The coming Obama landslide

Landslide? No. I think it will be 51-49, or 50.5-49.5, but which way it will go is still a toss-up. If you factor in the juicing of the polls, it’s an even split. It’s going to come down to who gets their people out to the polls, and whether blacks and college kids make up a disproportionately large share of the electorate the way they did in 2008.

It might even be like 2000, when the margin of victory in the crucial state(s) is less than the margin of error. Expect cries of “disenfranchisement” and “voter intimidation,” especially if Romney pulls it off. For me personally, the saddest day of the past two years was when Mitch Daniels pulled out of the race.

Maybe it’s because the Republican Party is run by a bunch of incompetents. Which is saying something when they’re up against the Democrats. When they should have gone bold, they went safe. While this will probably never happen – when Romney does lose, and loses horribly if this article is correct – the Republican Party leadership needs to resign en masse and never, ever again pretend that they know what is best. The last election may have been a toss up, but this one has been gift-wrapped for several years and they have fumbled it away by offering up to the American people an artificial choice. The most monstrously damning thing, though, is that the Republican Party faithful, who now say, “I never really liked Romney. He’s not a very good candidate,” swallowed the whole thing hook, line, and sinker. The Republican Party is political example par excellence that “there’s a sucker born every minute.”

The problem the Republicans are facing is the same problems the Democrats faced in 2004 — they have a bland, not very likable candidate who doesn’t capture anyone’s imagination. You cannot run a great campaign on “I’m not the other guy.” The Democrats should have driven Bush from the White House in 2004… but they didn’t give people anyone to vote *for*. Why would one for *for* Romney, other than the fact that he’s not President Obama?

Pollster John McLaughlin told reporter Jim Geraghty:

What Obama and his allies are doing now: “The Democrats want to convince [these anti-Obama voters] falsely that Romney will lose to discourage them from voting. So they lobby the pollsters to weight their surveys to emulate the 2008 Democrat-heavy models. They are lobbying them now to affect early voting. IVR [Interactive Voice Response] polls are heavily weighted. You can weight to whatever result you want. Some polls have included sizable segments of voters who say they are ‘not enthusiastic’ to vote or non-voters to dilute Republicans. Major pollsters have samples with Republican affiliation in the 20 to 30 percent range, at such low levels not seen since the 1960s in states like Virginia, Florida, North Carolina and which then place Obama ahead. The intended effect is to suppress Republican turnout through media polling bias. We’ll see a lot more of this.

Moral: don’t trust a good looking, polite rich guy from Massachusetts to unseat the incumbent.

In reality, the GOP wrote off the few inspiring candidates they could field from day one.

Dr. Veith, why you would believe that polls suggesting a Democratic turnout GREATER than the wave year of 2008 and Republican turnout LESS than the wave year of 2008 offer anything remotely realistic is beyond me. That’s just insane. But if you feel it necessary to believe something in those polls, look at how the Independents are going — decisively for Gov. Romney. While the President may certainly be ahead (likely 1-2 points), I find it incongruous that he should lose Independents and still win the election.

http://www.nationalreview.com/campaign-spot/328555/mornings-polls-project-more-heavily-democratic-electorates-2008

It is more likely that the poll results are the frantic efforts of a leftist media who is desperate to drag the President over the finish line by attempting to convince Republicans and Independents that the election is a foregone conclusion and that you should just stay home.

If you “hear” a little voice urging you to ignore your vocation of citizenship and skip voting, it is NOT God’s voice you hear.

16 Trillion Debt — 8.3% Unemployment — $4 Gas.

@SKPeterson
“Maybe it’s because the Republican Party is run by a bunch of incompetents. ”

Isn’t it a bit arrogant to suggest that you could do a better job than the Republican party leadership? They are an elite group of lawyers and professionals who have been in politics for much of their professional lives. But you know more about winning elections than they do?

Another thing to be aware of is what White House reporter, Jake Tapper, said about the main stream media helping Obama win over Hillary and then McCain in the 2008. During this 2012 election, they have taken off the mask of even attempting to appear objective and a majority of them have become nothing less than an arm of the Obama campaign. They aren’t even reporting or following up on Obama’s whoppers (eg: 90% of the debt isn’t my doing) and dereliction of his job as president during the fiasco of blaming a video for the Islamic terrorism and riots. The MSM has greatly handicapped Romney and the Republican party.

“… look at how the Independents are going — decisively for Gov. Romney”.

This assumes that all independents are going to pick either Romney or Obama.

A more reasoned statement would be “Of the independents who felt forced to choose between the two major parties, a higher percentage went for Romney. However, the results were well within the sampling error of the polls, meaning an effective tie”.

Just so you are aware of my leanings, I speak as an independent twenty-something who doesn’t see a way I would possibly vote for Romney.

Better a lame-duck Obama (second-term presidents are always plagued by scandals) than a liberal Republican who will mainstream the creepy religious cult that is Mormonism.

As for polls, the only one that counts is the one taken on November 6th.

There is a serious dearth of truthful information in the main stream media. As Elizabeth Scalia notes in her First Things column: http://www.firstthings.com/onthesquare/2012/09/us-press-plays-pravda

US PRESS PLAYS PRAVDA: “I’ve taken to wondering if perhaps Obama does not really want to be re-elected, and the press is thwarting him at every turn; perhaps an Obama victory is merely incidental to the press, as long as their narrative is sustained,”

And it really is all about the progressive narrative.

I think most voters aren’t even paying attention to the campaigns yet.

I think three things explain Obama’s position in the polls. (For other reasons, check out http://edition.cnn.com/2012/09/26/politics/romney-ohio/index.html?hpt=po_c1 )

1. The economy is in rough shape, but voters don’t blame the President for this unhappy reality. They recognize things were falling apart at the end of Bush’s second term. They also believe this President has done some positive things so what we’ve been going through hasn’t been as bad as it would have been otherwise. And the economy IS slowly improving. But overall, and maybe most importantly, voters don’t really believe EITHER candidate has the power to create new jobs.

2. Romney’s campaign has been bankrolled with more money, and that fact has influenced the approach taken to running the campaign. Over all, Romney and the Republicans have outspent the Democrats and the Obama campaign on mass media. Romney is relying on print and television ads to get across his message. (It’s true the President’s campaign is outspending the Governor’s in a few select markets, but this underscores how Democrats are managing their war chest more carefully; in most markets they are not even trying to keep up with Republicans.) Obama’s campaign is relying more (I think) on ground organization. Playing into this decision is the changing demographics of the U.S. As the percentage of the electorate increases that is non-Caucasian, I think ground organization that includes neighborhood get-out-the-vote efforts has greater effectiveness.

3. American voters identify the Republican party with a conservative social values agenda, which, despite the sentiments aired on this blog, most American voters are not really keen on. Not only is abortion a losing issue, so is DOMA. I imagine when all’s said and done, even Christians in America want to live in a country more socially left-leaning than what our Church leaders want. (Horrors!)

This means in 2016, if frustrated right-wingers manage to nominate a MUCH more socially and fiscally conservative candidate, that candidate will lose by an even wider margin than Romney probably will.

because they made the same mistake as 2008 an putting in a non conservative pro choice canidate. That why in 2008 and again this year I will be voting third party.

Rominy is not prolife his political past shows that.

Wow, a lot of shark jumping going on here. I still think Romney will win, but I think it’s pretty much a dead heat.

And yes, most of the polls you are seeing right now are horribly biased. They are classic push polls. Watch the Rasmussen daily tracking poll and the Gallup poll and pretty much ignore everything else.

At the state level, watch the partisan turnout breakdown and weight accordingly. Most of all, chill out a bit. I don’t have a lot of trust in Romney, but Obama is overrated as a debater (and a speaker and pretty much everything else).

” they have a bland, not very likable candidate who doesn’t capture anyone’s imagination.”

Ugh, the last thing we need is some exciting guy who captures people’s imagination with his baseless promises. Gag. That is the very definition of the guy you don’t want your daughter to marry, and also the one she chases. Why should we elect one as president? If male property owners 21 and older were the only voters, Ron Paul (or a clone of him) would be sitting in the Oval Office, the budget would be balanced or surplus, and we wouldn’t stupidly being trying to save the world’s fools from themselves in wars with pipe dream objectives.

From #7:

Isn’t it a bit arrogant to suggest that you could do a better job than the Republican party leadership? They are an elite group of lawyers and professionals who have been in politics for much of their professional lives. But you know more about winning elections than they do?

The originally poster probably does. Neither party is terribly good at winning elections — for instance, the Democrats got terribly lucky with Clinton (remember, he was not the one the party wanted; they wanted a faithful party member rather than someone with charisma they could not control) and had Obama shoved at them (going with him when it was clear that the voters wanted nothing to do with the party favorites). Neither party tries to actually win; both aim at getting as many on their hemisphere of the political globe (we must get rid of this spectrum nonsense, unless we admit that it wraps around) to vote for them. They aim to please donors and the party base, rather than try to find a candidate who might have appeal for the whole electorate. Thus the constant demand for at least lip service to party orthodoxy on the part of candidates who have no real interest in it, and the retreat on both sides to more and more radical positions.

I don’t have a lot of trust in Romney, but Obama is overrated as a debater (and a speaker and pretty much everything else).

Well he isn’t underrated in his appeal to the darling demographic of the marketing world, single women. He gets the largest share of the largest block of the least informed voters. What’s not to love if you are a marketer?

•Spring 1980 Carter ahead of reagan 43 to 35 gallup
•October 20 1980 Carter beating reagan 43-41
•Nov. 4 election Carter 41 Reagan 51

I remember this and have not trusted polls since.

Yes, that’s right, all you Republicans and all you would-be “anyone-but-Obama-voters.”

Obama just finished mailing out the last of his Inauguration Ball invitations.

There’s nothing more to do. So Just Stay Home on Nov6! Save that postage stamp! Don’t even bother to vote!

(This all smells very fishy for polls to start pulling away when history shows that no president with approval ratings and a record like the current president has ever won. Some devious collusion is going on between Team Obama, the pollsters and MSM, methinks.)

Seems to me I kept hearing about Carter and Mondale being ahead in the polls all the time as well…

Karl Rove explains

http://thehopeforamerica.com/play.php?id=13141

Random Lutheran at 17….got a chuckle out of what you wrote:

“Neither party is terribly good at winning elections”

Because at least one of those parties have won virtually 100% of most US elections 🙂

cheers!

@23: LOL. Seriously though, that means that’s it’s been a coin flip for some time…

@Other Gary

I would disagree with your assessment.

Obama may have been successful in misleading many people into blaming Bush for the consequences of his actions during his term in office, but that doesn’t make it true. The economy is not getting better. This last year, after 3 years of his presidency, the US credit rating has been downgraded. Twice. Since taking office, he has increased the US debt by $6 trillion dollars. His budget proposals have been so bad that %100 of congress rejected his budgets. That’s unanimous bipartisanship. While unemployment has been at over 8% for the last 44 months, studies that include those who gave up looking for work or who are working involuntarily as part time workers, put the real number around 17%. Household incomes have declined 8.2%. This week’s news is: 2Q GDP revised down to 1.25% from 1.7%… and: Durable orders drop 13.2% — ‘worst since Depression’… It should be well known that the claims that the economy is getting better are false and Bernanke has been busy trying to manipulate the economy to no avail. Hopefully, his latest moonbattery of buying $40 billion per month of mortgage securities and driving up the stock market prices in hopes that people will fall for the artificial inflation of stock prices (eg: sell their capital and spend it on consumer goods) will not end in complete disaster.

