Free Republic
Browse · Search
News/Activism
Topics · Post Article

Skip to comments.

FR EXCLUSIVE - Poll Weighting Demystified...and the Bug in Rasmussen Polling
FR | 10-18-08 | Bob J

Posted on 10/18/2008 9:51:46 AM PDT by Bob J

For those wanting to know the inside story on how and why polling companys weight their samples, Rasmussen has done a pretty good job of explaining it.

Here you can get an overall explanation of the methodology and rationale;

The Value of Party Weighting for a Tracking Poll

Next they explain why there may be large differences between polls;

Why Polls Sometimes Show Different Results ,p>

Rasmussen establishes the weighting for the polls by conducting a party affiliation poll of 15k every month. Here they show the results since 2004;

Summary of Party Affiliation

Here's the bug. Since 2004 their Party Affiliation numbers have see-sawed somewhat since they are based on political conditions at that time. IMO in 2004 they saw an increase in republican identification leading up to the election due to concerns over the war in Iraq/terror and doubts about Kerry. In 2006 there was a likewise trend toward dems due to electorate frustration about Bush over the WOT and the economy, leading to a dem victory in the midterms.

The party identification numbers in the last four years hav averaged...2004 = +2.4% dem, 2005 = +1.9% dem, 2006 was +4.1% dem and 2007 was +4.8% dem.

However in the first 2 quarters of 2008 there was a jump to +9% dem. Huh?

This jump in dem identification is not only historically without precedent, there is nothing from a national, cultural or financial perspective to explain it. In Jan 2008 we started to see the US winning in Iraq, the economy was basically sound, the DOW at 13000 and there were no pressing cultural issues. Rasmussen weakly explains this by saying the "Pubs were having a hard time".

Baloney. The only plausible explanation for such a dramatic and unprecedented change in party affiliation over such a short period of time is manipulation.

Now I don't know if this was being done by Rasmussen "Party Affilation" polling employees (how do you "weight" this poll, BTW?), Operation Chaos or dem/ACORN fraud, but for Rasmussen to use these obviously manipulated numbers as a basis for his entire polling methodology is at best suspect and at worst professional incompetence.

FReepers have been right all along. When looking at Rasmussens numbers we should skew at least 4% back toward McCain to balance the error in party affiliation.

At worst, McCuda are tied and probably up by 1-2%.


TOPICS: Free Republic; News/Current Events; Politics/Elections
KEYWORDS: poll; rasmussen
Navigation: use the links below to view more comments.
first previous 1-2021-4041-6061-80 ... 101-104 next last
To: Rockingham

Even if it was the case that GOP party ID is down — and this would flatly contradict actual election turnout data that shows ideology and party affiliation for both parties to be remarkably stable over decades — all this means is that some of the GOPers would switch to “Independent” and still vote reliably right of center. Most actual human beings rarely switch their political philosophies or party affiliations, and when they do it is very slowly and over a period of years rather than many times within the same election cycle.

The issue here is not whether there are more or fewer GOPers and Independent conservatives, but whether they will go out and vote. If they turn out at 1996 or 2000 levels (50% of voting age populatoin), the election will be very close. If they turn out at 2004 levels (55% of voting population), McCain will win. The sweet spot for Republicans and conservative Independents seems to be at the 53-56% turnout level, while Democrats and left-leaning Independents do better with lower turnout.

And it isn’t a case of how energized or enthusiastic GOPers are about McCain (although Sarah Palin helps tremendously to lessen this problem). The Gore voters of 2000 were also a grumpy, unenthusiastic lot, while the Bush voters were fired up and supposedly couldn’t wait to vote after eight years of Clintonism. But the Gore voters voted in droves anyway. And the Bush voters believed all the hype about Bush winning over 320 EV’s, perhaps even a landslide. Many of them ended up not bothering to vote. Many of them were the same newly registered young voters that Obama is now counting on so heavily in this election.

Even in 2006, Republicans had remarkably good turnout in what was a horrible year for them, which enabled them to lose most of their races by very narrow margins instead of in blowouts. Indeed, it is for this reason that I am hoping that if we get good turnout this year, the GOP will do much better in the House races than everyone seems to think they will (the Senate, unfortunately, may be another story).


41 posted on 10/18/2008 10:22:21 AM PDT by kesg
[ Post Reply | Private Reply | To 9 | View Replies]

To: MediaMole
Would someone who switched party to vote in a primary tell a polling operation that he was now a democrat? Operation Chaos was an attempt to help the GOP. It was not a shift of voters’ allegiance.

Turnout numbers for the primaries could cause a big shift in Rassmussen’s numbers, but self-identification shouldn’t shift that substantially.

It's not a question of whether they tell a pollster they are democrat. Pollsters ask which party you consider yourself a member of or which is closes to your beliefs.

