Rasmussen Polls Biased

[Toro]

This is a bunch of nonsense.

Given the actual results of 52 D (including the 2 independents) vs. 46 R (Alaska and Washington are not scored yet), Rasmussen's predictions look pretty spot on:

Election 2010: Senate Balance Of Power - Rasmussen Reports

And? I predicted 51 Ds and spent exactly $0 to come to that conclusion.

Of course the company is going to spin however they can. But if they were consistently over-estimating Republican support, I can guarantee you that there will be a whole lot of meetings at the company to figure out why. The partisans will jump to their defense, no doubt, but a company exists based on the quality of their product, not based on the pompom wavings of their cheerleaders.

 

[LiberalNut]

there will be a whole lot of meetings at the company to figure out why.

Scott already knows why. He's the one who added the patches, likely voter definitions and turnout models. The polls were DESIGNED to be biased toward Republicans.

a company exists based on the quality of their product

But you have to take into account who the consumer is. Where do you see Scott most on the TV box? FOX news perhaps?

 

[saveliberty]

Toro uses a blog as a source? One from the NYT no less. lol

 

[NYcarbineer]

Rasmussen's polls have a dirty little habit of boosting Republicans early on, then closing back in on reality close to the election. That way he can pump up Republicans and still get decent marks on accuracy when he's up against being compared to actual results.

 

[Chris]

fivethirtyeight.com is the best.

 

[Toro]

Toro uses a blog as a source? One from the NYT no less. lol

Nate Silver is an excellent and widely respected statistician.

This isn't 1962 BTW.

 

[skookerasbil]

Except for Nevada..........Rassmussen was spot on for the mid-terms............just like in 2008 when the lefties loved him.

 

[Annie]

Ok, here's 538's predictions:

Here's what's happened so far:

Rasmussen had 1 less pu in Senate than happened. 538 had 3 more pu's than happened.

I believe they had House pu at 55, when indeed it ended up at 60 as of right now.

All I'm finding is that he under counted the final tallies.

I don't really care which pollster comes out on top, it does give trends in general. It seemed to me in waning days that Gallup was doing the best, perhaps longevity comes into play when the election is so off the rails?

 

[Toro]

Ok, here's 538's predictions:

Here's what's happened so far:

Rasmussen had 1 less pu in Senate than happened. 538 had 3 more pu's than happened.

I believe they had House pu at 55, when indeed it ended up at 60 as of right now.

All I'm finding is that he under counted the final tallies.

I don't really care which pollster comes out on top, it does give trends in general. It seemed to me in waning days that Gallup was doing the best, perhaps longevity comes into play when the election is so off the rails?

538 doesn't do polling. 538 does polling analytics.

Some on here don't understand how forecasting works. Anybody could predict that the Democrats would win Hawaii. The science is predicting by how much. If you predict by 15 and its 50, its irrelevant that you picked the winner. Nobody is going to pay for that. Its like predicting that Tiger Woods is going to beat me at golf. Big deal.

I don't know if I've ever posted about Rasmussen before. I've never taken a position on whether or not they are a good or bad pollster. I've always thought they were as good as the others. But I posted this because the Right has made the firm an issue, claiming that they were the best pollster. Of course, the fact that Rasmussen is seen as a conservative pollster, and the fact that (in this election) they were biased towards the Republicans just confirms the beliefs to the ultra-partisans. Now, they are defending Rasmussen without understanding the issue. The same thing happened in Canada during the last election with a Liberal pollster. All the hyper-partisan Liberals claimed that this one firm was the most accurate pollster in Canada because they were closest in 2006. Of course, the pollster wound up being one of the most inaccurate pollsters in the 2008 election.

 

[saveliberty]

Seens to me your complaining Rasmussen had their numbers right too early. Maybe by a couple of weeks. Sounds like excellent polling to me.

 

[NYcarbineer]

Seens to me your complaining Rasmussen had their numbers right too early. Maybe by a couple of weeks. Sounds like excellent polling to me.

You can't have numbers 'right' a couple weeks early. Polls are based on the premise 'if the election were held today'.

 

[KissMy]

Rasmussen Polls are more accurate than Real Clear Politics. RCP had Democrat Russ Carnahan blowing out the Republican by 16%.

