Franklin & Marshall College Center for Opinion Research
Graded against the actual result across 33 races (from 55 polls, through 2022).
Head-to-head vs the VotePredictor model
The fair, apples-to-apples test: on the 33 races Franklin & Marshall College Center for Opinion Research 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 |
|---|---|---|
| Franklin & Marshall College Center for Opinion Research | 5.09 | 88% |
| VotePredictor | 3.54 | 85% |
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 (33)
Each race Franklin & Marshall College Center for Opinion Research 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+22.0 | D+14.8 | 7.2 | ✓ |
| 2022 PA Senate | D+4.0 | D+4.9 | 0.9 | ✓ |
| 2020 PA-7 House | D+7.2 | D+3.7 | 3.5 | ✓ |
| 2020 PA President | D+6.0 | D+1.2 | 4.8 | ✓ |
| 2018 PA Governor | D+26.0 | D+17.1 | 8.9 | ✓ |
| 2018 PA Senate | D+15.0 | D+13.1 | 1.9 | ✓ |
| 2016 CO President | D+7.0 | D+4.9 | 2.1 | ✓ |
| 2016 PA President | D+11.0 | R+0.7 | 11.7 | ✗ |
| 2016 CO Senate | D+11.0 | D+5.7 | 5.3 | ✓ |
| 2016 PA Senate | D+12.0 | R+1.4 | 13.4 | ✗ |
| 2014 PA Governor | D+7.0 | D+9.9 | 2.9 | ✓ |
| 2012 PA President | D+4.0 | D+5.4 | 1.4 | ✓ |
| 2012 PA Senate | D+9.0 | D+9.1 | 0.1 | ✓ |
| 2010 PA Governor | R+16.0 | R+9.0 | 7.0 | ✓ |
| 2010 PA-11 House | R+7.0 | R+9.4 | 2.4 | ✓ |
| 2010 PA-3 House | R+6.0 | R+11.4 | 5.4 | ✓ |
| 2010 PA-7 House | R+3.0 | R+11.0 | 8.0 | ✓ |
| 2010 PA-8 House | R+14.0 | R+7.0 | 7.0 | ✓ |
| 2010 PA Senate | R+8.0 | R+2.0 | 6.0 | ✓ |
| 2008 PA-10 House | D+15.0 | D+12.7 | 2.3 | ✓ |
| 2008 PA-11 House | R+5.0 | D+3.3 | 8.3 | ✗ |
| 2008 PA President | D+13.0 | D+10.3 | 2.7 | ✓ |
| 2008 US President | D+5.0 | D+7.4 | 2.4 | ✓ |
| 2006 PA Governor | D+25.0 | D+20.7 | 4.3 | ✓ |
| 2006 PA-10 House | D+12.0 | D+5.9 | 6.1 | ✓ |
| 2006 PA-6 House | R+3.0 | R+1.3 | 1.7 | ✓ |
| 2006 PA-7 House | D+1.0 | D+12.8 | 11.8 | ✓ |
| 2006 PA-8 House | R+9.0 | D+0.6 | 9.6 | ✗ |
| 2006 PA Senate | D+15.0 | D+17.4 | 2.4 | ✓ |
| 2004 PA-13 House | D+18.0 | D+14.5 | 3.5 | ✓ |
| 2004 PA-17 House | D+20.0 | D+20.2 | 0.2 | ✓ |
| 2004 PA President | D+5.0 | D+2.5 | 2.5 | ✓ |
| 2004 PA Senate | R+21.0 | R+10.6 | 10.4 | ✓ |
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 | 20 | 5.27 | +0.20 | 85% |
| 3–6 wk | 14 | 5.74 | 0.00 | 79% |
| 6–9 wk | 21 | 5.22 | -0.93 | 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) |
|---|---|---|---|
| 2004 | 9 | 4.8 | R+2.3 |
| 2006 | 8 | 6.1 | R+3.5 |
| 2008 | 7 | 6.1 | R+4.7 |
| 2010 | 8 | 6.0 | R+0.8 |
| 2012 | 4 | 1.5 | D+0.8 |
| 2016 | 6 | 8.3 | D+8.3 |
| 2018 | 4 | 4.9 | D+4.9 |
| 2020 | 3 | 4.4 | D+4.4 |
| 2022 | 4 | 2.7 | D+0.9 |
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.