Where the weighting comes in is this - if they've determined from voter registration that a population (like a state) is 60% dem 40% rep, and you identify yourself with the republican, your views count as .4, while if you say you're a dem you are counted as .6. Across a thousand voters, if 60% of those who answer the phone identify themselves with the republicans and say they are voting McCain, those 600 people will only count for 40% of the total. The 400 who identify themselves as democrats will count as 60% of the total because the Registration data says the self-identified democrats represent 60% of the state's registered voters.

That's not an exact explanation of weighting, but it should give you the picture. That is why the ACORN registrations matter in the polls, as well as why Operation CHAOS will have impacted the results of the polls. The polls may not be a very accurate reflection of actual voters this year because of those factors that impact the weighting.

42 posted on 10/18/2008 10:30:13 AM PDT by Kay Ludlow (Government actions ALWAYS has unintended consequences...)
[ Post Reply | Private Reply | To 28 | View Replies]

To: Bob J

Maybe my thinking is skewed, but ———

IF polsters are calling Democratic listings, and some of those answering are those who had changed over to Dem, (because of Operation Chaos), and are telling now tell the polster that they are for McCain, -——then wouldn’t that upset the Dem/Obama balance? In other words, less Dems on their lists for Obama?

Or weren’t there that many involved in OC to make a difference?


43 posted on 10/18/2008 10:34:25 AM PDT by Exit148 (Founder of the Loose Change Club. Every nickle and dime counts!!)
[ Post Reply | Private Reply | To 1 | View Replies]

To: Bob J

Before April, 2004, Rasmussen didn’t even use weighting in his polls.


44 posted on 10/18/2008 10:35:16 AM PDT by bw17
[ Post Reply | Private Reply | To 1 | View Replies]

..


45 posted on 10/18/2008 10:37:03 AM PDT by Chet 99 (Vote McCain/Palin, or this will be our future: http://www.youtube.com/watch?v=QTb5EFZmgbs)
[ Post Reply | Private Reply | To 1 | View Replies]

To: xzins

I am not sure that question controls in the way you are analyzing the polls. I may be wrong.

People will accurately — I guess— answer this question and not reflect back on their primary chaos activity. The problem as I understand is that once a pollsters has completed the poll, they need to decide who to include and not include in the sample they have taken. They have to control for unforseen biases in gathering the sample.

To do this, they impose formulas where they accept a percentage of dems, republicans and indpendents. If they accidentally polled 65% democrats, they can simply adjust their sampling to include a smaller percentage that reflects the norms they believe govern society.

If this is what is happening, and i may be wrong here, then the polls are hopelessly skewed.


46 posted on 10/18/2008 10:37:45 AM PDT by lonestar67 (Its time to withdraw from the War on Bush-- your side is hopelessly lost in a quagmire.)
[ Post Reply | Private Reply | To 35 | View Replies]

To: zebrahead
“Rasmussen and SurveyUSA have Obama up by an average of 7 points in Missouri. Only SUSA posts their internals and they weight 38% Democrats and 31% Republicans. Looking at the exit polls from the 2006 Senate race in Missouri, Republicans made up 39% and Democrats 37%. A drop in Republican support of 8% is exactly Obama’s margin overall.”

If I remember right, Operation Chaos was a hugh success in Missouri. I hope OC registration changes are large enough to skew these pols this much. I know I flipped here in Oregon and am waiting until after the election to flip back.

I feel dirty being registered as a Rat, and I am eating way too much cheese. Thank God for Joe the Plumber.

47 posted on 10/18/2008 10:38:47 AM PDT by mickey finn
[ Post Reply | Private Reply | To 33 | View Replies]

To: Bob J
“...They keep the 400 dem responses, throw out 50 pubs and 50 indys to get the 40-35-25 ratio....”

Question:

Which ONES do they throw out? The ones that voted for McCain or Obama....??? How do they make that decision

48 posted on 10/18/2008 10:39:49 AM PDT by forYourChildrenVote4Bush (Today, July 16th I no longer donate money for Israel)
[ Post Reply | Private Reply | To 36 | View Replies]

To: Exit148

I think OC had an impact of no more than 1-2%. I can’t answer your question and the only way to get any substance out of it is to watch the “crossovers” in those polls. If they heavily favor dem-—>McCain, I think we are seeing the effect of OC.


49 posted on 10/18/2008 10:40:30 AM PDT by Bob J (For every 1000 hacking at the branches of evil, one strikes at it's root.)
[ Post Reply | Private Reply | To 43 | View Replies]

To: Bob J

It would seem to me that weighting is counterproductive — especially when you consider the fact that in the last several presidential elections, the polls have not predicted the winners.

So, it would also seem to me that whoever is ahead in the polls turn out to be the winners.

