EMC Research
Graded against the actual result across 24 races (from 32 polls, through 2020).
Head-to-head vs VotePredictor Elections
The fair, apples-to-apples test: on the 18 races EMC Research actually polled, how its final poll's margin compared to what VotePredictor Elections predicted for those same races.
| Model | Avg miss (pts) | Called right |
|---|---|---|
| EMC Research | 7.65 | 56% |
| VotePredictor Elections | 3.92 | 78% |
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 EMC 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 |
|---|---|---|---|---|
| 2020 NC-11 House | D+3.0 | R+12.2 | 15.2 | ✗ |
| 2020 TX-25 House | R+2.0 | R+13.8 | 11.8 | ✓ |
| 2020 TX President | EVEN | R+5.6 | 5.6 | ✗ |
| 2014 MI-6 House | R+4.0 | R+15.5 | 11.5 | ✓ |
| 2012 CT Senate | D+6.0 | D+12.5 | 6.5 | ✓ |
| 2010 FL Governor | D+1.0 | R+1.1 | 2.1 | ✗ |
| 2010 SC Governor | R+5.0 | R+4.5 | 0.5 | ✓ |
| 2010 FL-24 House | D+2.0 | R+19.3 | 21.3 | ✗ |
| 2010 CT Senate | D+15.0 | D+9.3 | 5.7 | ✓ |
| 2008 FL-24 House | R+1.0 | D+16.1 | 17.1 | ✗ |
| 2008 FL-25 House | R+3.0 | R+6.1 | 3.1 | ✓ |
| 2008 FL President | D+4.0 | D+2.8 | 1.2 | ✓ |
| 2006 FL-13 House | D+11.0 | R+0.2 | 11.2 | ✗ |
| 2006 KY-2 House | R+8.0 | R+10.8 | 2.8 | ✓ |
| 2006 TN Senate | D+6.0 | R+2.7 | 8.7 | ✗ |
| 2004 TX-32 House | R+1.0 | R+10.3 | 9.3 | ✓ |
| 2004 FL President | R+2.0 | R+5.0 | 3.0 | ✓ |
| 2004 FL Senate | EVEN | R+1.1 | 1.1 | ✗ |
| 2002 FL Governor | R+5.0 | R+12.8 | 7.8 | ✓ |
| 2002 OK-4 House | R+1.0 | R+7.7 | 6.7 | ✓ |
| 2002 GA Senate | D+12.0 | R+6.8 | 18.8 | ✗ |
| 2000 WA Governor | D+24.0 | D+18.7 | 5.3 | ✓ |
| 2000 WA President | D+6.0 | D+5.6 | 0.4 | ✓ |
| 2000 WA Senate | R+4.0 | D+0.1 | 4.1 | ✗ |
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 | 12 | 6.75 | +1.68 | 50% |
| 3–6 wk | 14 | 8.05 | +2.31 | 43% |
| 6–9 wk | 6 | 8.70 | +2.55 | 67% |
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 | 4 | 2.9 | D+0.0 |
| 2002 | 3 | 11.1 | D+11.1 |
| 2004 | 4 | 5.1 | D+5.1 |
| 2006 | 7 | 8.2 | D+8.2 |
| 2008 | 3 | 7.1 | R+4.3 |
| 2010 | 5 | 7.1 | D+4.6 |
| 2020 | 4 | 12.2 | D+12.2 |
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.