Post-Game Analysis

Primary and special election polling is usually unreliable and erratic and this Tuesday, it was no exception. We declined to allocate undecideds in most of the primaries, due to erratic polling. Our metric was off by 8 points at most, which made the difference in close races such as PA-12 (special), KY-Sen (Dem) and PA-Sen (Dem). We called every other winner.

This post-game analysis will give an insight on the polling averages vs. the final outcomes and the best pollster of the race.

KY-Sen (GOP):
Metric: Paul 51 - Greyson 32 (Paul +19)
Final: Paul 59 - Greyson 35 (Paul +24)

Our metric was off by 5 points. We allocated undecideds and were off by only 2 points. But overall, the pollsters, even the best in the business, missed the margin. Magellan, a partisan firm, surprisingly nailed the "Rand-slide".

Most Accurate Pollster: Magellan (R) - Paul 55 - Greyson 30 (Paul +25)

KY-Sen (Dem):
Metric: Mongiardo 38 - Conway 34 (Mongiardo +4)
Final: Conway 44 - Mongiardo 43 (Conway +1)

Our metric was off by 5 points. We declined to allocate undecideds in this race. Conway surged just a few weeks before election day and hit a wall. But he managed to squeak through with a 1 percent win and was arguably the 2nd most shocking outcome of Tuesday. No one showed Conway with a lead, but Survey USA did show a 1 point race.

Most Accurate Pollster: Survey USA - Mongiardo 38 - Conway 37 (Mongiardo +1)

PA-Sen (Dem):
Metric: Specter 43 - Sestak 42 (Specter +0)
Final: Sestak 54 - Specter 46 (Sestak +8)

Our metric was off by 8 points. Specter led most polls and in the last week, showed a consolidation of his base, but that wasn't enough. The undecideds broke for Sestak in a big way. Heavy rain in Philadelphia suppressed turnout for Specter and his opponent's performance may have been overstated. Suffolk almost nailed the result.

Most Accurate Pollster: Suffolk - Sestak 49 - Specter 40 (Sestak +9)

PA-Gov (Dem):
Metric: Onorato 36 - Wagner 10 (Onorato +26)
Final: Onorato 45 - Wagner 24 (Onorato +21)

No surprises here. Our metric was off 5 points, but accurately predicted a win for Onorato in the 20's. Polling was all over the map. Rasmussen came closest but they were still 4 points off.

Most Accurate Pollster: Rasmussen - Onorato 34 - Wagner 17 (Onorato +17)

AR-Sen (Dem):
Metric: Lincoln 44 - Halter 34 (Lincoln +10)
Final: Lincoln 45 - Halter 43 (Lincoln +2)

Our metric was off 8 points. The surprises here lay with the margin of victory. Almost all pollsters showed a huge lead for Blanche Lincoln. Talk Business was the most conservative in its estimate for Lincoln's lead, but they were still 5 points off the final result.

Most Accurate Pollster: Talk Business - Lincoln 38 - Halter 31 (Lincoln +7)


AR-Sen (GOP):
Metric: Boozman 46 - Holt 14 (Boozman +32)
Final: Boozman 53 - Holt 17 (Boozman +36)

Our metric was off 4 points. Talk Business did a great job in the GOP side. They were off by only 2 points.

Most Accurate Pollster: Talk Business - Boozman 46 - Holt 8 (Boozman +38)

OR-Gov (GOP):
Metric: Dudley 39 - Alley 23 (Dudley +16)
Final: Dudley 40 - Alley 32 (Dudley +8)

Our metric was off 8 points. Survey USA, the nation's most reputable pollster, was surprisingly off. They overstated Dudley's support in a big way. Portland Tribune's poll, conducted around the same time, nailed the margin much better and were off by only 2 points.

Most Accurate Pollster: Portland Tribune - Dudley 33 - Alley 23 (Dudley +10)

OR-Gov (Dem):
Metric: Kitzhaber 57 - Bradbury 24 (Kitzhaber +33)
Final: Kitzhaber 66 - Bradbury 30 (Kitzhaber +36)

Our metric was off by 3 points. Survey USA nailed the Dem side of the Governor race. They were off by only 2 points.

Most Accurate Pollster: Survey USA - Kitzhaber 59 - Bradbury 25 (Kitzhaber +34)

PA-12:
Metric: Burns 45 - Critz 45 (Tie)
Final: Critz 53 - Burns 45 (Critz + 8)

The metric was off by 8 points in this special election. PPP released the last poll of the cycle, showing a Burns lead of 1 point. Our metric actually showed a very slight Critz win, but we decided to allocate the undecideds and gave Burns an edge. The heated primary between Sestak and Specter drove up turnout in the district and Critz won by 8 points, essentially picking up 100% of undecideds according to the metric. None of the recent polling nailed the race, but a month old poll from Global Strategy nailed the final result.

Most Accurate Pollster: Global Strategy - Critz 45 - Burns 37 (Critz +8)