While it is true that government does not create jobs – especially by increasing government jobs or subsidizing industries (eg: wind farms) which by definition do not add to GDP. Government policies and actions do affect the economic climate with stability or uncertainty. ObamaCare is but one example of creating instability and uncertainty. Businesses have no idea how much it will cost or how to make long range risk planning for it.

I still have my money on Romney given historico-empirical trends regarding the fate of incumbents during periods of economic stagnation.

That said, Romney is a terrible candidate, and he has run a campaign bordering on utterly incompetent. Thus, I won’t be shocked if Obama wins. I think the popular vote will be close in either case, though. Predicting the outcome of the election based on today’s polls is premature, at best. As ChrisB suggests @12, it’s unlikely that most voters are even cognizant that an election is happening soon (cf. last Saturday’s SNL skit regarding the “undecided voter”).

One note, though: leftists in particular have gleefully misinterpreted the likely effect of Romney’s “47%” gaffe. Contrary to elite opinion, I guarantee that those remarks appealed to the conservative base–probably the only thing he’s ever said that could energize his base. I personally know lots of working class folks who responded to those comments with the equivalent of an “Amen, brother!”

Also, anyone who thinks Obama will beat Romney in a debate is smoking something strong. Obama is a terrible public speaker, and the fawning adulators who have suggested otherwise over the past five years are smoking something even stronger.

That said, sg’s comment deserves repeating: the world/America would be a better place if the franchise had never been extended beyond property-holders.

/heresy!

Why would you vote for Romney with all his warmongering talk, secrecy about his policies, I am amazed that Christians with all their preaching about getting a Christian into the white house would even consider Romney, a devout Mormon who believes that Jesus came to the United States from Palestine. Mormons are far more generous than mainstream followers of Christ, it surprises me that he chose Ryan since his church absolutely is totally against what Ryan preaches. It seems to me to all add up one thing, MR is determined to bepresident and will fudge, dissemble, and hide to do this. Why would we want someone like that? He seems to be rather megalomanic; which doesnt mean he is not a kind person.

Hanni@28:

Why would you vote for Obama with all his actual warmongering?!

Your assignment is to read this: http://www.theatlantic.com/international/archive/2012/09/every-person-is-afraid-of-the-drones-the-strikes-effect-on-life-in-pakistan/262814/

Michael @ 7 – Is it arrogant? Yes. I admit it. So what? I think they’re idiots and that I could do better. I think a large swathe of the rank-and-file supporters of the Republican Party would do better. They’ve been doing this their whole professional lives and this, this!, is what we get? I didn’t have to be a professional automotive designer or engineer to discern that the Pontiac Aztek was a piece of crap, despite the glowing gush of positive commentary from the GM leadership, and that I could have designed something less goofy. Most four year-olds would have abandoned the Aztek as childish frippery, as well. Eventually that level of incompetence landed GM in the loving arms of the federal government.

If the “professionals” continue to botch things on this level, do we really need to continue to rely on the professionals? Maybe I’ll be proven wrong. After all I’m not someone who’s been up to my gills in politics for the last 25 years. But right now, I’m not seeing how I could be doing worse than what the Republicans are putting forth.

“Why would you vote for Obama with all his actual warmongering?!”

Because impressions and marketing hype are more persuasive to the weak-minded than facts.

He seems to be rather megalomanic; which doesnt mean he is not a kind person.

Seems?

Can you explain why he seems so?

@Hanni,

Ditto with Cinannatus’ question and reading assignment. I would ask you what policies is Obama proposing? And do you think Romney/Ryan are running for public office or church office? Obama is the one who is enforcing policies that attack the freedom of exercise of religion and wants to enforce abortion without any limitations and make gay marriage the law of the land.

Cin at 27…..perhaps the analogy to a public corporation is a better model for how to distribute the franchise?

Only stockholders can vote (if you want to vote for who leads the corp, you must own stock in the corp), and it get’s even better…you get not just one vote/stockholder, but you get one vote/each share!

If the “business” of america is business, perhaps we should run it like one 🙂

cheers!

The fact that Romney is a member of a church that is somewhat unpopular means that he would be more disposed to favor religious liberty.

@34 I think you are thinking of feudalism not the US Constitution as originally written and approved by states who had a different share of the votes quite literally.

Here is one of the “47%” explaining why she’s voting for Obama and what is wrong with Romney.

Is Obama going to use his teleprompter during the debates?

Oh Carl @ 37! Those crazy coloreds are always good for a laugh, eh? But it’s time we level-headed white folks got back to the serious discussion, here, of media and polling conspiracies.

@Carl Vehse

Another poster child for those who believe in the government fairy and think he will magically pay for their mortgage, gas, food, phones, college education, health care, ad infinitum. I’ve got some beach front property in Arizona they might be interested in…

@Jon- Kudos. ; )

Race cards aside, Tom@39, is there anything this Obamaniac from Ohio (one of the critical electoral states) said that indicates the pollsters are overestimating Barry’s support?

And just look at the enthusiasm of these Obama voters in Chicago!

It’s a hoot to see all you righties reinventing reality as the election
goes to President Obama.

Lots of bad news for you.

The latest: Catholic voters are breaking for Obama, 54-39.

Most Americans weren’t fooled by the Catholic extremism of Paulie Ryan and the Right’s obsession with religious freedom.

Tough break. Awwwww.

http://www.huffingtonpost.com/2012/09/27/obama-surges-catholic-voters_n_1918120.html

Here’s something to consider:

Taxpayers spent $1.4 billion dollars on President Obama and his family last year. In comparison, British taxpayers spent just $57.8 million on the royal family.

http://dailycaller.com/2012/09/26/taxpayers-spent-1-4-billion-on-obama-family-last-year-perks-questioned-in-new-book/#ixzz27gscJLqp

Anon@42:

Your pointless gloating aside, Catholics have been a reliable Democratic voting bloc since their Italian, Irish, and Polish ancestors got off the boat.

Some folks were suggesting that would change this year, given Obama’s assault on the Church via the HHS mandate, but I doubt it.

I don’t know who will win the election, but it does look dim for Romney. The reason:

Those who KNOW what the Mormons believe, and are knowledgeable in the Word of God are not going to vote for Romney, they will sit the vote out for president and vote for the rest of the prop’s and others running in their state, etc.

For some, the moral courage to face the future, knowing that it won’t be easy if Obama wins, but knowing full well that voting for a man who believes he will one day be a god, is mocking the real LORD God ALMIGHTY.

This worlds comforts are not worth, SELLING OUT!

Obama ahead among those who self identify as Roman catholic or among those who attend weekly mass?

Spin, spin, spin.

It is easy to say you are Roman catholic. Getting up and going, well that is different.

Among Catholics who attend Mass at least weekly, Romney holds a 51%-42% lead.

Basically, the closer they are to the church, the farther they are from Obama.

http://www.catholicculture.org/news/headlines/index.cfm?storyid=15711

President Barack Obama holds a 54%-39% advantage over Mitt Romney among Catholics, according to the latest voter survey conducted by the Pew Research Center for the People & the Press.

Among Catholics who attend Mass at least weekly, Romney holds a 51%-42% lead. Catholics who attend Mass “monthly” or “yearly” favor Obama by a 53%-39% advantage, while Catholics who attend Mass “seldom” or “never” back Obama by a 61%-32% margin.

The survey also found that white Catholics favor Obama by a 47%-46% margin.

Protestants favor Romney by a 50%-42% margin; among white evangelical Protestants, the pro-Romney advantage is 74%-19%, while the two candidates are in a 46%-46% dead heat among mainline Protestants. Black Protestants favor Obama by an overwhelming 95%-2% margin.

The Pew survey, conducted September 16, found that Obama holds a 51%-42% lead among all registered voters—a significantly larger lead than the 3.7% advantage Obama currently holds in the RealClearPolitics.com average of surveys.

For all the talk of racism against Obama, I really don’t see reports of voters who agree with Obama’s policies but won’t vote for him because he is half African. On the other hand, if Romney were, say, Methodist, he would probably get a few more votes from people who largely agree with him and tend to disagree with Obama. So, if anyone is suffering from prejudice, it is probably Romney.

@44 Cincy, don’t forget (Catholic) Latinos. To the fastest growing ethnic group in the US (most of whom are not immigrants), the GOP can only say ‘self deport.’ Yet, after Romney loses, the party’s cry will be, “We didn’t move far enough to the right!”

Grace @46: “… but knowing full well that voting for a man who believes he will one day be a god, is mocking the real LORD God ALMIGHTY.”

So is voting for a lying, murdering traitor who, along with his entitlement morlocks, thinks (and acts like) he is god!

Grace @45,

The polls (for what they’re worth) don’t support that conclusion. Evangelicals, who would in theory be most likely to understand Mormonism and not like the Gov Romney on the basis of his false faith, are flocking to him (something like a 45% lead!). The President is behind in every poll (again, FWIW) of those who “attend church regularly.” As a group Evangelicals are quite animated about the HHS regulations because they fear what else may come down the pike. They won’t sit this out — even though they think Gov. Romney is going to hell.

Yes, to orthodox Christians, Mormonism is a heretical cult and its followers stand condemned for their unbelief. We may question their judgment for believing such nonsense. But we are not electing someone to be the Chairman of our Board of Elders. Since my vote on Nov 6 has nothing to do with the faith (or lack thereof) of the persons running for president, the only question that matters is which individual I believe will better lead the country. The presidency has nothing to do with the practice of our faith — except when it seeks to overstep its authority and impose something (HHS regulations) that violate our conscience. We must not conflate the Two Kingdoms. Yes, God rules both – but that doesn’t mean He doesn’t use the heathen to do His work (see Dr. Veith’s excellent work on Vocation). 🙂

Yes, there should be concern about “normalizing” Mormonism. But frankly, the Senate Majority Leader is a Mormon and that fact has not suddenly created a groundswell of agreement among Christians that Mormonism is no longer a cult. As long as churches continue to preach Law and Gospel, the truth will be proclaimed.

Cincinnatus’ point @44 is spot on — though I am hopful the Roman Catholics will forego their default position to support the social gospel via government action and take a good look at how the HHS regulations will severely impact the practice of their faith. Especially as the election draws closer and voters start paying more attention.

Jon H.:

I don’t approve of the Republican immigration platform, but since when has that platform involved telling non-immigrant Latinos to “self-deport”? Since never, I think? Heck, the W. administration was famously–and well-nigh traitorously–laissez-faire about illegal immigration.

Anyway, Latinos who are actually eligible to vote lean 60% /40% in favor of Democrats. But I don’t think that’s because Republicans are telling them to self-deport, because Republicans aren’t actually doing that. I am more than certain that the disparity is due to the fact that many Latinos are poor, dependent upon the same social programs that inspire quite a lot of other poor voters to continue choosing Democrats.

@51
Not hopful. Still at work (though it is lunch). Sadly, no beer here. 🙁

However I am hopeful…

Why is it that Republicans are always asking me to believe in vast conspiracy theories?

Most traditionally, they are always asking me to believe in a monolithically liberal media, one that is nothing but an arm of the Democratic party, its only goal the electing of Democrats, truth be damned.

Now I’m being asked to believe the same thing about pollsters. Another vast conspiracy, with nary a fact to be found in the data. Just lies for the sake of electing Democrats.

