Talk Elections

It’s interesting to see the change in PPPs’ polls since they became the official pollster for the Daily Kos.

In the Colorado Senate race, their most recent PPP poll has the Democrat leading by one point, while the more recent Rasmussen and Ispos polls have the Republican leading by 1 to 3 points.

In the Nevada Senate race, PPP has the Democrat up by 2 points while Rasmussen has the Republican up by 3 points and Mason-Dixon has the Republican leading by 2 points.

In the Pennsylvania Senate race, PPP has the Democrat up by one point while Quinnipiac has the Republican up by 2 points and Rasmussen has the Republican leading by 10 points in their most recent polls.

In the Missouri Senate race, PPP has the Republican up by 7 points while Opinion Research has him up 13 points and Rasmussen shows the Republican up by 9 points.

In the West Virginia Senate race, PPP has the Democrat up by 3 points, with Rasmussen showing the Republican up by 7 points and Opinion Research has the race as a tie.

In the Florida Governor’s race, PPP has the Democrat leading by 5 points while Rasmussen has the Republican up by 6 points and Opinion Research has the Republican leading by 3 points.

In the Pennsylvania Governor’s race, PPP has the Republican nominee leading by 2 points while Morning Call has the Republican leading by 11 points, Rasmussen shows a Republican lead of 14 points and Magellan Strategies has the Republican nominee up by 10 points.

Now, if you had asked me six months ago to rate the survey research organizations listed, I would have rated PPP better than Ispos, Quinnipiac, Morning Call or Magellan.

However, recently, PPP has pretty consistently leaned more favorable to the Democrats than other polls, or than it did six months ago.

Most of the differences you've pointed out are well within the margin of error.

EPIC FAIL

Nice selective choices in terms of the polls used, by the way.

Certainly not "since they became the official pollster for the Daily Kos".

This most recent batch though... have to wait and see. Does look a little odd.

Of course Rasmussen is a Republican pollster (in the same sense that PPP is a Democratic pollster), and Magellan is a Republican pollster (in the ignore if you want to be on the safe side sense).

As usual, you simply disagree without citing any facts whatsoever.

I should point out that PPP does not poll every race, and recent (October) polls only are being used.

Now if PPP were operating within the MoE then they should have results which average higher than other pollers for Republicans in a number of faces.

LOL@the Missouri example.

Go back to 2009 (Virginia and New Jersey), when they (PPP) was very good (i.e. more  accurate than most surveys).

In the last few months PPP's polls have drifted to the Democrats.

The most telling is the Nevada poll.  Mason-Dixon is a top not6ch outfit, which limits itself to states it really knows.

PPP has been very accurate and fair, not every poll they put out is favorable to Democrats. The only way to find out who is most accurate is to wait for the election night results.

I think the only poll that was exaggerated was the Alex Sink poll. I think that Scott may have a narrow advantage at this point.

probably it's rasmussen which is leaning republican. PPP is my favourite pollster, by far. and they aren't more favorable to democrats than other polls (see washington senate, wisconsin senate or FL senate)

Nate Silver tends to agree (http://fivethirtyeight.blogs.nytimes.com/2010/10/19/pennsylvania-revisited/) with jmfcst, and indeed that might be the source of his material. :P

PPP has been by far the most accurate pollster when it comes to polling actual elections in the last year.

?

jmfcst hasn't posted in this thread.

I don't think you understand the difference between a house effect and inaccurate or "bad" polling. Just because you don't like the results of one firm's findings, it doesn't make them wrong. Yes, PPP has a house effect. So do all other polling firms.

What you don't address is longer term trends, which is probably a more accurate way of assessing pollster accuracy. If PPP were to find trends completely different from every other pollster, that means one of two things. Either they are catching on to something and everybody else is wildly off, or vice versa. If they are showing the same trends as other pollsters, albeit with a house effect (or, as Mr. Silver puts it, "systematic differences in the the way that a particular pollster's surveys tend to lean toward one or the other party's candidates"), that doesn't indicate bad polling.

http://www.pollster.com/blogs/how_pollsters_affect_poll_resu.php?nr=1

Many folks have accused Rasmussen of leaning Republican, and others PPP leaning Democratic.
Here are 9 races and the current results from PPP vs Rasmussen.
In 12 days we'll get election results and I will come bask to see which is more accurate:

NC-Sen Ras: Burr +14
NC-Sen PPP: Burr +8

PA-Sen Ras: Toomey +10
PA-Sen PPP: Sestak +1

WV-Sen Ras: Raese +7
WV-Sen PPP: Manchin +3

NV-Sen Ras: Angle +3
NV-Sen PPP: Reid +2

NV-Gov Ras: Sandoval +19
NV-Gov PPP: Sandoval +9

CO-Gov Ras: Hickenlooper +4
CO-Gov PPP: Hickenlooper +14

PA-Gov Ras: Corbett +14
PA-Gov PPP: Corbett +2

IL-Gov Ras: Brady +8
IL-Gov PPP: Brady +1

FL-Gov Ras: Scott +6
FL-Gov PPP: Sink +5

Are these actually the final polls of these states by each poster though, or will they be releasing another one?  A couple of years ago, I read a column by Mark Blumenthal about how the polling from different pollsters tends to converge towards the same answer in the final week, even while they diverged wildly throughout the rest of the campaign.  Since the final poll is the one that people use to measure "accuracy", it becomes impossible to measure which pollster was actually "right".

Unfortunately, there's no way to check the accuracy of a poll taken a month before the election.

First, welcome to the forum.

Second, yes I do know the difference between bad polling and "house effect."

What I pointed out was a significant shift in PPP's results in a consistent direction relative to other survey research firms.

As the change in PPP's results is relatively recent, it is impossible to deal with "long term trends," as there is not enought data.

I'd like to bump this one to the top because, in hindsight, it's absolutely hilarious.

http://fivethirtyeight.blogs.nytimes.com/2010/11/04/rasmussen-polls-were-biased-and-inaccurate-quinnipiac-surveyusa-performed-strongly/

PPP didn't lean Democratic at all. In fact, they did just the opposite.

Looks like PPP won!

Overtaken by events, CARL.

Haha, that's hilarious.

Never let the facts ruin a good story (Republican proverb).

Threads like this are awesome in retrospect. Unfortunately, two years from now, the same characters will be saying the same things.

Yeah, too bad PPP's very last batch of polls were so close to Rasmussen's. This is still hilarious though.

Accuracy fluctuates from election to election. IIRC, wasn't Rasmussen one of the most accurate pollsters a few elections back?

But, yeah, they did very well this election season.