Quinnipiac Pollster Admits: ‘Probably Unlikely’ That Electorate Will Feature Massive Dem Skew

NewsBusters ^ | 9-26-2012 | Matthew Sheffield

[smoothsailing]

LOL, you picked up on that! That was the subtle point I was trying to make. Glad you saw it! :)

[Perdogg]

I am not sure I know what you mean.

[marktwain]

People who voted Democrat are not the same as Democrat voters in the polls being discussed.

I do not think that you are talking about the number of votes for Obama vs McCain, are you?

Certainly the overall vote was much higher for Obama, but that is not what the discussion is about.

[muawiyah]

If you call Manhattan only numbers ~ if such a thing is actually possible anymore ~ you have 1 in 10 who are Republicans. A poll taken there would look quite reasonable with 9 Dems for each 1 Rep. If there were a 50/50 split, that would be unreasonable.

[Perdogg]

Even in the bogus polls, The GOP is going more for Romney than the Dems are going for 0bamugabe

[Red Steel]

Not true, at least according to Dick Morris. DM claims Ras has a +2.5% Dem skew.

I believe that's true too. Once a pollster sets his baselines for an election, they don't usually change their methodology.

[muawiyah]

i find it difficult to believe that the Second Coming won't happen between now and then either ~ but it might not and then where are we eh!

Polls like the ones we are discussing are NOT PREDICTIVE ~ PRESCRIPTIVE PERHAPS ~ telling Romney NO MORE MR NICE GUY ~ OR ~ Start the campaign for gosh sakes.

[smoothsailing]

Read the article, it will help you answer your questions.

Helpful hint: The answer is in the title.

[marktwain]

You cannot sit down and randomly call just Democrats or just Republicans unless you have lists composed exclusively of such folks and all you want is a random sample of current Democrat or Republican thought.

Lists of registered Democrats, Republicans, and Independents are readily available. Calling those lists randomly in predetermined percentages designed to show what you expect turnout to be, based on party affiliation, is easily done.

[muawiyah]

A discussion is whether or not it is rational for a pollster to accept as probably meaningful a poll where there are 9% more Democrat respondents than Republican respondents.

We have people who seem to believe THAT CAN'T BE but, of course, it's always true ~ because there are more Democrats.

[tatown]

Except there ain’t D+9 or D+13 party affiliation numbers nationally that support the LV models the pollsters are using. Have you noticed that Romney’s numbers have gotten ‘worse’ since the pollster have went to LV models? It’s because they have gone more dem that a ‘random’ same just like the example you just gave.

http://m.rasmussenreports.com/public_content/politics/mood_of_america/partisan_trends

Using numbers party ID numbers from the best pollster on the planet, can you tell us why a D+9 or D+13 model is statistically valid? Well neither can the pollsters that are publishing them.

[muawiyah]

Alas, not universally true, and in many places where they do, in fact, have such lists you can't necessarily get access to them without handing over lots of money.

In states without party registration, there are no such lists.

[marktwain]

but, of course, it's always true ~ because there are more Democrats.

Not any more, according to Rassmussen. That organization shows more Republicans than Democrats for the last nine months.

[Smokin' Joe]

Well, that shows the demographic would work out to be biased and that could affect the outcomoooiiiiioo. Numerous prefixes would give similar results, and some Area Codes...

[italianquaker]

Maybe for once see a poll that reflects the accurate party breakdown on planet Earth for once. The only time they did more Rs was to keep Atkins in the race.
“Ahhh they didnt think the could fool a Corleone did they “lol

[tatown]

It’s been posted many time but he refuses to acknowledge it.

[marktwain]

Do you think Rassmussen is skewing data toward Republicans? Is there some dynamic that I am missing there?

[muawiyah]

Michael Reagan was on FOX today explaining that there have been more Democrats than Republicans for a long time ~ and his dad overcame the odds.

He won by being a friendly face with some good ideas and he got more Republicans to turn out than you can shake a stick at.

Mike knows how he won, and so did he.

So, the Democrats have turned to using Michael Reagan as a source for talking points?

Well, just darned eh!

[muawiyah]

What you are missing are the subtle distinctions between one guy's choice of how to handle non responsive non responses and non responsive responses.

You need to distort anything ~ just count all your calls, or exclude some.

Rasmussen also seems to screen for age ~ which adds another factor to the poll ~ which means he has to make more calls.

He claims it improves his effectiveness ~ others disagree ~ they say more smaller polls and that will give you a trendline that's as close as Rasmussen by election day.

[muawiyah]

let me put it this way ~ they’re still gd democrats no matter what they say. they’re trying to hide from the abusive democrat robocall system ~ but we also have pollsters determining through questioning how the subject/victim actually voted in previous elections ~