And, wouldn’t you know it, if it weren’t for all those vast conspiracies out there, Republicans would be doing better. They’d be winning more elections, solving more of the world’s problems, and there would be more rainbows in the sky.

I’m not denying that bias exists here or there. It does — though, as is almost universally the case with supposed conspiracies, “bias” is usually better understood as incompetence, of which there is no shortage.

But you know what? Republicans so often just come across as whiny. They’re always whining about the system. “Boo hoo! It’s gamed against us! If only things were fair, we’d win all the time!”

You know what? If Republicans weren’t so good at botching the game, maybe the supposed influence of biased polls wouldn’t be much of an issue, would it? If Republicans could nominate a candidate that even 20% of their party base were excited about, maybe the supposedly singular influence of the “MSM” wouldn’t matter so much.

In short, maybe you’re resorting to these conspiracy theories because you’re trying to form a cogent argument in favor of a terrible candidate.

Man up, Republicans. Learn how to play the game.

tODD:

Isn’t Carl the only one claiming conspiracy? Which is like a meteorologist claiming that it’s going to rain in Portland next week, so you haven’t really made a worthwhile observation.

Cincinnatus (@55), what?!

By my count, here are the people apparently alleging that there is some kind of conspiracy in the polling this year:

* Dick Morris (in Veith’s update)
* Matthew (@1): “the juicing of the polls”
* Susan (@4): “So [Democrats] lobby the pollsters to weight their surveys to emulate the 2008 Democrat-heavy models. They are lobbying them now to affect early voting.”
* NavyChaps (@6): “It is more likely that the poll results are the frantic efforts of a leftist media who is desperate to drag the President over the finish line…”
* Steve Billingsley (@15): “most of the polls you are seeing right now are horribly biased. They are classic push polls.”
* Jon (@20): “This all smells very fishy for polls to start pulling away. … Some devious collusion is going on between Team Obama, the pollsters and MSM, methinks.”
* HippoAugustine (@22), via Karl Rove

@55, But see @15.
Also, DonS will chime in soon … waiting … waiting.

Cincinnatus (@55), you also said:

Which is like a meteorologist claiming that it’s going to rain in Portland next week

A humorously unfortunate comparison, as we’re still in our dry season. We’ve had less than a quarter of an inch, total, since July. Most of next week looks to be sunny and in the 70s, as this week has been. So, yeah.

tODD@56:

Hm, guess I should read more carefully.

On the other hand, I’m actually somewhat sympathetic to claims of conspiracy. These media polls–as opposed to some of the more neutral but still worthless polls–are often weighted arbitrarily in favor of Democrats by 8 or 9 points (!)*, and regularly oversample from Democrats. “Conspiracy” probably isn’t the right word, but “hype” certainly is.

*There are various reasons for this. One of them is that media elites still maintain the romantic delusion that many likely Democratic voters are hardworking ‘Murkans working 7 jobs but still can’t afford a phone. They imagine, for no particular reason, that this class of non-existent voters represents up to 10% of the population of likely voters.

IT WAS A CONSPIRACY!!!! Do you not know that the NFL referees strike has been a coordinated attempt to distract the populace! Look it up! See – right there in black and white – the mainstream media lapdogs have distracted and hoodwinked the people for Obama. Hah! Now that it is over, you fools will all suffer the WRATH of Romney and his plethora of Mormon wives who will occupy the White House! Soon you will be driven into the SEA!

Also–and not to jump too eagerly on the conspiracy train–but anyone who thinks flagship media outlets haven’t been heavily biased against Romney during this election cycle are probably willfully blind.

A friend of mine has been keeping track of the campaign coverage provided by the Washington Post. In the past few weeks–even during the whole Middle East debacle, which should really have tarnished Obama’s image in the foreign policy department–the Post has featured no fewer than 5-7+ stories on Romney on the front page, all of them negative. Usually no stories on Obama, and certainly none of them negative. I found this interesting.

Still don’t care about Romney’s fate, though.

SKP, you forgot to add:

WHA HA HA!!

You know a party is in trouble when the main response from its supporters as to its dismal performance contain the word “Conspiracy”.

KK@63:

Hold up. I’m by no means claiming that the Republicans are “performing” poorly on account of a conspiracy. The Republicans are performing poorly–to the extent that they are performing poorly–because the party leadership is a collection of incompetent, elitist fools.

The polls, though, are skewed, often deliberately–and usually in favor of Democrats. Every time Veith–or anyone–posts polls like these, I point out this fact. But people still use them as grounding for their respective hysterics.

Cin @ 64 – You, sir, are an arrogant bastard for claiming the Republican Party leadership is made up of elitist incompetents. I ask you, sir, could you nominate someone better than Romney? I daresay not! Ignore the conspiracy if you will, but do not tarnish the hard-won reputations of your betters in your blind ignorance.

tODD @ 54

Gee, I remember being told by the now Secretary of State that it was the Vast Right Wing Conspiracy (R) that was out there controlling everything making life hard on the pure and righteous Democrats.

Seems to be plenty of whine on the Dem side.

But to your complaint: Since 94-97% of the media votes Democrat, mightn’t that be an indication that there is a problem in the way the media reports? You are right that incompetence is partially to blame, but it is more than that. They intentionally create “narrative” that is pro-President Obama and anti-Governor Romney. They WANT the President to be re-elected and so they tank the news that might be harmful to him.

Incompetence explains buying into the absurd idea that the attacks in Benghazi were purely the result of a protest gone out of control. Being in the tank explains the pathetic lack curiosity as to why MORTARS and combined arms tactics were so effectively used (andy by whom). And to the polls under discussion: Is there ANY evidence that their polling model has any basis in reality? The oversampling is insane. Incompetence explains the entering assumption that the Dems will turnout 2-3% above the wave of 2008. Being in the tank explains the doubling down when called on their incompetence instead of correcting for their error.

It’s not a conspiracy. But they are in the tank and they adjust their reporting to meet THEIR narrative that has the President winning in a landslide. Reality is an just inconvenient truth that gets in the way of their dream, so they ignore it until they have to spin it.

I will posit this: there are three points of view being expressed here.

1) The pro-Obama gloat pov
2) The Romney Is a Victim of Media Conspiracy pov
3) The Republican Party Leadership (and Romney) are incompetent pov

Now, other than 1), there is a range of adherence to either 2), 3) or some mix between. Now, I hold to 3) almost exclusively, and here’s my argument: 2) is more or less a given, but if you can have the inverse of 3), i.e. campaign/leadership competence, then 2) can be easily overcome. For examples, see Reagan and Bush II in recent memory. For examples of how 3) tends to play out, see Bush I part 2, Dole and McCain.

Carl @50 “So is voting for a lying, murdering traitor who, along with his entitlement morlocks, thinks (and acts like) he is god!”

Voting for either one of them isn’t a choice for most Bible Believing Christians –

NavyChaps @ 51

Sitting out the vote!

Abstain from all appearance of evil.
1 Thessalonians 5:22

Cincinnatus (@59), it appears you’ve forgotten the context in which this discussion is taking place:

These media polls–as opposed to some of the more neutral but still worthless polls–are often weighted arbitrarily in favor of Democrats by 8 or 9 points

Okay, first thing: we’re referring to the RealClearPolitics aggregate tracking poll. Which, yes, includes some “media polls” (including the Wall St. Journal, which generally isn’t held up as an icon of liberalism, but it runs a joint poll with NBC, FWIW), but also several not — notably, Rasmussen and Gallup.

Moreover, RealClearPolitics, if it has a bias itself, leans slightly conservative, in my experience. Which isn’t much of a bias, is it? But it’s hardly a member of the “MSM” agit-prop conspiracy.

So it’s one thing to argue that this or that poll, run by a media outfit that is at least plausibly left-leaning, is wrong. But when an arguably right-leaning site, collecting data from several pollsters, at least some of which are respected and historically vetted, comes to the conclusion that Obama is ahead in a statistically significant fashion, I really have a hard time believing it’s all due to conspiracy.

Grace @69

So are you saying that Daniel’s willing and faithful service to pagan kings was a sin? In a sense, he “voted” to serve as an advisor (as opposed to rejecting the offer).

What about Esther? Did her marriage to a pagan king violate your principle?

Are Christians thus forbidden to work with and/or support non-Chrisitians as they fulfill their vocations in the world (since God is also working through them). What about the work of various organizations or corporations — does that requirement prevent any association with them as well?

Again, Two Kingdoms.

As just one example (if you want to gainsay me, do your own legwork), Gallup (again, a constituent member of the RCP aggregate poll) right now has Obama ahead by 6 points, with a margin of error of 2 points.

I think Gallup has done pretty well, historically, so they’re hard to discount as part of a conspiracy. They certainly were within their margin of error for the past two elections.

Of course, at some point, one has to realize the failing of national polls like this. At best, they tell us the result of a mythical popular-vote election. Which isn’t the election we’re going to have.

As a case in point, Gallup correctly predicted in 2000 that Bush would win, but in so doing, they kind of miffed the actual results of the national totals, since, as you’ll recall, Gore won that. They missed Gore’s total by 2.4 percentage points. So should that count as their getting the election right, or wrong?

Interesting side note: Rasmussen has Obama and Romney tied right now, with a margin of error of 3 points. And I think they have a good track record, historically. Still, it’s worth noting that RCP’s average right now has Obama at 48.9 points and Romney at 44.9. Note that this fits within both Rasmussen’s and Gallup’s margins of error.

A further side note: conservative people will often complain (this year, at least) about pollsters using the wrong survey screen. Well, from what I can see, RCP’s constitent pollsters all use “likely voters” as their screen, except for Gallup, who uses “registered voters”. So I don’t see that as an issue here.

There is a new website named Unskewed Polls: http://www.unskewedpolls.com

It consistently shows Romney ahead and high disapproval ratings for Obama. Natch, it’s not perfect, but it’s something to consider when talking about polls.

NavyChaps (@66), as for your opening paragraphs, it would have been more efficient for you to simply write “Tu quoque“, don’t you think?

Since 94-97% of the media votes Democrat…

Oh good grief, do you really think you can pull a statistic like that out of your … head? I’m gonna call on that one. Citation, please. Because even the rather right-wing Media Resource Center doesn’t allege numbers anywhere like that … at least, since 1964. According to them, in a 2004 survey of 300 journalists (most recent I could find on their site), 156 voted for Kerry, and 57 for Bush. Lopsided? Yes. 94%? Um, hardly. And, again, this is the data the conservative media critic site chooses to publish. I’m gonna go with: your claim is full of it.

But that’s only part of the problem. What you’re asking me to believe is that people who vote Democratic are incapable of being objective. Were I to apply that same logic to you, I could simply dismiss your claims out of hand because, hey, you consistently vote Republican/right-wing/conservative (a guess, but a likely one, isn’t it?). Instead, I’m actually replying to your arguments.

So let’s see, that’s two logical fallacies so far: tu quoque and ad hominem. What else you got?

tODD@72:

I’m quite aware of the source, though I didn’t know RCP was noticeably conservative. I thought it was more or less an aggregator–which comes with its own can of potential biases, but not editorial in nature.

Anyway, the polls we’re looking at–Quinnipiac, Franking and Marshall, Gallup, Rasmussen, etc.–often skew anywhere from 2 to 11 points in favor of Democrats. The ones on RCP right now are NBC, AP, etc., so I reckon problems with skew are even worse. Rasmussen is probably the least skewed in general–and, surprise, that’s the one that shows them in a dead heat! As I said, this is less conspiratorial in nature than it is hype.

