Alaska Survey Research
Graded against the actual result across 21 races (from 38 polls, through 2024).
Head-to-head vs VotePredictor Elections
The fair, apples-to-apples test: on the 20 races Alaska Survey 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 |
|---|---|---|
| Alaska Survey Research | 7.17 | 70% |
| VotePredictor Elections | 4.57 | 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 Alaska Survey 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 |
|---|---|---|---|---|
| 2024 AK President | R+10.0 | R+13.4 | 3.4 | ✓ |
| 2022 AK-1 House | D+1.6 | D+3.0 | 1.4 | ✓ |
| 2020 AK-1 House | D+2.0 | R+9.1 | 11.1 | ✗ |
| 2020 AK President | R+4.0 | R+10.1 | 6.1 | ✓ |
| 2020 AK Senate | R+4.0 | R+12.7 | 8.7 | ✓ |
| 2018 AK Governor | R+0.2 | R+7.0 | 6.8 | ✓ |
| 2018 AK-1 House | D+1.0 | R+6.6 | 7.6 | ✗ |
| 2016 AK-1 House | R+15.0 | R+14.3 | 0.7 | ✓ |
| 2016 AK President | R+5.5 | R+14.7 | 9.2 | ✓ |
| 2016 AK Senate | R+37.7 | R+32.7 | 5.0 | ✓ |
| 2014 AK-1 House | D+5.5 | R+10.0 | 15.5 | ✗ |
| 2014 AK Senate | D+7.9 | R+2.1 | 10.0 | ✗ |
| 2010 AK Governor | R+18.3 | R+21.4 | 3.1 | ✓ |
| 2010 AK-1 House | R+32.8 | R+38.4 | 5.6 | ✓ |
| 2010 AK Senate | R+22.1 | R+12.1 | 10.0 | ✓ |
| 2008 AK-1 House | D+8.0 | R+5.2 | 13.2 | ✗ |
| 2008 AK President | R+11.0 | R+21.5 | 10.5 | ✓ |
| 2008 AK Senate | D+1.0 | D+1.2 | 0.2 | ✓ |
| 2006 AK Governor | R+11.0 | R+7.4 | 3.6 | ✓ |
| 2004 AK Senate | D+1.6 | R+3.0 | 4.6 | ✗ |
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 | 6.94 | +1.87 | 58% |
| 3–6 wk | 26 | 6.02 | +0.28 | 89% |
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) |
|---|---|---|---|
| 2008 | 7 | 6.0 | D+6.0 |
| 2010 | 3 | 6.3 | R+0.4 |
| 2016 | 3 | 5.0 | D+1.2 |
| 2018 | 9 | 7.1 | R+1.7 |
| 2020 | 3 | 8.6 | D+8.6 |
| 2022 | 5 | 4.3 | D+3.2 |
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.