(Vanity) Rasmussen's Polling Techniques

Self | 10/22/08 | Self

[mn-bush-man]

For those who allow the tone for their day to be set by the lates poll information, please consider how much manipulation is actually occurring.

For example, Rasmussen has changed his measure of Rep/Dem in his methodology on a regular basis.

From his own reports: "In September, the number of Americans who consider themselves to be Republicans increased a percentage point from 33.2% in August to 34.4% in September."

"During that same period, the number of Democrats declined slightly from 38.9% to 38.4.%. The remainder, 27.2%, are not affiliated with either major political party."

"That gives the Democrats a net advantage of 4.0 percentage points, down from 5.7 percentage points in August."

Now look at how he has shifted his sampling with his most recent polling effort:

"For polling data released during the week of October 19-25, 2008, the partisan weighting targets used by Rasmussen Reports will be 39.7% Democratic, 33.0% Republican, and 27.3% unaffiliated."

A 5.7% Dem advantage in August, a 4.0% Dem advantage in September, and a 6.7% advantage in October, plus, one has to wonder whether those who profess to be "unaffiliated" are TRULY unaffiliated.

Can you say SECRET SAUCE!!

[Maverick68]

Please go away. We only allow bad news as it relates to polling as many find the tast of perpetual misery intoxicating. Just kidding, thanks for the info...

[pgkdan]

Maybe somebody can enlighten me. Why adjust for party id at all? Why not simply take a random sample and publish that?

[ncalburt]

exactly, Ras is peddling the Obama media talking points and Obama skewed polls !
its a fraud.

if McCain wins ,his first act should be to BAN polling 30 days before an election like the rest of the world.

[forYourChildrenVote4Bush]

“...A 5.7% Dem advantage in August, a 4.0% Dem advantage in September, and a 6.7% advantage in October...”

Does this number reflect Virginia, CO, Florida, Ohio, or is this National.... I really dont understand these polls????? How are STATE polls "sauced"

[Ikemeister]

What is really starting to bother me is the citing of the RealClearPolitics “Poll Average” by the media (including FOXNEWS) to show how far ahead Obama is. This average includes polls that are obvious outliers and also data that is out-of-date.

[GoMonster]

Only places that do random sample is SurveyUSA. I think Gallup had in past used random sample as well.

[SoConPubbie]

if McCain wins ,his first act should be to BAN polling 30 days before an election like the rest of the world.

How about we leave the totalitarian stuff to the libtards, k?

[andy58-in-nh]

Let's put it this way: if the Obama camp believed these polls, would they still be campaigning in Pennsylvania or Wisconsin?

[mn-bush-man]

Agreed — even Drude is citing and linking to it! Terrible!

[norton]

Be sure to share this with your liberal associates when they claim that Acorn and it's kin are not involved in vote fraud: inflated registration numbers mean inflated wighting of polls, which means they are a direct means of influencing voters that their puppet is the sure winner - that "Resistance is futile".

[top 2 toe red]

exactly, Ras is peddling the Obama media talking points and Obama skewed polls ! its a fraud.

Worth repeating...exactly!

Every time I have to listen to Scott Rasmussen spout the DNC/obama talking points, I have to restrain myself to keep from throwing something thru the TV!

[Ikemeister]

The leaked Obama internals and the Rendell request for him to return to PA supports your point. PA is in play.

[I_hate_politicians]

The only poll that matters is the one on November 4th. It’s no longer about McCain or Obama. It’s about America. I will gladly hold my nose and vote for John McCain because the alternative is just dreadful.

[I_hate_politicians]

Don’t go to the Drudge website. Let it wither and die on the vine. I’ve written off Fox as well after their performance at the last debate. No news, entertainment, sports, etc. Let it hurt them in the pocketbook.

[JLS]

I was thinking some about the reported polling result after a couple of post analyzing some of the consistent bias against the GOP. Then I remembered that pollsters are finding it harder and harder to get people to answer their polls.

So when they report results like McCain 44%, Obama 50%, undecided 6%, the real results were McCain 22%, Obama 25%, undecided 3%, refused to answer, not home, hung up the phone etc 50% or maybe even far worse than 50%. If to get a sample of 500 answers it takes 1000 phone calls, then what do you have? I am not sure.

If you take the leap of faith that refused to answer not home, hung up the phone etc are randomly distributed in the population, then I guess you can argue for the point estimates. But are the sampling errors still the same?

[Gypssy]

Let's put it this way: if the Obama camp believed these polls, would they still be campaigning in Pennsylvania or Wisconsin?

Sadly, because Obama has so much money, he could actually campaign full blast in every state and still not be able to spend his money.

[andy58-in-nh]

It's not the money - it's that he is physically demanded there. Of course they have enough money to make advertising buys in every state - it is the choice of which states to visit that reveals the campaign's thought process and areas of concern.

[PghBaldy]

It also has no margin of error which would be quite difficult to get...it’s meaningless except for maybe a trend but it is so close, it is meaningless...

[latina4dubya]

if McCain wins ,his first act should be to BAN polling 30 days before an election like the rest of the world.

the best thing would be if conservatives got into the polling business... i have a 12-year old son who is good with numbers... perhaps this is an opportunity for him in the future... we need our children to become pollsters, journalists, movien directors, etc...