Muhlenberg College Institute of Public Opinion
Graded against the actual result across 24 races (from 64 polls, through 2022).
Head-to-head vs VotePredictor Elections
The fair, apples-to-apples test: on the 23 races Muhlenberg College Institute of Public Opinion actually polled, how its final poll's margin compared to what VotePredictor Elections predicted for those same races.
| Model | Avg miss (pts) | Called right |
|---|---|---|
| Muhlenberg College Institute of Public Opinion | 4.17 | 87% |
| VotePredictor Elections | 3.63 | 87% |
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 (24)
Each race Muhlenberg College Institute of Public Opinion 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 PA Governor | D+14.0 | D+14.8 | 0.8 | ✓ |
| 2022 PA-7 House | D+1.0 | D+2.0 | 1.0 | ✓ |
| 2022 PA Senate | EVEN | D+4.9 | 4.9 | ✗ |
| 2020 PA-7 House | D+13.0 | D+3.7 | 9.3 | ✓ |
| 2020 PA President | D+5.0 | D+1.2 | 3.8 | ✓ |
| 2018 PA Governor | D+21.0 | D+17.1 | 3.9 | ✓ |
| 2018 PA-7 House | D+7.0 | D+10.0 | 3.0 | ✓ |
| 2018 PA Senate | D+14.0 | D+13.1 | 0.9 | ✓ |
| 2016 PA President | D+4.0 | R+0.7 | 4.7 | ✗ |
| 2016 PA Senate | R+1.0 | R+1.4 | 0.4 | ✓ |
| 2014 PA Governor | D+12.0 | D+9.9 | 2.1 | ✓ |
| 2012 PA President | D+3.0 | D+5.4 | 2.4 | ✓ |
| 2012 PA Senate | D+6.0 | D+9.1 | 3.1 | ✓ |
| 2010 PA Governor | R+7.0 | R+9.0 | 2.0 | ✓ |
| 2010 PA-15 House | R+17.0 | R+14.6 | 2.4 | ✓ |
| 2010 PA Senate | R+4.0 | R+2.0 | 2.0 | ✓ |
| 2008 PA President | D+6.0 | D+10.3 | 4.3 | ✓ |
| 2006 PA Governor | D+25.0 | D+20.7 | 4.3 | ✓ |
| 2006 PA-8 House | R+5.0 | D+0.6 | 5.6 | ✗ |
| 2006 PA Senate | D+8.0 | D+17.4 | 9.4 | ✓ |
| 2004 PA-15 House | R+33.0 | R+19.2 | 13.8 | ✓ |
| 2004 PA President | D+2.0 | D+2.5 | 0.5 | ✓ |
| 2004 PA Senate | R+22.0 | R+10.6 | 11.4 | ✓ |
| 2002 PA Governor | D+11.0 | D+9.0 | 2.0 | ✓ |
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 wk | 5 | 2.75 | -1.53 | 100% |
| 1–3 wk | 27 | 3.42 | -1.65 | 78% |
| 3–6 wk | 19 | 3.98 | -1.76 | 100% |
| 6–9 wk | 13 | 4.01 | -2.14 | 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) |
|---|---|---|---|
| 2004 | 3 | 8.6 | R+8.6 |
| 2006 | 5 | 6.4 | R+4.5 |
| 2008 | 8 | 2.3 | D+0.4 |
| 2010 | 14 | 3.0 | R+1.2 |
| 2012 | 10 | 2.5 | R+0.8 |
| 2016 | 8 | 4.0 | D+4.0 |
| 2018 | 5 | 2.9 | D+1.7 |
| 2020 | 3 | 6.3 | D+6.3 |
| 2022 | 5 | 2.1 | R+2.1 |
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.