Ah, yes, “Unskewed Polls”. A site where a man takes other people’s polls and reweights them according to a single set of data points on voter registration. What could possibly go wrong with that approach?

Yes, if you want to hear about how Romney is going to win the popular vote by anywhere from +3 to +11 percentage points (that last, highly probable number coming from a Web survey run by the same guy running “UnSkewed Polls” — because how could there be a bias in that?), then head on over to UnSkewedPolls.com.

And while you’re at it, why not place a friendly, Romney-sized $10,000 bet with me on whether Romney will win the popular vote with greater than 5 percentage points — a highly likely outcome, according to “UnSkewed”? I’ll take “no”, because I could use the money.

But yes, thank goodness, there is finally a site out there that “unskews” the inherent liberal bias in the polling done by the liberal stalwarts at Fox News.

Yes, Fox News, as part of the conspiracy, are lying to you when they say that Obama is ahead of Romney by 5 percentage points. Lies! Liberal lies! But hey, when “unskewed”, Romney is now magically ahead by 3 points. And isn’t that good news?

tODD@76:

Uh, no. In the Rasmussen poll–which is more reliable than, say, the NBC poll–it’s a dead heat. I highly doubt that Romney is 11 points ahead of his opposition. I am almost certain that the race is currently within the polling margin of error, though. So, again, hysteria about polls is inappropriate. That’s my point.

Look, it’s a fact: a trend in recent polling has been to skew the polls in favor of Democrats by a few points. Again, just a fact. The alleged reason for this–i.e., the one passed off by those who publish skewed polls–is that, especially in swing states, Republican turnout was allegedly inordinately high in 2010 and 2011 (and in the recall elections here in 2012). So they’re trying to make up for that “discrepancy” by tossing a few points in the Democratic direction. In fact, as much was explicitly admitted on the television news here in Wisconsin (a swing state) last night. I’ll leave it up to you to decide whether that tactic is or is not scientifically dubious. But it is what it is.

Anyone interested in the methods and models that unskewed.com uses, can find a number of links on the sidebar at their website.

Todd @ 76 (and anyone else who think that those complaining about the polls are a bunch of conspiracy theorists)

Look, I will make this really simple. Just ask yourself this question. Do you think there are more registered and/or self-identified Democrats today than there were in 2008? If so, then of course Obama will be re-elected and it won’t even be close.

If, however, partisan identification is roughly equal now (the only pollster I know that is really even asking that question consistently is Rasmussen, he asks it every week in the form of the generic Congressional balloting. The latest offering had Republicans leading by 1 pt., which is consistent with where it’s been for the past 2 months) and the electorate in November has a roughly even breakdown by party – then guess what – it will be a close election and it will come down to independents.

It’s not that hard. It’s not a conspiracy theory. It’s just math.

By the way – this isn’t new. In every election since 1976 the media has consistently in its polling (including Gallup most of the time) overstated support of Democrats in all polling right up until the last week or two before the election. It has run anywhere from 2 pts to 10 pts.

Again, a lot of angst and hand wringing about something that is pretty predictable, not new and easily explainable.

Victor Davis Hanson has an interesting write-up about how the 1980 election that highlights the serious gaffes Reagan made on the campaign trail, the angst among Republicans, polling and more.

http://www.nationalreview.com/articles/328582/election-1980-style-victor-davis-hanson

tODD,
Ok, got my number a little high — fuzzy reading at the end of the day. Will you stipulate to “Overwhelmingly Democrat” as in “it would be more than a landslide blowout”? If not, you’re not being honest with yourself. Regardless, the underlying point still stands.

93% — http://www.usnews.com/news/blogs/washington-whispers/2011/06/16/book-liberal-media-distorts-news-bias

81-94% — http://conservapedia.com/Liberal_bias

Admission of such in their own words (yes, it is the MRC — but this is the media in their own words): http://www.mrc.org/node/29251

Actually, I would suggest that they cannot be objective. No one can. The question is whether they will own their bias, openly admit to it, seek a counter-balance to ensure they don’t miss the real story instead of their desired narrative, and then adjust when they blow it. Most of the media simply will not. I freely admit my bias: conservative. Then again, I’m not a journalist.

And yes, I DO believe that they shape the news in such a way that it benefits the President and harms the Governor. The media echo chamber is patently obvious to all but the most left-leaning partisans. Good grief, though rarely, the honest ones among them admit it themselves (again see the page from the MRC).

So unless you got something that proves their reporting neutrality in the face of their obvious bias, I don’t believe you have much of an argument.

Since you bring up ad hominem, just for fun, Fox News and MRC only look like the Right Wing when you are more than 2 standard deviations to the Left. 😉

And by the way, you are right about Rasmussen. Probably the fairest of them all. Even when I don’t like his results.

Prager University did a study on media bias and it’s consequences. UCLA Professor of Political Science and Economics, Tim Groseclose has a short video that gives an overview: 93% of D.C. journalists vote Democrat and that percentage is greater than the percentage in Berkley, Ca. It’s a worthwhile 6 minute report.

http://www.youtube.com/watch?v=PpBW6Tx2bGs

I ssee a hanging chad in our future.

NavyChaps @71

Using two situations both from the Old Testament, ie: Esther and King David as examples, has nothing to do with the present situation, which is two men, both with reasons mentioned beforehand to be president of this country.

God ALMIGHTY will direct the outcome of this country, HE allowed this country to come this far, HE has a plan. My conscience will not allow me to choose one evil over another.

The Mormon issue is too big, and most of all contrary to God to even consider. Obama is not an option, his views on abortion, taxes, health bill are a disaster. One is left with no one to vote for.

YOU WROTE: → “Are Christians thus forbidden to work with and/or support non-Chrisitians as they fulfill their vocations in the world (since God is also working through them). What about the work of various organizations or corporations — does that requirement prevent any association with them as well?”

Anyone can decide who they will or will not work for. There are many corporations and businesses I would decline for employment.

“Unskewed Polls”

Tee-hee.

What a horselaugh.

Orwell wins.

Black is white in Susan’s world.

And white is black.

Cincinnatus (@75, 77), I am legitimately confused by what you’re saying here.

Anyway, the polls we’re looking at–Quinnipiac, Franking and Marshall, Gallup, Rasmussen, etc.–often skew anywhere from 2 to 11 points in favor of Democrats.

First of all, what do you mean “we’re looking at”? Second of all, what do you mean by “skew”? That the final outcomes are 2 to 11 percentage points higher than they “should” be?

The ones on RCP right now are NBC, AP, etc.

What? Look, I don’t know what algorithm RCP is using to come up with its final results, but they list a lot more than just media-backed polls in their polling data.

And I don’t get your first paragraph (@77) at all, given that you’re responding to my previous post (@76) which said nothing about Rasmussen. Did you miss that I was mocking the “UnSkewed Polls” methodology, which is where that Romney-by-11-percentage-points nonsense came from?

Look, it’s a fact: a trend in recent polling has been to skew the polls in favor of Democrats by a few points. Again, just a fact.

Then it won’t surprise you if I ask for a source for this fact. Facts are easy to cite.

So your claim here is that this alleged pro-Democratic “skewing” is only since 2010?

Anyhow, from the research I’ve done into this, you’re wrong. Major pollsters simply do not weigh party identification when they conduct their polls. Now, one could argue that they should, in fact, start doing this, but that would be an entirely different argument than you or others are making.

Point being, when pollsters weight their data, it is to provide a sample that is statistically representative in terms of race, education, marital status, and so on (but not party ID). Thus, there is no reactionary skewing or fudging in order to match some historical antecedent. I cite the following articles (which I fully realize come from arguably left-leaning sources, but let’s try and focus on the actual claims being made, people):

[1] tnr.com/blog/electionate/107778/the-polls-dont-use-2008-turnout-model
[2] guardian.co.uk/commentisfree/2012/sep/27/unskewing-polls-party-id-bunk
[3] guardian.co.uk/commentisfree/2012/sep/17/weighting-polls-party-identification

Steve Billingsley (@79) said:

In every election since 1976 the media has consistently in its polling (including Gallup most of the time) overstated support of Democrats in all polling right up until the last week or two before the election. It has run anywhere from 2 pts to 10 pts.

Okay. Show me the evidence of this.

Steve (@79), VDH also called Romney “a confessed conservative”. Ha! That’s awesome. I wanna live in his world.

Here is an interesting video with two professional pollsters discussing the challenges and problems with weighting polls (and yes, the subject here was addressing the weight given by party component). http://video.foxnews.com/v/1773842238001/campaign-insiders-8-6-2012

NavyChaps (@80), first off, you should know that more than 2 links or URLs in your comment lands it in the quarantine, which only gets checked sporadically. (The one exception to this is if you chop off the bits WordPress uses to identify URLs as URLs — namely, the “http://” or “http://www.” bits — as I did at the end of comment #85.) FYI.

Will you stipulate to “Overwhelmingly Democrat” as in “it would be more than a landslide blowout”? If not, you’re not being honest with yourself.

Come on. Did you even read my comment (@74)? Here, let me repeat this part for you again:

In a 2004 survey of 300 journalists (most recent I could find on their site), 156 voted for Kerry, and 57 for Bush. Lopsided? Yes. 94%? Um, hardly.

(Side note: I should’ve said “admitted to voting” instead of “voted”, but whatever.)

But the data you’re giving me is, frankly, kinda lousy.

The 93% number you tossed out is hardly supported by the link you provided. The relevant bit says:

Groseclose opens his book quoting a well-known poll in which Washington correspondents declared that they vote Democratic 93 percent to 7 percent…

When was that poll taken? How was it taken? Fine. You made me research your own claim. That poll is from (surprise) a Media Research Center survey from 1995 that polled 139 “Washington bureau chiefs and congressional correspondents”. The footnote in the book literally says that the figure comes from the same MRC page I already linked to (@74).

Okay, then. So on to your “81-94%” claim, with a link to (shudder) Conservapedia. Okay, so in the section marked “Statistical data”, Conservapedia repeatedly makes claims about what percentage of media professionals “voted for” this or that candidate. The dates of these claims? 1964-1976. So that’s fresh data. All the footnotes for these claims say they come from “Robert Lichter and Stanley Rothman’s 1981 survey of 240 journalists at top media outlets”. That’s because those footnotes and bullet points are all cribbed straight from this FrontPageMag.com article. However, correct me if I’m wrong, but neither FrontPage nor Conservapedia state where that survey was published. It’s a mystery.

However, the top footnote on the FrontPage article says that

Summaries and analyses of most of the research cited in this article can be accessed from the Media Research Center, which does an outstanding job of documenting media bias and its many ramifications.

Hey look, it’s the Media Research Center! Again.

I’m gonna go ahead and take a wild guess that, honestly, the MRC is the sole source behind all these figures. And most of the claims out there about media voting habits.

Which isn’t to say that the claims are, as such, wrong. But you do realize that MRC has an (admitted) axe to grind, right? And, as such, it’s possible that they’re presenting the data that most makes the case that is the very raison d’être for their existence?

I’m very uncomfortable with the fact that they appear to underlie every claim you’ve offered so far. I’d feel a little better if there were more than one source behind these claims — especially one that didn’t depend so much on one particular result.

And, once again, I’d like to see some fresher data. Because, again, the most recent data from the MRC points to more of a 3-to-1 split among journalists in favor of voting Democratic.

