GfK Group
Historical accuracy is measured against certified results. Current-cycle records come from the canonical, deduplicated archive and do not by themselves change the sealed forecast.
- Historical polls
- 21
- Mean absolute error
- 3.90percentage points
- Historical bias
- D+0.39
- Represented cycles
- 52008–2016
- Current archive
- 0canonical records
- History span
- Oct 18, 2008through Oct 22, 2016
Known names and relationships
- Known aliases
- No additional reviewed aliases
- Organization type
- Single canonical organization
- Cycles represented
- 2008 · 2010 · 2012 · 2014 · 2016
21 eligible historical polls across 5 represented cycles passed the published history identity checks. Historical bias is descriptive and is kept separate from sponsor or client affiliation.
Current-cycle records
Archive source date 2026-09-09 · bundled copy. Archive records may be newer than the sealed forecast; inclusion here does not mean the forecast uses them.
0 shown · 0 total
No record in the currently authorized canonical archive is attached to this identity.
This is not evidence that the pollster released no 2026 polling; the source-audit gaps remain documented.
Eligible historical observations
21 shown · 21 total
| Cycle | Race | Type | Field end | Sample | Poll | Result | Error |
|---|---|---|---|---|---|---|---|
| 2016 | 2016_Pres-G_US | Pres-G | 2016-10-22 | 1,212 | D+13.5 | D+2.1 | D+11.4 |
| 2016 | 2016_House-G_US | House-G | 2016-10-22 | 1,212 | D+5.0 | R+1.1 | D+6.1 |
| 2014 | 2014_House-G_US | House-G | 2014-10-18 | 968 | R+8.0 | R+5.9 | R+2.1 |
| 2012 | 2012_Pres-G_US | Pres-G | 2012-10-21 | 839 | R+2.0 | D+3.9 | R+5.8 |
| 2010 | 2010_House-G_US | House-G | 2010-10-16 | 600 | R+7.0 | R+6.8 | R+0.2 |
| 2008 | 2008_Sen-G_VA | Sen-G | 2008-10-24 | 601 | D+26.0 | D+31.3 | R+5.3 |
| 2008 | 2008_Sen-G_NH | Sen-G | 2008-10-24 | 600 | D+6.0 | D+6.3 | R+0.3 |
| 2008 | 2008_Sen-G_NC | Sen-G | 2008-10-24 | 601 | D+4.0 | D+8.5 | R+4.5 |
| 2008 | 2008_Sen-G_CO | Sen-G | 2008-10-24 | 626 | D+12.0 | D+10.3 | D+1.7 |
| 2008 | 2008_Pres-G_VA | Pres-G | 2008-10-24 | 601 | D+7.0 | D+6.3 | D+0.7 |
| 2008 | 2008_Pres-G_PA | Pres-G | 2008-10-24 | 607 | D+12.0 | D+10.3 | D+1.7 |
| 2008 | 2008_Pres-G_OH | Pres-G | 2008-10-24 | 607 | D+7.0 | D+4.6 | D+2.4 |
| 2008 | 2008_Pres-G_NV | Pres-G | 2008-10-24 | 628 | D+12.0 | D+12.6 | R+0.6 |
| 2008 | 2008_Pres-G_NH | Pres-G | 2008-10-24 | 600 | D+18.0 | D+9.6 | D+8.4 |
| 2008 | 2008_Pres-G_NC | Pres-G | 2008-10-24 | 601 | D+2.0 | D+0.3 | D+1.7 |
| 2008 | 2008_Pres-G_FL | Pres-G | 2008-10-24 | 600 | D+2.0 | D+2.8 | R+0.8 |
| 2008 | 2008_Pres-G_CO | Pres-G | 2008-10-24 | 626 | D+9.0 | D+8.9 | D+0.1 |
| 2008 | 2008_Gov-G_NH | Gov-G | 2008-10-24 | 600 | D+56.0 | D+42.6 | D+13.4 |
| 2008 | 2008_Gov-G_NC | Gov-G | 2008-10-24 | 601 | 0 | D+3.4 | R+3.4 |
| 2008 | 2008_Pres-G_US | Pres-G | 2008-10-18 | 800 | D+1.0 | D+7.3 | R+6.3 |
| 2008 | 2008_House-G_US | House-G | 2008-10-18 | 800 | D+6.0 | D+11.1 | R+5.1 |
Method · historical final-three-week D/R general-election rows plus verified historical supplement; min n=5
Definition · bias_D is poll D-R margin minus actual D-R margin; positive overstates Democrats.
A missing sample, sponsor or source is labeled unavailable; it is never reconstructed.
Poll data via VoteHub, used under CC BY 4.0. Poll data from The New York Times, used under CC BY 4.0.