The goal, then, would be to be lagging in the polls. Those are where the winners are found.


50 posted on 10/18/2008 10:43:35 AM PDT by quintr
[ Post Reply | Private Reply | To 1 | View Replies]

To: bw17

Really? Is there any way to find out his results from elections prior to that? In other words, was polling without weighting as accurate?

BTW - I’m having a hard time accepting that there was no weighting prior to ‘04...


51 posted on 10/18/2008 10:47:00 AM PDT by Bob J (For every 1000 hacking at the branches of evil, one strikes at it's root.)
[ Post Reply | Private Reply | To 44 | View Replies]

To: quintr

“...whoever is ahead in the polls turn out to be the winners”

should read: whoever is behind in the polls turn out to be the winners...”

sorry about that.


52 posted on 10/18/2008 10:47:16 AM PDT by quintr
[ Post Reply | Private Reply | To 50 | View Replies]

To: xzins
If someone calls you on the phone and asks if you consider yourself to be a dem, rep, or indy, what do you say?

I say *click*! I can't hang up on unsolicited callers fast enough, especially pollsters or politicians.

53 posted on 10/18/2008 10:47:58 AM PDT by 6SJ7 (Welcome PUMAs!)
[ Post Reply | Private Reply | To 35 | View Replies]

To: lonestar67

If I randomly call a thousand people in the US, and just accidentally get all my phone numbers with California area codes, will that affect the opinions I get? I think so. The same as if some weird group of numbers came up all out of Montana. I need to make sure my sample is balanced in a lot of ways, but there seems to be this emphasis with Rasmussen on making sure it has 39% democrats and 33% republicans. He says he gets that from phone surveys he takes in which he simply asks folks what they consider themselves to be.

Some don’t try to balance for political party at all.

The bottom line, though, is that the national poll means nothing. It’s a state-by-state election.


54 posted on 10/18/2008 10:48:00 AM PDT by xzins (Retired Army Chaplain Pro Deo et Patria)
[ Post Reply | Private Reply | To 46 | View Replies]

To: Bob J

I’d like to add another piece of anecdotal evidence about polls. A friend of my daughter participated in a poll by phone. Questions were asked about issues: immigration, taxes, etc. At the end the pollster wrapped up and said thank you. The poll-taker said, “But wait. Aren’t you going to ask who I am supporting in the election?” The pollster said, “Oh we don’t need to. From your answers, we know you are voting for Obama, so I already put your vote down for him.” The poll-taker responded that there was no way in ___ that she would ever vote for Obama, but that didn’t matter to the pollster.


55 posted on 10/18/2008 10:48:14 AM PDT by Faith
[ Post Reply | Private Reply | To 1 | View Replies]

To: forYourChildrenVote4Bush

EXACTLY. I’m going to assume they are thrown out randomly but if there is an ability for one of their “employees” to do it selectively, that could be a huge potential for fraud.


56 posted on 10/18/2008 10:49:18 AM PDT by Bob J (For every 1000 hacking at the branches of evil, one strikes at it's root.)
[ Post Reply | Private Reply | To 48 | View Replies]

To: 6SJ7

And someone should keep track of the refusals.

Call 2000 homes and report the real results.


57 posted on 10/18/2008 10:50:52 AM PDT by xzins (Retired Army Chaplain Pro Deo et Patria)
[ Post Reply | Private Reply | To 53 | View Replies]

To: quintr

I believe that by election day many of the polls are withing the margin of error. That the MOE always seem to favor dems is a question that has never been answered... ;-}.


58 posted on 10/18/2008 10:51:20 AM PDT by Bob J (For every 1000 hacking at the branches of evil, one strikes at it's root.)
[ Post Reply | Private Reply | To 50 | View Replies]

To: kesg

The Democratic registration and turnout machine is the best of the modern era. With Democratic registration up sharply and likely to show up and vote, a stable GOP turnout rate means that we fall behind in actual votes produced. We are going to take some hard licks even if we manage to avoid a disaster.


59 posted on 10/18/2008 10:52:36 AM PDT by Rockingham
[ Post Reply | Private Reply | To 41 | View Replies]

To: Bob J

“It’s not who votes that counts. It’s who counts the votes” Stalin.


60 posted on 10/18/2008 10:52:58 AM PDT by RetSignman (DEMSM: "If you tell a big enough lie, frequently enough, it becomes the truth")
[ Post Reply | Private Reply | To 51 | View Replies]


Navigation: use the links below to view more comments.
first previous 1-2021-4041-6061-80 ... 101-104 next last

Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.

Free Republic
Browse · Search
News/Activism
Topics · Post Article

FreeRepublic, LLC, PO BOX 9771, FRESNO, CA 93794
FreeRepublic.com is powered by software copyright 2000-2008 John Robinson