With Start of Early Voting, Election Day Becomes Election Month

September 27, 2012
By JEFF ZELENY

DES MOINES – “As eight bells rang from the clock tower of the Polk County Courthouse, the doors to the election office opened on Thursday morning and voters began casting the first ballots of the presidential race in this battleground state.

“It seems like we’ve been waiting for this day for a long time,” said Nancy Bobo, 60, who stood in line and voted for President Obama. “I’m just thrilled to get out here and vote as soon as I possibly could.”

Less than a week before the president’s first debate with Mitt Romney and a month before the closing arguments of a campaign traditionally would be made, a steady stream of voters walked into election offices across the state to cast their ballots. They will be joined by voters in Ohio next week, along with 30 states where some type of voting is already under way.“

READ THE REST:

http://thecaucus.blogs.nytimes.com/2012/09/27/with-start-of-early-voting-election-day-becomes-election-month/?hp

The article above @90 was taken from The New York Times.

NavyChaps (@80) said:

Actually, I would suggest that they cannot be objective. No one can. The question is whether they will own their bias, openly admit to it, seek a counter-balance to ensure they don’t miss the real story instead of their desired narrative, and then adjust when they blow it.

Oh good grief. Did you even notice how you laid out a rubric for objectivity right after you claimed it can’t be done?

Look, I’m not making a case for pure and complete objectivity, whatever that is. I’m just arguing against the claim that is necessarily at the heart of your argument: that admitted Democrats are literally incapable of writing an objective news article. That they necessarily must be biased according to their own personal voting habits.

Again, if “they” (that is to say, everyone) cannot be objective, then neither can you, and I simply cannot trust anything you say. You’ll notice I’m doing you the favor of not responding according to that “logic”, however, instead checking your facts, not according to who you are as a person (ad hominem), but according to the available data.

The problem with the MRC is that they cherry-pick their data. They see what they want to see. You want to see the same thing for the left-wing side? Go to Media Matters. Hey presto, the media has a horrible right-wing bias now! Quelle surprise!

And yes, Fox News the channel does have an obvious right-wing bias. But their news programs aren’t quite so bad. As always, it’s the opinion segments that are patently biased. Come on. You can admit it. Just like I can admit MSNBC’s talking heads are raging left-wingers. Not that I watch either, as I don’t have cable.

Also, I’m not the liberal you probably think I am. I’m much more in line with the libertarians, especially those you’ll find on this blog. Didn’t vote for Obama last time, and probably won’t this time. But right-wing partisans on here tend to mistake me for a line-toeing left-winger. Whatever.

But now this conversation is well and truly derailed.

93% and media bias quote:

“In a typical presidential election Washington correspondents vote about 93-7 for the Democrat. This was first documented in a 1996 Freedom Forum report. In 2008, the New York Times’ John Tierney found similar numbers. That is, in an “unscientific poll” of his colleagues, he found that Washington correspondents preferred Obama over McCain by a ratio of 92-8.”

http://ricochet.com/main-feed/Political-Views-of-Journalists-My-Feud-with-Eric-Alterman-Part-2

And if that isn’t satisfactory, here is an exhaustive list of references that deal with the subject:

Quarterly Journal of Economics article
http://www.sscnet.ucla.edu/polisci/faculty/groseclose/pdfs/MediaBias.pdf

REFERENCES
Alterman, Eric, What Liberal Media? The Truth about Bias and the News (New
York: Basic Books, 2003).
Baron, David, “Persistent Media Bias,” Journal of Public Economics, LXXXIX
forthcoming (2005).
Black, Duncan, The Theory of Committees and Elections (London: Cambridge
University Press, 1958).
Bozell, L. B., and B. H. Baker, That’s the Way It Isn’t: A Reference Guide to Media
Bias (Alexandria, VA: Media Research Center, 1990).
Crouse, Timothy, Boys on the Bus (New York: Ballantine Books, 1973).
Djankov, Simeon, Caralee McLiesh, Tatiana Nenova, and Andrei Shleifer, “Who
Owns the Media?” Journal of Law and Economics, XLVI (2003), 341–381.
Franken, Al, Lies and the Lying Liars Who Tell Them: A Fair and Balanced Look
at the Right (New York: Dutton, 2003).
Goff, Brian, and Robert Tollison, “Why Is the Media so Liberal?” Journal of Public
Finance and Public Choice, I (1990), 13–21.
Goldberg, Bernard, Bias: A CBS Insider Exposes How the Media Distort the News
(Washington, DC: Regnery, 2002).
Groeling, Tim, and Samuel Kernell, “Is Network News Coverage of the President
Biased?” Journal of Politics, LX (1998), 1063–1087.
Groseclose, Tim, Steven D. Levitt, and James M. Snyder, Jr., “Comparing Interest
Group Scores across Time and Chambers: Adjusted ADA Scores for the U. S.
Congress,” American Political Science Review, XCIII (1999), 33–50.
Groseclose, Tim, and Jeffrey Milyo, “Response to ‘ “Liberal Bias,” Noch Einmal,’ ”
Language Log, viewed December 20, 2004,

Sigh. Do you all work for the Media Resource Center?

Susan attempts another go (@93) at defending “93% and media bias quote”. So she links to a Ricochet article written by … (drum roll, please) … Tim Groseclose!

Yes, Tim Groseclose, the very same person Susan already cited (@81)! And, what’s more, the exact same Tim Groseclose that wrote the book in the US News article that NavyChaps linked to (@80)!

Yes, that’s right, the very same Tim Groseclose who, in that book, notes that his 93% statistic came from … the MRC! Something I already pointed out (@89).

Groseclose even flubs his own facts, however. Here he is in Ricochet:

In a typical presidential election Washington correspondents vote about 93-7 for the Democrat. This was first documented in a 1996 Freedom Forum report.

And here he is in Left Turn, his book:

In a typical presidential election, only about 7 percent of Washington correspondents vote for the Republican. Perhaps the most famous study to document this was conducted in 1995 by Elaine Povich, a fellow at the Media Research Center’s Freedom Forum and a former reporter for Newsday and the Chicago Tribune. She surveyed 139 Washington bureau chiefs and congressional correspondents. Seven percent said they’d voted for George H. W. Bush, while 89 percent said they’d voted for Bill Clinton.

Okay, so ’95, ’96, whatever. But in Left Turn, there’s a footnote that, again, goes to that MRC article. That single source certainly is getting a workout today.

But whatever, I kept reading in Left Turn, and it turns out there’s totally a second data point! Two! Yes, here’s the other one, from a 2004 New York Times article:

But do journalists really want John Kerry to defeat George W. Bush? It depends where they work and how you ask the question, at least according to the unscientific survey we conducted last weekend during a press party at the convention. We got anonymous answers from 153 journalists, about a third of them based in Washington. When asked who would be a better president, the journalists from outside the Beltway picked Mr. Kerry 3 to 1, and the ones from Washington favored him 12 to 1.

So do the math there. Roughly 50 people were surveyed, unscientifically. Anyone wanna guess what the margin of error is on that 92% number? Anyone?

Still, more problematic is that Groseclose would attempt to make a claim (at least twice!) about what happens “in a typical presidential election” based on two data points.

Sigh.

Susan @93 regarding your list

I don’t agree with tODD all that often, but you’re taking this too far.

I don’t speak for everyone, but it’s boring, you’ve made no point except that you must LOVE GOOGLE!

Klasie at 63
You know a party is in trouble when,
— it tries to limit the franchise.
— it ignores the legacy of its former leader.
— it offers such mediocre candidates as Caine, Bachmann, Gingrich and Perry for consideration in the Primaries

@tODD

One of the reasons you struggle is that you prefer to complain, criticize, nitpick, and condemn rather than take the time to study the material and find all of the data points and learn the arguments in full for yourself. Another fault is that you have failed to recognize that Groseclose work was smeared by undue attacks. His story is similar the recent studies by the professor whose findings about gay couples and children didn’t fit the politically correct narrative. Both men were attacked viscously even going so far as to call for them to be fired. Both have been vindicated by the academic community.

If you are still in the mood to sigh and carp, please start on the fairly exhaustive list of references posted and don’t miss the footnotes in the linked journal article. There is more than enough to there to answer your demands. Or perhaps you would like to continue to demand that what is good (eg: NavyChaps observation about objectivity) be destroyed in the pursuit of perfection on a blog?

@Grace

That was one of the points of the list. Research is tedious and time consuming. Blogs aren’t research libraries.

Susan (@97):

One of the reasons you struggle is that you prefer to complain, criticize, nitpick, and condemn rather than take the time to study the material and find all of the data points and learn the arguments in full for yourself.

Dearest Pot, meet my friend. His name is “Kettle”. I’m sure you’ll be best of buddies.

Another fault is that you have failed to recognize that Groseclose work was smeared by undue attacks.

I have two words to say in response to that sentence: “Um” and … well, the second one isn’t so much a word as a kind of expression where I stare at you with my mouth slightly open and screw up my eyebrows a bit, trying to work out why anyone would think that’s a reasonable reply … to anything, really. Anyhow, in the real world, we call that “criticism”, and it’s a fact of life. You may even find yourself employing criticism from time to time! No, really!

If you are still in the mood to sigh and carp, please start on the fairly exhaustive list of references posted and don’t miss the footnotes in the linked journal article.

Oh, please, honey. I know a snow-job when I see one. You’re not trying to convince me of anything. You just want to give me enough busy-work so I go away.

Waah. At the end of the day, you and NavyChaps are both shopping around arguments that pretend to come from many multiple sources, and desperately want to sound like they have extensive data behind them, but it’s actually two data points.

What’s more, now that I look back at NavyChaps’ original point (@66), he managed to mangle what little data there is. He said:

94-97% of the media votes Democrat

But the whole point of both the MRC and NYTimes data points was how the political correspondents in Washington, D.C. voted (in the mid-90s and in 2004). Both sources went on to note that political correspondents outside of D.C. didn’t vote anywhere near that monolithically in those two instances, and when you expand that to include other journalists, the numbers get more and more Republican. So claims about “the media” are entirely unjustified, based on these two very-well-worn data points.

But hey, don’t let facts get in your way.

Susan @98

I am a researcher, that’s what I mainly do, for over 7 years now. However, I don’t waste my time with the kind of research you’re serving up on this thread.

Susan, if you really knew how to research, you woldn’t have addressed the topic as you.

You know a party is in trouble when,
– it tries to limit the franchise.

pause and let that sink in

There is no democracy without a limit on the franchise, starting with one man, one vote and secret ballots.

Just to give John C. the benefit of the doubt, perhaps he would like to qualify that statement.

RE: post 100

My last sentence should have ended with ‘did.

In 2008 a smaller fraction of eligible white voters turned out. Millions of eligible white voters stayed home who had voted in 2004 and 2000.

Now the demographics of this group suggest they’d go fairly heavily Republican. If these voters turnout it will be a close election. If they do not then Obama will win based on a small unenthusiastic electorate.

It is somewhat of a moot point since we’re headed into another recession and we’re likely to be even worse off in coming years than we are now. Whichever party wins this election will be devastated in 2014 and 2016.

So true, tODD, don’t let the facts get in your way (eg: 93% reference is for D.C. not national journalists), it’s so much easier to dish out contempt and pat yourself on the back for senseless insults.

@Grace – you missed the point I was making by posting the reference list.