AAF/Ayers (R) 8/16 - 8/20 . 400 LV 54 38 Carnahan +16
WeAskAmerica 8/17 - 8/17 1089 RV 48 39 Carnahan . +9

Now the election was so close that it is being contested.

Actual Results
RUSS CARNAHAN (D) . . . . . . . 39,991 Votes 48.52%
EDWARD MARTIN (R) . . . . . . . 39,754 Votes 48.23%

ACORN activist rolled a bunch of inner city votes in the final reporting precincts to bring Russ Carnahan from way behind to squeak out a narrow win.

 

[saveliberty]

Seens to me your complaining Rasmussen had their numbers right too early. Maybe by a couple of weeks. Sounds like excellent polling to me.

You can't have numbers 'right' a couple weeks early. Polls are based on the premise 'if the election were held today'.

Moving purpose for a poll now? I thought it had already been established that polls are valued for their accuracy to the actual election results. When you lefties make up your mind what you want pools to mean, get back to me.

 

[NYcarbineer]

Seens to me your complaining Rasmussen had their numbers right too early. Maybe by a couple of weeks. Sounds like excellent polling to me.

You can't have numbers 'right' a couple weeks early. Polls are based on the premise 'if the election were held today'.

Moving purpose for a poll now? I thought it had already been established that polls are valued for their accuracy to the actual election results. When you lefties make up your mind what you want pools to mean, get back to me.

No they are not. Where in hell did you ever get that idea? A poll in July is supposed to predict the election???

lol

 

[saveliberty]

You can't have numbers 'right' a couple weeks early. Polls are based on the premise 'if the election were held today'.

Moving purpose for a poll now? I thought it had already been established that polls are valued for their accuracy to the actual election results. When you lefties make up your mind what you want pools to mean, get back to me.

No they are not. Where in hell did you ever get that idea? A poll in July is supposed to predict the election???

lol

You are truly an idiot. It is meant to predict the election and the reason why those results are likely. Then, the candidates can modify their message or communication levels to create a better outcome. The next poll measures the success toward that goal.

 

[NYcarbineer]

Moving purpose for a poll now? I thought it had already been established that polls are valued for their accuracy to the actual election results. When you lefties make up your mind what you want pools to mean, get back to me.

No they are not. Where in hell did you ever get that idea? A poll in July is supposed to predict the election???

lol

You are truly an idiot. It is meant to predict the election and the reason why those results are likely. Then, the candidates can modify their message or communication levels to create a better outcome. The next poll measures the success toward that goal.

Only the final polls are measured against the actual election for their accuracy.

 

[JimH52]

They are a republican backed organization. They have good polling but they are biased in which polls they publicize. That is why I follow RealClearPolitics - Opinion, News, Analysis, Videos and Polls

They give several different pollings, identify if they are D or R, and then they take the average.

 

[edthecynic]

If RatMuffin wasn't biased Right the CON$ wouldn't cite him.

 

[saveliberty]

If RatMuffin wasn't biased Right the CON$ wouldn't cite him.

Actually, the last several polls I quoted were liberal leaning ones.

I was surprised to hear NBC start off election coverage last night by acknowledging the House was in Republican hands. It was obviously a little forced, but they did a passing job of being balanced last night.

 

[boedicca]

I don't know if I've ever posted about Rasmussen before. I've never taken a position on whether or not they are a good or bad pollster. I've always thought they were as good as the others. But I posted this because the Right has made the firm an issue, claiming that they were the best pollster. Of course, the fact that Rasmussen is seen as a conservative pollster, and the fact that (in this election) they were biased towards the Republicans just confirms the beliefs to the ultra-partisans. Now, they are defending Rasmussen without understanding the issue. The same thing happened in Canada during the last election with a Liberal pollster. All the hyper-partisan Liberals claimed that this one firm was the most accurate pollster in Canada because they were closest in 2006. Of course, the pollster wound up being one of the most inaccurate pollsters in the 2008 election.

Rasmussen uses a different sampling method in which he identifies Likely Voters. The other pollsters generally use Adults or Registered Voters. The Left doesn't like Rasmussen's method because it some of their constituents have poor voting participation patterns, hence they prefer the other methods which overstate their likelihood of winning.