Glengariff Group
Graded against the actual result across 20 races (from 33 polls, through 2024).
Head-to-head vs VotePredictor Elections
The fair, apples-to-apples test: on the 20 races Glengariff Group actually polled, how its final poll's margin compared to what VotePredictor Elections predicted for those same races.
| Model | Avg miss (pts) | Called right |
|---|---|---|
| Glengariff Group | 4.52 | 85% |
| VotePredictor Elections | 3.06 | 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 (20)
Each race Glengariff 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 |
|---|---|---|---|---|
| 2024 MI-10 House | R+3.8 | R+6.1 | 2.3 | ✓ |
| 2024 MI President | D+3.0 | R+1.4 | 4.4 | ✗ |
| 2024 MI Senate | D+3.7 | D+0.3 | 3.4 | ✓ |
| 2022 MI Governor | D+8.6 | D+10.5 | 1.9 | ✓ |
| 2022 MI-10 House | R+7.9 | R+0.5 | 7.4 | ✓ |
| 2022 MI-7 House | D+5.7 | D+5.4 | 0.3 | ✓ |
| 2020 MI President | D+7.7 | D+2.8 | 4.9 | ✓ |
| 2020 MI Senate | D+9.6 | D+1.7 | 7.9 | ✓ |
| 2018 MI Governor | D+12.5 | D+9.6 | 2.9 | ✓ |
| 2018 MI Senate | D+16.7 | D+6.5 | 10.2 | ✓ |
| 2016 MI President | D+11.6 | R+0.2 | 11.8 | ✗ |
| 2014 MI Governor | R+5.0 | R+4.1 | 0.9 | ✓ |
| 2014 MI Senate | D+15.1 | D+13.3 | 1.8 | ✓ |
| 2012 MI-3 House | D+1.0 | R+8.4 | 9.4 | ✗ |
| 2012 MI President | D+2.7 | D+9.5 | 6.8 | ✓ |
| 2012 MI Senate | D+14.9 | D+20.8 | 5.9 | ✓ |
| 2010 MI Governor | R+18.0 | R+18.2 | 0.2 | ✓ |
| 2010 MI-15 House | D+19.0 | D+16.7 | 2.3 | ✓ |
| 2006 IL Governor | D+9.0 | D+10.5 | 1.5 | ✓ |
| 2006 MI-7 House | R+8.0 | R+4.0 | 4.0 | ✓ |
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 | 14 | 4.36 | -0.71 | 86% |
| 3–6 wk | 15 | 5.57 | -0.17 | 87% |
| 6–9 wk | 4 | 3.49 | -2.66 | 100% |
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) |
|---|---|---|---|
| 2010 | 3 | 1.5 | D+0.1 |
| 2012 | 7 | 6.1 | R+2.1 |
| 2014 | 4 | 2.7 | R+1.8 |
| 2018 | 4 | 6.7 | D+6.7 |
| 2020 | 4 | 5.6 | D+5.6 |
| 2022 | 4 | 4.1 | R+0.6 |
| 2024 | 3 | 3.4 | D+3.4 |
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.