Susan @105

You didn’t have a point. You didn’t expect anyone to accept your reading assignment did you? Also adding:

“And if that isn’t satisfactory, here is an exhaustive list of references that deal with the subject:” LOL

@Grace – irony doesn’t appear to be your strong suit.

Susan,

“Irony” ? You must be kidding, but I doubt it.

SAL @103: “Millions of eligible white voters stayed home who had voted in 2004 and 2000.”

And they stayed home in 2008 because…?

Maybe could it be because they were unenthusiastic about McCain’s candidacy four years ago? If that’s true, do you really think those same voters are going to be more inspired by Mr. Romney? McCain’s history included the price he paid for this country as a POW for 5 years. The current candidate can’t point to anything like that. Will the heart-warming story of Romney’s time with Bain Capital endear him to those white voters who stayed home in 2008?

I also think it’s interesting how you seem to be pinning the GOP’s hopes on getting out the “white” vote. Perhaps you haven’t noticed, but quite a few white people vote for Democratic candidates.

This does, however, bring up my second point I made above (@13), that getting out the vote by means of a motivated ground organization is something I think Democrats are better at, and yes, I think that effectiveness is demonstrated more in minority communities. I find it interesting that no one has challenged or in any way engaged that particular point.

Or maybe that’s because people here believe the Republicans have a terrific ground organization themselves, one that’s getting ready to really show itself…… when? Hmmm? Yeah, that’s what I thought.

Didn’t read all the posts, and certainly don’t pretend to be able to analyze things the way you fine folks can, but truly, why should things turn around for the Democrats so drastically when 2010 was such a landslide for the GOP?

Has everyone changed his mind-or are more people simply losing theirs?

Susan 110

→ YOU WROTE: “Didn’t read all the posts, and certainly don’t pretend to be able to analyze things the way you fine folks can, but truly, why should things turn around for the Democrats so drastically when 2010 was such a landslide for the GOP?“

ANSWER: Romney wasn’t on the ticket. Romney being the presidential candidate is the problem – his cult membership.

→ YOU WROTE: “Has everyone changed his mind-or are more people simply losing theirs?

ANSWER: Losing ones mind because they refuse to vote for a man who believes he will be a god is reason enough.

“The Lord created you and me for the purpose of becoming Gods like Himself; when we have been proved in our present capacity, and been faithful with all things He puts into our possession. We are created, we are born for the express purpose of growing up from the low estate of manhood, to become Gods like unto our Father in heaven.“

A Discourse by President Brigham Young,
Delivered in the Tabernacle, Great Salt Lake City, August 8, 1852
Volume 3, page 93

Dr. Veith

Voting for a man who believes he will be a god, is contradicting Scripture. He was a Bishop at one time, his father was one of the 12. That’s about as high as one gets. For that reason the Evangelicals are not willing to put their comforts ahead of God Almighty and HIS love for us, in sending HIS Son to die on the Cross for our sins.

I believe it to be a sin, to vote for such a man who opposes God, and contradicts Scripture.

Who hath wrought and done it, calling the generations from the beginning? I the LORD, the first, and with the last; I am he Isaiah 41:4

Thus saith the LORD the King of Israel, and his redeemer the LORD of hosts; I am the first, and I am the last; and beside me there is no God.
Isaiah 44:6

Your mindset is the goal. Virginia, no. Fla, no. NC, no. Ohio, likely not. But, you never know til u know.

sg at 101
In late August, a federal three Judge panel blocked the Texas voter ID law passed by the Republican Legislature, saying the new restriction would place an unfair burden on Latinos, racial minorities and the poor.

tODD, you are in rare form today. I mean that as a compliment. That is all.

what Susan says 🙂

cheers!

Well, I missed a good discussion today, but I’ll chime in late with a few thoughts, based on the original post and the polls, not the 119 comments since then, obviously.

Nothing has changed in my own thoughts. Romney should and probably will win this election. It is still early in the general campaign, and the polls are reflecting the usual volatility as pollsters experiment with turnout models and weighting issues. No one knows yet who is really going to turn out to vote in this election, but the combination of historic economic problems at home and growing foreign policy disasters in the Middle East do not point to a scenario of Obama pulling away in this election. Polls that say he is make no sense, logically. View them with great caution.

Obama beat McCain 53-46, a 7 point spread. That is the starting point for Romney in this race. There is no way that Obama is going to get more than 53%, or beat Romney by more than 7 points. 2008 was a high water election for Democrats, with a historically high turnout of the demographic groups most likely to vote for Obama. He also received the votes of a lot of Republicans and independents who were tired of Bush, of war, and thought it was a good thing to change the tone of Washington by electing the first African-American president and electing a platform based on hope and change. Those factors are no longer in play. We all feel and know it — there are a good number of 2008 Obama voters who are going to vote for Romney this time, and one can find very few McCain voters who say they will vote for Obama this time. This will be a close election. Any poll that tells you Obama is going to win by more than 5 points, either nationally or in one of the true swing states (eg OH, VA, IA, FL, CO) is certainly an outlier. If anyone wants to bet me on this, I will gladly take your money.

Rasmussen is showing a tied race. He is polling every day, and has an outstanding record of accurately predicting presidential elections, including both the 2004 and 2008 elections. His polling is also showing a stable electorate, which makes sense at this juncture in the campaign, where there haven’t been any “game changers” yet. Gallup showed a tied race until two days ago, among registered, not likely voters (this usually favors the Democrats by a couple of points). The recent jump to a 6 point Obama lead has happened before, in both directions, and usually settles out after another cycle (7 days) of polling. We’ll see if it does this time, as they also shift to a likely voter screen on October 1 or so.

Dr. Veith, state polls are terrible predictors at this point in a campaign. The quality of state pollsters, such as Marist and Quinnipiac, is uneven (neither pollster has a particular solid record in presidential polling), and samples are too small. I guarantee you that the winner of the national popular vote will win the electoral college. At this point in a campaign, the national tracking polls are your best bet for understanding the state of a race.

Don’t pay attention to things like “Unskewed Polls”. You can’t fix bad data by making it worse. Throw the bad poll away, don’t try to fix it, especially when you do not know or understand all of the assumptions the pollster made in constructing the poll.

IF the electorate in 2012 is similar in composition to the electorate in 2008, then the polls you have been seeing showing, on average, a 4-7 point Obama lead are probably in the ballpark. However, that is why I believe the polls are off — it’s not a conspiracy, it’s an error in judgment. Most pollsters don’t weight for party identification, but they do weight for demographic characteristics. This is good polling science because party id is an attitude and demographic characteristics are immutable. BUT, they are weighting demographics to 2008 exit polling data. In 2008, African Americans comprised about 14% of the electorate, even though they are only 11% of the population, and usually vote at about that proportion or lower. Young adults comprised 18% of the electorate in 2008, whereas they are usually 12%. Single women were also overrepresented in 2008, though I don’t have the numbers in front of me. By weighting up their polling samples to reflect these demographic groups, they are weighting up the strongly Democratic portions of the electorate. That is why the party ID for a lot of these polls has been so screwy, and why I think the polls are skewing toward Obama, based on the belief of many that these groups of voters will not turn out in anywhere near the numbers in 2012 as they did in 2008.

Another thing pollsters are doing, for the first time in a presidential election, is trying to poll cellphone-only users. Now, this is a good idea, as they comprise an ever-increasing proportion of voters, but there is no data as to how to weight this demographic group, which also tends to skew Democratic. So each pollster is using his own judgment as to an appropriate weighting. It remains to be seen who will be correct.

Republicans historically do better in national elections than polls predict. Many polls in 2008 predicted an Obama win of 12-15%. He won by 7. This is typical. Recall that in late October 1980, Gallup showed Carter ahead of Reagan by 7 points. He lost by 9, two weeks later.

So, I don’t believe there is any conspiracy going on here. Just some bad polling assumptions. This will be a turnout election. The key is to work hard and turn out your voters in a greater proportion than those of the other guy.

When the news is bad Don, I turn to Fox news. It’s news you can trust.
Fox has Obama ahead by 4.

DonS wins the thread. Everything he says about polling is true, and you should listen to him. Seriously.

Good comments, DonS.

Todd @ 123
Wow, really? After all of this back and forth you basically agree with most of what DonS said (you want sources, OK fine, if I have a few hours today, which I won’t, I can put together a set of links).

DonS is right, the polling discrepancies aren’t a conspiracy, they are possibly errors in judgment, possibly groupthink or maybe they are right after all

As far as weighting demos to 2008 exit polling data, I don’t know that they are doing that, but based upon the partisan breakdown of these polls that would make the most sense. In fact, it is the most generous interpretation because it means that these pollsters are using a fixed, known data point to tie their weighting to (as opposed to just making stuff up).

At the end of the day, we will all know in about 6 weeks anyway and unlike many conservatives, I don’t think an Obama re-election is the end of the world. I don’t think it would be the best outcome, but I think a Romney victory would only be marginally better. If Romney wins and if (a big if), his administration takes the painful steps that are necessary to put public finances on better footing, it won’t magically cure anything. The economy might recover some of its footing, but I still think there is plenty of economic pain that we are facing, even under the best of options. Our economic footing has been poor for some time and a lot of it has very little to do with Obama (or Bush or any other President).

tODD @ 123: I’ll see what I can dig up in the way of specific links — I posted last night based on memory, as this has been a crazy week, and I sucked up too much time commenting about football officials on Wednesday 🙁

I was a math minor, and took some statistics courses in college, but am by no means a polling expert. The people I respect in this field are Michael Barone, Jay Cost, and Mark Blumenthal. Both Barone and Cost have written and given interviews on the weighting issue, and the usage of 2008 exit polling data to weight their samples, and I have confirmed in a couple of the recent polls that publish their internals that the weighted demographic percentages seem to roughly line up with 2008 numbers.

Barone was an elections expert for U.S. News & World Report for the bulk of his career, but is clearly conservative, as is Cost. Blumenthal was a Democratic pollster who now blogs for the Huffington Post @ pollster.com. In 2008, Blumenthal wrote a lot of insightful articles that countered some of the hysteria that prevailed on the right concerning those polls, and he was a big advocate for polling cell phones. Unfortunately, so far this year he seems to be buying the polls he’s seeing without comment. I am waiting for him to address the charges of those on the right that the sampling is suspect with some scientific counter-evidence, but so far he has not.

Things to be cautious about:

1) Those whose candidate is behind in polls always complain about poll accuracy. Those whose candidate is ahead accept them without question. All in all, I’d rather be ahead in the polls.

2) Anyone who is claiming that an election is over in September because one candidate has assumed a 4-7 point lead in the polls is a hack. A partisan one at that. That includes supposedly “objective” media writing Romney doom stories currently. All they are doing is providing credible evidence for those who see a media polling conspiracy. This race is far from over, and the polls will swing in both directions between now and November 6. The Benghazi issue, for example, has yet to register with the electorate, but it will as more facts come out.

3) Until the last couple of weeks of a campaign, the “horserace” numbers are far less important than trends and polling on specific issues, which guide campaign advertising strategy. Clearly, in the past week, the polls have been almost uniformly trending toward Obama. That trend, to some extent, is probably real, though the horserace numbers are likely inaccurate. The question is whether this is just a normal cyclical trend or whether it is something deeper. My own suspicion is just that as the race gets later the lower information voters are starting to engage, and register preferences in the polling, and they lean heavily Democratic. If so, then the trend isn’t a big deal.

