So umm.... what the hell happened with the polling?

[Jimmy Higgins]

The Republicans won in '14. Won pretty well. What is odd is that the polling was off on almost all the key races, and significantly so... about at least 4 or 5 points in Colorado, Georgia, Iowa, Virginia, Kansas, Kentucky (way off). New Hampshire was right and North Carolina was close.

So what the heck happened? Did the polls over-estimate liberal turnout for each and every one of those states?

 

[Trausti]

Let Nate Silver explain: http://fivethirtyeight.com/features/the-polls-were-skewed-toward-democrats/

 

[Jimmy Higgins]

My recollection from '12, '10, '08, '04 was the polling was relatively accurate. There were outfits that may have been problematic (Rasmussen), but overall, the polling as shown at Electoral-Vote.com has been pretty darn close. One race may go one way, one race could go the other, but in general, you had a good feel for the predicted results. This time around, the polls were wickedly over biased for the Democrats... repeatedly.

I don't particularly care for the links analysis as it takes the last 21 days (it also doesn't answer my question). I'm interested in the last few days. KY, AR, IA, SD, GA were insanely off. Georgia and South Dakota can be explained in part due to the third party involvement.

From the link, what is really weird is the error for Senate in Colorado, but not the Governor!

 

[Tom Sawyer]

Well, polling is based on a guesstimate about which voters are going to bother to turn out. The GOP did a much better job of motivating their people to come out to the polls than the Dems did. If that superior campaign wasn't properly taken into account, the numbers will be off accordingly.

Also because most Republicans voted twice. VOTER ID LAWS NOW!!!!

 

[Jimmy Higgins]

Well, polling is based on a guesstimate about which voters are going to bother to turn out. The GOP did a much better job of motivating their people to come out to the polls than the Dems did. If that superior campaign wasn't properly taken into account, the numbers will be off accordingly.

Also because most Republicans voted twice. VOTER ID LAWS NOW!!!!

Certainly turnout was equivalent in NH and NC though.

I'm not suggesting malfeasance. I'm more curious as to the one-sided bias in this polling in many states. Or were the Republicans benefiting from a late wave of support from moderates because of the Ebola pandemic that is raging in the US?

 

[Tom Sawyer]

Well, polling is based on a guesstimate about which voters are going to bother to turn out. The GOP did a much better job of motivating their people to come out to the polls than the Dems did. If that superior campaign wasn't properly taken into account, the numbers will be off accordingly.

Also because most Republicans voted twice. VOTER ID LAWS NOW!!!!

Certainly turnout was equivalent in NH and NC though.

I'm not suggesting malfeasance. I'm more curious as to the one-sided bias in this polling in many states. Or were the Republicans benefiting from a late wave of support from moderates because of the Ebola pandemic that is raging in the US?

IIRC, the Democrats were counting on the impact of illegal immigrants, dead people and felons voting for them in order to put them over the top in close races. Since that's not actually a thing that happens, any polls which included all the dead people in their models would have been inaccurately skewed as a result.

 

[Jimmy Higgins]

Certainly turnout was equivalent in NH and NC though.

I'm not suggesting malfeasance. I'm more curious as to the one-sided bias in this polling in many states. Or were the Republicans benefiting from a late wave of support from moderates because of the Ebola pandemic that is raging in the US?

IIRC, the Democrats were counting on the impact of illegal immigrants, dead people and felons voting for them in order to put them over the top in close races. Since that's not actually a thing that happens, any polls which included all the dead people in their models would have been inaccurately skewed as a result.

But the dead people helped re-elect the Democrat Governor in Colorado... but not the Senator. The polling for the Governor was dead on, but off by 5 points for the Senator. What... dead people decided to not vote Democrat for the Senate race?

 

[Tom Sawyer]

IIRC, the Democrats were counting on the impact of illegal immigrants, dead people and felons voting for them in order to put them over the top in close races. Since that's not actually a thing that happens, any polls which included all the dead people in their models would have been inaccurately skewed as a result.

But the dead people helped re-elect the Democrat Governor in Colorado... but not the Senator. The polling for the Governor was dead on, but off by 5 points for the Senator. What... dead people decided to not vote Democrat for the Senate race?

Because Mitch McConnell is a zombie and they wanted to help a brother out.