Why Polls Tend to Undercount Democrats

[Adam Griffin]

You must be logged in to read this quote.

http://www.nytimes.com/2014/10/30/upshot/why-polls-tend-to-undercount-democrats.html

[Senate Minority Leader Lord Voldemort]

Now that the Democrats are behind, it's their turn to whine about bad polling.

[Maxwell]

We'll see if that's true this time.

[Bandit3 the Worker]

I don't think polls undercount Democrats in midterms. Aren't the results in midterms much more Republican than the polls predicted? I know that in 2010 they were way off. The GOP won some seats that weren't even close in the polls (for instance, the seat in Duluth and Boucher's seat in Virginia).

[Attorney General, LGC Speaker, and Former PPT Dwarven Dragon]

Last time, it was the republicans saying even in the final days they'd defy the odds and win the presidency (in a landslide) and the senate. Instead, the democrats swept everything. Now, it's the democrats saying they'll defy the odds and hold the senate (and depending on who you talk to, have house net pickups and make significant net pickups among the governors (>3)). Seems like we're going to see a very good night for republicans.

[backtored]

So, basically, Ken Buck?

[Adam Griffin]

Now that the Democrats are behind, it's their turn to whine about bad polling.

I don't think polls undercount Democrats in midterms. Aren't the results in midterms much more Republican than the polls predicted? I know that in 2010 they were way off. The GOP won some seats that weren't even close in the polls (for instance, the seat in Duluth and Boucher's seat in Virginia).

Last time, it was the republicans saying even in the final days they'd defy the odds and win the presidency (in a landslide) and the senate. Instead, the democrats swept everything. Now, it's the democrats saying they'll defy the odds and hold the senate (and depending on who you talk to, have house net pickups and make significant net pickups among the governors (>3)). Seems like we're going to see a very good night for republicans.

You must be logged in to read this quote.

False equivalency, guys. Unlike the Republicans' claims in 2012:

1) there is actually a trend in polling that backs up this for Democrats
2) several real pollsters are actually expressing agreement with the sentiment

But I wasn't necessarily saying that this would make the difference in Senate control. If we had a 2010-like discrepancy, however, then it likely would.

[Bandit3 the Worker]

Also, polls and the actual results are both very different from what most Americans really think. Polls are weighted for past election results, and we know how elections are rigged. This has gone on my whole adult life, which is many election cycles, so God only knows how far off it really is.

[Mr. Reactionary]

http://unskewedpolls.com/

[King]

I'll believe it when I see it.

[Mr. Reactionary]

1) there is actually a trend in polling that backs up this for Democrats

Wow ... 2 whole elections ... what a trend.

[Chancellor Tanterterg]

So, basically, Ken Buck?

Next time, try reading the article.

Also this:

I'll believe it when I see it.

[politicus]

Also, polls and the actual results are both very different from what most Americans really think. Polls are weighted for past election results, and we know how elections are rigged. This has gone on my whole adult life, which is many election cycles, so God only knows how far off it really is.

In what way?

[Mehmentum]

I wouldn't discount the possibility completely because there's always a chance of this happening (in either direction), but like King said, I'll believe it when I see it.

[Bandit3 the Worker]

Also, polls and the actual results are both very different from what most Americans really think. Polls are weighted for past election results, and we know how elections are rigged. This has gone on my whole adult life, which is many election cycles, so God only knows how far off it really is.

In what way?

Voter suppression, for one thing.

[IceSpear]

I like how people ignore the actual hard data to whine about "unskewed polls" or whatever.

You must be logged in to read this quote.

These are facts. Nobody disputes this. Does this mean it's guaranteed to happen again in 2014? No. Does it mean it's more likely for polls to underestimate Democrats than to underestimate Republicans? Certainly.

I think the bottom line is the "likely voter" screens. While LVs are much more reliable than RVs close to an election, an election isn't just likely voters. It's likely voters plus some "unlikely voters". These "unlikely voters" tend to skew Democratic, thus Democrats outperform the polls. It also depends on the pollster. Most of them don't even bother to disclose their likely voter screens so we have no idea how strict they are. Some of them have absolutely ridiculous ones like "MUST have voted in 2010" or "MUST have voted in the last 10 elections" which are obviously going to give heavily Republican results.

[IceSpear]

I don't think polls undercount Democrats in midterms. Aren't the results in midterms much more Republican than the polls predicted? I know that in 2010 they were way off. The GOP won some seats that weren't even close in the polls (for instance, the seat in Duluth and Boucher's seat in Virginia).

Huh? Everybody knew those were toss ups going into election day (though it wasn't known until late in the cycle).

http://www.realclearpolitics.com/epolls/2010/house/va/virginia_9th_district_griffith_vs_boucher-1390.html

http://www.realclearpolitics.com/epolls/2010/house/mn/minnesota_8th_district_cravaack_vs_oberstar-1712.html

[Bandit3 the Worker]

I don't think polls undercount Democrats in midterms. Aren't the results in midterms much more Republican than the polls predicted? I know that in 2010 they were way off. The GOP won some seats that weren't even close in the polls (for instance, the seat in Duluth and Boucher's seat in Virginia).

Huh? Everybody knew those were toss ups going into election day (though it wasn't known until late in the cycle).

http://www.realclearpolitics.com/epolls/2010/house/va/virginia_9th_district_griffith_vs_boucher-1390.html

http://www.realclearpolitics.com/epolls/2010/house/mn/minnesota_8th_district_cravaack_vs_oberstar-1712.html

That wasn't really how I remember it. Back in 2010, I just assumed the Democrats would win those seats, plus others they ended up losing. I was absolutely shocked by the results that year, especially considering the way public opinion was going.

Now I know to just assume that very close races will go Republican - unless there's very strong evidence otherwise.

[Fmr President & Senator Polnut]

http://unskewedpolls.com/

Actually, no. Part of the reason why I, and many Obama supporters were confident he'd win was because in a number of key states, Colorado and Nevada being key, and their history of underestimating Democrats for a number of reasons.

Unskewed Polls was based on GOP fantasy and an inability to read trends and party ID graphs.

I do think Democrats Will out-perform expectations - but not to the extent of changing too much in the end.

[memphis]

Long story short, Republicans are still using landlines? I wonder if they also send telegrams.

[Bandit3 the Worker]

Anyone else remember them good ol' days when cellphone folks were overwhelmingly GOP? Rush Limbaugh even used to brag about accepting cellphone calls.

[ElectionsGuy]

Now that the Democrats are behind, it's their turn to whine about bad polling.

[Fmr President & Senator Polnut]

Now that the Democrats are behind, it's their turn to whine about bad polling.

It's easy to be dismissive, except the complaints about Democratic under-polling are based in fact.

But we'll see.

[KCDem]

Can't wait until the Republicans and fake independents on this forum get taken to the woodshed on Tuesday.

[DrScholl]

It's turnout, turnout, turnout. In close races where get out the vote operations contacting every possible voter, there is no way that likely voter polls can pick up on last minute voters or even have contact information.