4) The race will tighten substantially in the polling in October. Polls almost always tighten in favor of a Republican candidate as Election Day approaches, because likely voter screens get tighter (almost everyone says they will vote in September) and Republicans almost always vote in greater proportions to their registrations than Democrats do (2008 being the one exception that I can think of).

Steve (@125):

Wow, really? After all of this back and forth you basically agree with most of what DonS said

Well, yes, with the reservations I noted. Why does that surprise you? Did you read the conversation? Most of the comments weren’t actually about polling methods. They either alleged a conspiracy without offering up any evidence, or they were about more general claims of “liberal media bias”.

the polling discrepancies aren’t a conspiracy, they are possibly errors in judgment, possibly groupthink or maybe they are right after all

Sure. I hesitate to say that we’ll know definitively after the election, because it’s short-sighted to think that your model works based on one data point. But we’ll at least have one more data point then.

Still, my understanding is that the lack of party-ID weighting is nothing new. As such, it’s already proven its worth as an approach — at least, for pollsters who correctly predicted things in the past within their margins of error.

it is the most generous interpretation because it means that these pollsters are using a fixed, known data point to tie their weighting to (as opposed to just making stuff up).

Well, even using a fixed data point is still “making stuff up”, because you’ve chosen to hang your hat on that data point, even if it is determined by someone else.

Polling isn’t a pure science — if it were, we’d be getting much more consistent results. No, there’s an art side to it, as well.

I’d be surprised if pollsters were simply trying to make their current sample’s demographics equal those of 2008 voters. Some pretty obvious objections to that approach leap out to me. But then, I’d be surprised if any pollster — much less most of them — would reveal what they’re doing in this regard with any level of detail.

Again, feel free to show me where I’m wrong.

I was looking at polls a couple of days ago, including this Public Policy Polling (PPP) poll of Iowa. Right on the first page, I read this:

“Barack Obama is starting to pull away in Iowa,” said Dean Debnam, President of Public Policy Polling. “Voters there just really don’t like Mitt Romney. It would be more competitive if the Republicans had a half decent candidate.”

Now, if this particular pollster wants to be take seriously as an objective non-partisan, she does not say that it would be more competitive if the Republicans has a “half decent” candidate. That is a pejorative beyond belief. This is the kind of stuff that fuels conspiracists’ fires, and understandably so.

You would think pollsters would work very hard to justify and explain their work, the reasons for making the assumptions they are making, and to appear objective. But, at least in many cases, you would be wrong.

@ Other Gary #13

http://m.washingtonexaminer.com/abortion-advocacy-hurts-obama-in-swing-states/article/2509113

DonS @ 128

But pollsters are neutral, none of them have an agenda! (Sarcasm intended)

PPP is very much a Democrat outfit. That being said, they are generally pretty good with their polling. But the comment by Dean Debnam is pretty inappropriate to be sure.

DonS (@128) said:

Now, if this particular pollster wants to be take seriously as an objective non-partisan, she does not say that it would be more competitive if the Republicans has a “half decent” candidate.

Well, first of all, he. Second of all, Debnam has made no secret of his Democratic leanings, so I don’t know where you get the idea that he wants to be tought of as a “non-partisan”.

Nor do I think that Daron Shaw — who served as a consultant for the Bush-Cheney campaign and the Republican National Committee — at Shaw & Company, which does polling for Fox News, is terribly worried about coming across as non-partisan.

This is just another variation on the oft-heard right-wing complaint that someone cannot be objective if they have political views. Or, rather, what you actually hear is that acknowledged Democrats are not capable of being objective, and must by nature skew their work in favor of their political opinions.

It is, at its heart, an ad hominem fallacy.

Steve @ 130, tODD @ 131 — Oops. “He” it is.

I do understand that PPP is a Democratic outfit, but if you want to be taken seriously in the objective world, which they do, and which they are, they should refrain from partisan statements in their polling reports.

I have never seen Shaw make such statements in his reports, and FoxNews polling generally falls into line with other polls. I’m just sayin’, a poll with a statement like that should be disregarded, or at least discounted, on its face.

Objectivity is a myth, in general, but DonS actually makes a fair point. If you claim to be a relatively impartial reporter of political reality, you probably shouldn’t go trumpeting your ideological biases if you can help it. In fact, a political scientist here at UW (Charles Franklin; founded Pollster.com, then sold it for millions) claims that he doesn’t even vote. Ever. He claims it would compromise his objectivity.

This might be a tad extreme, but yeah. A reliable pollsters profession involves reporting statistics in as neutral a fashion as possible. At least maintain the appearance of neutrality.

Two last bullet points regarding polling and partisan identification….

http://www.rasmussenreports.com/public_content/archive/mood_of_america_archive/partisan_trends/summary_of_party_affiliation

A few snapshots.

November 2006 – Democrats had a 6 point advantage in the Rasmussen party ID poll, makes sense that year in the Congressional elections they won the national popular vote in House races by 8 points (52-44)

November 2008 – Democrats +7.6% – Obama won by 7 points

November 2010 – Republicans +1.3% – Republicans outperformed this by winning the popular vote by over 6 points (51-45)

September 2012 – Republicans +4.3% (this is a bit of an outlier, the previous quarter they had averaged a +1.8%, down from a +3.3% the quarter before)

So tell me, how realistic is for pollsters to be using a Democrat +8, +10 or +11 turnout model?

Presidential approval (According to Rasmussen)
Six weeks before the midterm elections (September 27, 2010)
Approve – 48, Disapprove 51, -3
Strongly Approve 27, Strongly Disapprove 41, -14

Six weeks before the Presidential election (September 27, 2012)
Approve – 48, Disapprove 51, -3
Strongly Approve 27, Strongly Disapprove 41, -14

So again, does this look like an Obama landslide waiting to happen?

Here’s a link to a recent article by a Pennsylvania pollster, Susquehanna, regarding turnout models in 2012: http://sprblog.wordpress.com/2012/09/24/why-our-recent-polling-is-accurate-and-shows-a-close-election-for-president-by-james-lee-president-of-spr/

It is useful because it explains some of the things I was saying in my comment last night, regarding demographic turnout models, and the assumptions pollsters have to make when weighting their raw survey responses to reflect expected voter turnout.

Susquehanna is a Republican polling firm, but it has a good record. Here is their mid-October poll for 2008, which was pretty spot-on to the final PA vote breakdown, just to show that they aren’t always drinking the Kool-Aid:

http://www.publicpolicypolling.com/pdf/2011/PPP_Release_IA_926.pdf

Susquehanna had Obama by 8 in PA at this point in time. The final RCP average was 7.3, and he won by 10.4.

I did find a good weighting article by Mark Blumenthal from earlier this year: http://www.huffingtonpost.com/2012/06/17/gallup-poll-race-barack-obama_n_1589937.html

It’s pretty statistically intense, but highly interesting, whether or not you agree with Blumenthal’s polling approach and assumptions. That’s why I respect him as a pollster. It should help to illuminate the universe of assumptions that go into polling, and that we are at a crossroads in polling, going into uncharted territory because of cell phones and changing demographics. For example, weighting survey respondents in an “all adults” poll to the Census statistics is perfectly reasonable. But how do you adjust those weightings for registered, and even more importantly, likely voters? Even the experts don’t agree.

DonS (@132), Cincinnatus (@133), you appear to be making claims about best business practices for pollsters. And, sure, if a pollster largely earns his money by doing political polls for non-partisan outfits, then yes, it might not be best to make statements as Debnam has here.

I’ll be honest, I’m not entirely sure how most pollsters make money. But it doesn’t appear to be via these national presidential tracking polls — which, and I’m guessing here — appear to be more of a way to get your name out there and earn a reputation as a pollster.

A Politico article mentioned that:

The liberal website Daily Kos, with sponsorship from the Service Employees International Union, has commissioned more than 100 PPP polls to be conducted and published over the course of the 2012 election cycle. But that’s the exception, not the rule, to how the company usually makes its money.

But I’m going to go out on a limb and guess that neither of those organizations think Debnam’s “half decent” comment casts him in a negative light.

But DonS (@132) said:

a poll with a statement like that should be disregarded, or at least discounted, on its face.

Once again, this is an ad hominem fallacy. This man’s data cannot be accurate, you argue, because of his personal beliefs. I can’t say it much more plainly than that.

Contrary to Cincinnatus’ claim (@133), I would say that a reliable pollster’s job is to provide reliable data. If the pollster is biased, and providing the data to those who are similarly biased, then attempts at neutrality in the presentation of said data are probably not worth much. It’s the data people want.

Of course, if one’s personal beliefs made one’s data unreliable, then that would again be a business case for (at least attempts at) neutrality. But no one’s really arguing that here, at least from the actual data. You’re asking me to believe that the data is unreliable because of Debnam’s statement.

Steve (@134), what, exactly, is the point of your analysis there? You seem to be trying to draw a direct line between party affiliation and actual votes cast.

But your analysis seems also to rely entirely on Rasmussen’s party affiliation data. If one goes over to Gallup, one finds fairly different data.

Rasmussen doesn’t say what their affiliation question is like Gallup does, but, using your snapshot times, whereas Rasmussen finds R affiliation higher than at any point since 11/2004, Gallup shows R affiliation declining 10 percentage points over the same period.

Gallup also shows a marked increase in independent affiliation over that same period, from 27% to 41% today. Rasmussen, in contrast, has “other” affiliation holding mostly steady between 24% and 29% since 11/2004.

And, most to your point here, Rasmussen has seen the R-D delta change a whopping 12 percentage points in favor of Republicans since 11/2008, whereas in that same period, Gallup only records a 2-point shift towards Republicans.

It would seem that most of the Republican complainants are relying solely on Rasmussen’s party-affiliation data. But is it reliable? Or, rather, is it so much more reliable than Gallup’s data that we must assume that it alone is correct, and Gallup’s data is trash?

And what would the case for that argument be?

tODD @ 137: Two polling outfits which were prominently featured in the media earlier in this century have been almost totally discredited since 2008. One of them was Strategic Vision, a Republican consulting firm that did a lot of media polling in 2000-08. Another was Research 2000, which did polling for the Daily Kos in 2008. It turned out that both outfits were cutting corners, and in some respects faking their polls. No one knew that, of course, until after the fact. It helps us to remember and understand that our relationships with pollsters is a matter almost entirely of trust — there is no way for us to know whether or not the pollster is constructing its samples scientifically, making reasonable turnout assumptions, or actually conducting the interviews they say they are conducting.

In such an environment, where we have an increasingly partisan media commissioning the polls, and an ever greater number of polling companies offering their product for our consideration, I don’t see any reason for me, as a conservative, to give any heed to a pollster who is so partisan that they cannot even keep their partisan statements out of their polling reports. You can, if you want to. But I do not, and will not, trust and give any credibility to a PPP poll.

I would say that a reliable pollster’s job is to provide reliable data. If the pollster is biased, and providing the data to those who are similarly biased, then attempts at neutrality in the presentation of said data are probably not worth much. It’s the data people want.

This statement is true. But, unfortunately, we cannot know if the pollster is providing reliable data until after the election. So, it is trust that is at issue. I am not asking you to believe anything. I am telling you why I think there is good reason to distrust PPP.

DonS (@135), first of all, your second link up there is to the wrong article, as you can plainly see.

But what is the point of your first link? Yes, there’s an example of a state polling firm that actually tries to fit their current dataset to “turnout models from both the 2008 and 2004 presidential elections”, but also, notably, this firm says that they use “party affiliation” as one of their demographic considerations.

