AK Governor | R+16.0 | R+15.1 | 0.9 | ✓ |
| 2002 UT-2 House | D+8.0 | D+0.7 | 7.3 | ✓ |
| 2002 OR Senate | R+17.0 | R+16.6 | 0.4 | ✓ |
| 2000 MO-2 House | R+2.0 | R+12.9 | 10.9 | ✓ |
| 2000 WA-2 House | R+4.0 | D+4.1 | 8.1 | ✗ |
| 2000 NE President | R+24.0 | R+29.0 | 5.0 | ✓ |
| 2000 NE Senate | R+2.0 | D+2.2 | 4.2 | ✗ |
| 1998 MT-1 House | R+15.0 | R+8.6 | 6.4 | ✓ |
| 1998 WA Senate | D+2.0 | D+16.8 | 14.8 | ✓ |
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 | 10 | 5.71 | +0.64 | 60% |
| 3–6 wk | 31 | 7.34 | +1.60 | 65% |
| 6–9 wk | 7 | 4.97 | -1.18 | 71% |
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 | 5 | 6.3 | D+0.1 |
| 2002 | 3 | 2.8 | D+2.0 |
| 2004 | 4 | 3.7 | R+3.7 |
| 2008 | 4 | 6.8 | R+6.8 |
| 2010 | 3 | 10.1 | D+0.4 |
| 2012 | 4 | 5.0 | R+2.4 |
| 2014 | 5 | 9.9 | R+3.7 |
| 2016 | 6 | 5.8 | D+1.9 |
| 2020 | 3 | 7.5 | D+7.5 |
| 2022 | 7 | 6.6 | R+6.6 |
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.