Charlotte?
St. Louis??

Wicked interesting!

it makes sense to me to have your conventions in purple states which makes minneapolis the odd city out here....too bad florida isn't in the mix....if the dems can keep that blue, the gop is toast in the race to 270

It's Rasmussen, so take the results with a grain of salt

You keep writing this, and sure there's an obvious Rasmussen bias relative to other polls. But is there any evidence that Rasmussen's polls are more inaccurate than others?

for multiple reasons why you should take rasmussen results with a grain of salt click here:

http://www.fivethirtyeight.com/search/label/rasmussen

for multiple reasons why you should take rasmussen results with a grain of salt click here:...

538 has a pollster ranking with a "pollster introduced error". Rasmussen's PIE is 1.74. Gallup's is 1.66. That's pretty close to identical. Both did worse than Survey USA (1.19) and better than CBS/NYTimes (1.94) and Newsweek(2.19).

Based on that, it looks like Rasmussen is right in the ballpark with the other major polling firms.

that study wasn't of polling done 6 months out from the election, as you probably noticed...

for a good piece on the hows, whys and size of rasumssen's house effect this is worth the look

http://www.fivethirtyeight.com/2010/04/use-of-likely-voter-model-does-not.html

nate's conclusion for those pressed for time: "Moreover, most objective attempts to rate pollsters, including ours, rely on an evaluation of the accuracy of polls in the week or two immediately preceding an election (when pollsters have strong incentives to "behave" themselves). They may reveal little or nothing about the accuracy of polls months ahead of one. "

Rasmussen tends to structure their polling in a way that, according to some, overweights or overrepresents conservative responses. But even while they occasionally produce a whopper of an outlier, in the aggregate, they tend to be pretty good. Paul is clearly leading in KY, but actually not by as much as you might expect in such a conservative state. Hopefully the DNC and Senate campaign committee will invest plenty in highlighting Paul's wacky wingnuttery over the next few months.

Make it Cleveland! Give that poor thing a shot in the arm already, also my mother won't have to travel as far! You want to talk about recession effects?

that study wasn't of polling done 6 months out from the election, as you probably noticed...

I did, but I don't believe that any poll 6 months out from an election is very predictive of the election day result.

Steve and others still talk about them everyday, though. And by writing "It's Rasmussen, so..." every-single-time, he's implying that Rasmussen is particularly bad at measuring public opinion 6 months out. I don't see Nate making that argument.

If Rasmussen is much worse than others months out, but in the mainstream within weeks of the election, we should be able to see it in polling results over time. Their predictions would have to show a rapid movement towards Democrats in the weeks before the election relative to their earlier forecasts in order to end up with the same PIE as the other pollsters.

The alternative is that other pollsters show rapid movement towards Republicans in the final weeks. In which case Rasmussen would have the better argument.

I haven't seen anyone make that case yet.

rick scott says he was not charged, not even questioned about the $1.5 billion medicare scam his company committed while he was ceo. ok i believe him.

but please tell me this: how far out of touch do you have to be as a ceo to not know a thing about such a mammoth scam going on in your company? and do you want such a slug as your governor?

the argument that nate has made repeatedly is that rasumussen's polls in this cycle have an obvious house effect...."So far in the 2010 cycle, their polling has consistently and predictably shown better results for Republican candidates than other polling firms have."

and "the degree of the "house effect" seems to vary according to the nature of the candidate. When the candidate is a mainstream or liberal-ish Republican, Rasmussen's numbers don't seem to differ all that much from other pollsters; maybe a couple of points, on average. But if the candidate is more of an "insurgent" conservative with ties to or roots in the tea-party movement, you can see some really big differences."

and

"Rasmussen, thus far, has a Republican-leaning house effect of about 5 and 1/2 points. So if Rasmussen, for example, has a Republican leading by 7 points in a particular race, an average pollster would have the Republican ahead by only 1 or 2 points."

he also notes that the house effect has been much bigger than it has historically been from ras and also goes to great pains to explain "this does not necessarily imply bias."

i can't speak for steve, but that's what i think about when i read his "grain of salt" comment..

the argument that nate has made repeatedly is that rasumussen's polls in this cycle have an obvious house effect.

Sure, but we won't know whether this is really a house effect until after the election. My concern is that it's always too easy to simply dismiss polls that we don't like, which is what Steve is doing, here. This happens in virtually every election, Karl Rove's "math" being one of the more famous examples.

If you just want to convince yourself that you're on the winning side of the debate, then by all means, pick your favorite pollster and ignore all the rest. But for anyone who's interested in actually winning elections, it's really better not to ignore the worst case scenario.

i don't think steve is dismissing a poll he doesn't like, he's [i'm guessing] alerting us to suspicions he has about rasmussen...or that 5 point wobble nate has noticed....cheers

Jinchi, could you clarify the difference between what you wrote:

sure there's an obvious Rasmussen bias

and what Steve wrote:

It's Rasmussen, so take the results with a grain of salt

sure there's an obvious Rasmussen bias

I'm using the word in the sense of "instrumental bias" not "personal preference".

In other words, the results of Rasmussen's polls consistently show better numbers for Republicans relative to other pollsters, but that doesn't mean they are deliberately skewing the results, or that their results are wrong.

I'm also not plugging for Rasmussen. I'd just rather see some proof that their numbers are consistently wrong and not just that they make Sarah Palin happy.

I'm using the word in the sense of "instrumental bias" not "personal preference".

For a pollster, how is an "instrumental bias" any more acceptable than a "personal preference"?

I'm also not plugging for Rasmussen.

You're actually doing a better job of discrediting Rasmussen than Steve is.

Let me add, Jinchi, that any independent, professional pollster worth his or her salt would detect (and correct) any instrumental bias in their methodology long before you or I could see it.

Unless they had a personal preference for the results.

any independent, professional pollster worth his or her salt would detect (and correct) any instrumental bias in their methodology long before you or I could see it.

Every polling firm has a bias or a house effect if you prefer a different term. Those house effects are well known and predominantly result from the assumptions that go into things like "likely voter" models. Rasmussen's ends up showing Republicans doing better than other firms. CBS/NYT is distinctly more favorable to Democrats. That's why people do polls of polls, which give a better sense of overall trends. But there's no guarantee that the real answer is the average of all polling firms. We don't learn who was right until the day after an election. To date, Rasmussen has been about as accurate as other major firms, which is why I don't understand Steve's constant snark at them.

http://www.pollster.com/blogs/how_pollsters_affect_poll_resu.php

as nate silver notes, the "likely voters"/"all voters" methodology difference between pollsters doesn't explain the size of ras's house effect all be itself..

as for steve's "snark," since scott is a republican pollster who claims he's non-partisan and since all of his polls are outliers that always favor only one party, there is a lot of suspicion among non-republicans that he polls in such a way to get results that will win him drudge links and fauxnews appearances...that argument shows up on such a regular basis on various comment threads over at nate's place..and pretty much everywhere ras gets discussed on teh innertubes that's it's almost a cliche at this point....cheers

I don’t know if your still following this thread, Jinchi, but I wanted to respond.

Every polling firm has a bias or a house effect if you prefer a different term.

Not as consistently pro-GOP as Rasmussen (as your own link confirms, Rasmussen represents the "extreme" in Republican "house effect"). It's not just one particular election season either, it's year-after-year that his polling has this effect. For some strange reason, this effect seems larger the farther away the poll is from an election. And while I may be naïve for believing that a serious independent pollster would want to do something about it, I think its just as naïve to think its an unintended consequence.

I don't understand Steve's constant snark at them.

I don't think its snark; I think Steve is serious. But how about this, instead of saying:

"It's Rasmussen, so take the results with a grain of salt,"

Steve says, instead,

"It's Rasmussen, who consistently has anywhere from a +3 to a +4 pro-Republican house-effect, which Pollster.com's Charles Franklin indicates is the highest GOP skew among major independent pollsters, although we have no evidence that this is, in anyway, an intentional outcome that Scott Rasmussen personally desires,"

Cool?

"he's no-less accurate than other pollsters"

Those are election-eve polls whose accuracy can be immediately determined. This a poll six-months out; there is no way to determine how better, or worse, Rasmussen is at that, unless you get everyone together tomorrow in Kentucky and hold a mock election. With these type of polls, as with issue polls (unless they proceed a ballot-initiative), we have no way of knowing who is most accurate.

We do know, however, that Rasmussen has been caught manipulating his poll questions to favor Republicans over Democrats. Here's an example.

With a poll like the Kentucky senate race, where no other pollster is in the field yet, Rasmussen can effect early media coverage and public opinion. Do you think any media reference to this poll is going to include an explanation of “house effect”?