Thus, Susquehanna is not an example of what you were talking about earlier when you said (@120), “Most pollsters don’t weight for party identification, but they do weight for demographic characteristics.”

Here’s a column by Mark Blumenthal (whom you respect) on why trying to adjust a sample according to what you think the party affiliation should be is a bad idea. And here’s another column from Gallup’s editor-in-chief saying much the same thing.

So that’s two pretty respectable voices (one by your own admission) saying that Susquehanna’s party-affiliation practices are a bad idea. And if Rasmussen also factors in party-affiliation to its results, well … you know.

So has anyone pointed out yet that polls are stupid and useless? Because I think someone should point that out. That would be pretty neat.

tODD @ 140: Oops. Here’s the right link. Too many documents up on my browser at the same time, I guess.: http://www.pittsburghlive.com/images/video/2008_pdfs/Poll%20Memo-PA-Statewide-Oct08.pdf

I don’t have time to address your substantive comments until later this weekend, and I haven’t found links to the demographic weighting issue yet that are primary sources — just haven’t had time to look today. Hopefully more later.

I think we’re getting close, Cincinnatus @ 141. At least they’re stupid and useless for declaration an election over IN SEPTEMBER, based on a single digit horse race lead!

No, I am making 2 points.

1. Rasmussen has been fairly reliable and should be respected (which is really only a secondary point and a tangent to the second point) – BTW, Rasmussen polls likely voters and Gallup tends to poll registered voters. Rasmussen also records “leaners”, that is to say independents who say the “lean” Republican. That is a different screen so their numbers tend to look different. They are respected pollsters because they are consistent in their methodologies and have over the past decade had the best results. And as far as actual votes are concerned, there was a 8-10 point delta swing in votes from 2008 to 2010.
2. The thought that Obama is on is way to a landslide is on very shaky ground. The polls that show him with a wide lead have assumptions built into the weighting of the data that are counter to historical data and recent trending. This looks to be a close election, period, full stop.

(an addendum, party affiliation has a lot to do with actual votes cast, over the past decade Presidential candidates, even losing ones, get 85-90% of the vote from voters affiliated to their party. If party affiliation is roughly even – then whoever wins the independents wins the election – which makes the worst of these polls even more suspect. In every one of the polls that I have looked into the cross-tab data, Romney has a slight or in some cases a substantial lead among independents, yet in some of these same polls Obama has a 5+ lead. The only way that is possible is if Democrats are either surging among likely voters in affiliation, which is a data point that I haven’t seen verified anywhere else independently or they are oversampling Democrats in these polls – what other conclusion would you make?)

What is the point of making your guy look like he is doing better? I mean psychologically speaking, if it looks like a dead heat, wouldn’t that make voters feel that they really need to get out there and support their guy? Because you know, when it is close, every vote really counts, or something like that?

Well, as predicted, the polls tightened substantially over the weekend. The question is whether this is as a result of the normal political cycle, because pollsters are tweaking their likely voter screens and demographic modeling, or because the Benghazi foreign policy disaster, and the administration’s lying about it being a mere protest of a silly video that had been on the Internet for months, is beginning to penetrate the minds of the voters. Or maybe it’s the Fast & Furious scandal that only Univision has had the willingness to cover adequately.

Back to polling: I haven’t found any pollsters who will specifically say how they model their voter demographics. As tODD says, that is their “secret sauce”. However, there are plenty of commentators who have been in the polling business in the past and are looking at the reported demographics of some of these reported media polls, leading them to believe that they were modeled, at least mainly on something approaching the 2008 election exit polls.

Gallup says on their site that they take their adult sample, and then weight that to 2011 census data for immutable demographic characteristics. This is exactly as they should do. Presumably, to obtain their registered voters sampling subset, they simply accept the word of those they poll as to whether they are registered, and then re-weight that subsample to the census statistics. That would be a reasonable thing to do, though it would almost certainly result in a registered voters subsample which is too big, because it is undoubtedly the case that people will say they are registered to avoid embarrassment.

The likely voters screen is where things get tricky. If you merely take your already too large registered voters pool and then ask them whether they plan to vote, using a set of questions designed to yield more accurate results, you are still going to get a very high percentage of “yes” answers, particularly in the earlier stages of the campaign. That is why most pollsters do not merely accept those answers as the final word, they also model their likely voter pool to an anticipated electorate. Some merely use immutable demographic characteristics to do that, and some also use voter id demographics. This is why we get wildly disparate results. Logically, if you see a sample that appears far more weighted toward one party than any recent voter id polls show, or than have occurred in any recent similar election, then you should question your sample and your modeling, at least against what you sense is actually going on in the electorate.

Michael Barone published a very interesting article over the weekend: http://washingtonexaminer.com/barone-when-it-comes-to-polls-readers-beware/article/2509360#.UGnKnK7wOSo

He was a Democratic pollster back in the ’70’s, and a political analyst for U.S. News & World Report for many years. His big concern with polling is the sharply dropping response rate. According to a recent report by Pew Research, only 9% or so of those polled are agreeing to respond. That survey is here: http://www.people-press.org/2012/05/15/assessing-the-representativeness-of-public-opinion-surveys/

In 1997 the response rate was 36%! Even in 2008, the response rate was 15%. So, we are really in uncharted territory here. At what point are we talking about self-selection bias, where the low number of respondents is not demographically representative at all? Who responds to pollsters? I know I don’t. The other thing that is going on now is cellphone polling, which has very low response rates (helping to drive down the overall response rate to 9%). Do you ever answer your cellphone if you don’t recognize the number? I don’t. Who does? Are those that are willing to do that representative of the rest of us?

These are questions pollsters are still wrestling with, but that the media won’t explain when they confidently trumpet that the results of a few September polls mean one candidate or the other has lost the race.

“At what point are we talking about self-selection bias,”

Isn’t that like accounting for lying? If the population lie at a constant rate then it can be factored in and adjusted for, right?

sg @ 147: Self-selection bias can’t really be adjusted for. At some point, your response rate is so low that you may only be getting extroverts, or lonely people, or jobless people, or seniors, for example, to respond to your polling. Busy, productive people or introverts may seldom respond, or may constitute an insignificant number of your sample size. At that point, you have to ask yourself, as a pollster, if your polling sample, those willing to talk to a pollster for 15 or 20 minutes on the phone, is representative of the population at large. Simply re-weighting it to reflect census demographic numbers won’t help to fix self-selection bias.

fyi 🙂

http://www.politico.com/blogs/under-the-radar/2012/10/a-pollster-under-oath-137100.html

cheers!

Most Political Science election forecasters have Obama squeaking by in a very narrow, far from certain, victory. Here’s a link to their results:

http://journals.cambridge.org/action/displaySpecialArticle?jid=PSC&bespokeId=4484

Troy@150:

Yes, that’s absolutely right. Of course–and perhaps ironically–political scientists have an utterly dismal record in forecasting elections.

@dust

Great link verifying the manipulation of polls plus with only 9% of the population being polled, it doesn’t make much sense to pay attention to them.

Oh good grief, Susan (@152):

Great link verifying the manipulation of polls…

Did you even read the article?

It did not show that any polls were “manipulated” as you have been maintaining here. The whole point is that the Edwards campaign was intentionally touting results it didn’t believe in. It wasn’t doing the polls, it was just announcing them. But they were shoddy polls, not good statistical work — “dog and cat and baby-sitter polls”, as Hickman said.

If this is news to you, you’re kind of naive. News flash: political campaign strives to make candidate look better than he actually is by cherry-picking the facts! Full story at 11!

But, again, that has nothing to do with the actual “manipulation of polls”. Which remains a baseless claim by conspiracy-minded folks who don’t like what the polls are saying.

…with only 9% of the population being polled, it doesn’t make much sense to pay attention to them.

Um, no way has 9% of the population ever been polled for anything, except the Census. That would require 28 million surveys. Not happening.

Still, you don’t get statistical sampling, do you? I mean, it’s one thing to question the value of polls themselves, but to dismiss them without understanding sampling doesn’t make sense.

DonS, you haven’t come out and said so, but I’m afraid I’m going to have to say you were wrong when you claimed (@120) that “they are weighting demographics to 2008 exit polling data”.

Sorry, but this is a fairly specious claim:

there are plenty of commentators who have been in the polling business in the past and are looking at the reported demographics of some of these reported media polls, leading them to believe that they were modeled, at least mainly on something approaching the 2008 election exit polls.

As we have now both noted, Gallup not only isn’t taking party affiliation into account, it also isn’t weighting the demographics it does consider according to anything but the most recent Census data. Which, as you note, is how it should be.

And Gallup currently has Obama beating Romney by 6 percentage points — well outside of the margin of error, and an increase of 1 point for Obama since last week.

Sure, some less reputable pollsters (especially at the local level) might consider party ID, but at least some of those are actually Republican (or Republican-leaning) pollsters, which kind of inverts the argument you and others have been making here.

tODD @ 154: I’m not sure you understand the arguments I have been making here. I won’t speak for others.

My first statement (“BUT, they are weighting demographics to 2008 exit polling data”) was written in haste, and a little bit imprecise. I did not mean that they were weighting exactly to 2008 exit polls, but that a number of polling commentators, former pollsters, are saying that they are using those polls to model their expected 2012 turnout. I re-stated this more precisely @146, which you quoted above, labeling it “fairly specious”. I’m not sure I get why you think that statement, as you’ve quoted it, is “a fairly specious claim”. Read some of the links I provided, and, for that matter, the Dick Morris links that Dr. Veith posted for evidence that commentators, with respectable credentials in the political and polling worlds, are saying this. I have not linked Jimmy Carter’s pollsters, Pat Caddell and Doug Schoen, but they are saying it as well. In their view, the most likely way these pollsters are obtaining such consistently lopsided voter id samples in their polling is by weighting the likely voter electorate to something like the 2008 demographics, meaning that they believe the electorate in 2012 will have a similar demographic composition.

As we have now both noted, Gallup not only isn’t taking party affiliation into account, it also isn’t weighting the demographics it does consider according to anything but the most recent Census data. Which, as you note, is how it should be.

Well, Gallup isn’t really an issue here, because it is only still reporting registered voters. We don’t know how they will model their likely voter turnout screens, though we can be fairly certain, based on history, that when they switch to likely voter reporting the results will tighten to what we are seeing from the other polls — Obama + 1-4 points.

Sure, some less reputable pollsters (especially at the local level) might consider party ID, but at least some of those are actually Republican (or Republican-leaning) pollsters, which kind of inverts the argument you and others have been making here.

Now, this is where I particularly don’t think you understand my point. I’m not talking about weighting to party id. Some pollsters do it, some don’t. In traditional polling science, it’s not done, though a lot of pollsters, including Rasmussen, do it now, in a scientific way, because they believe doing so compensates for the distortions being introduced by the extremely low response rates achieved in recent polling, with its worrisome self-selection bias. BUT, if you weight to Census demographics, and you don’t also further weight your likely voter pool to a particular turnout model, you will definitely have a result that favors the Democrat significantly. Minority voters bias heavily to the Democratic side in their voting patterns, but they typically do not vote in proportion to their demographic share of the population. Hence the Democratic bias in polling results, which has been evidenced in election cycle after election cycle.

@tODD

It is time to stop dishing out the insults and snarks at other commenters. It’s not necessary – all it does is coarsen the conversation and discourage some people from commenting.

Here is a link to Pew regarding 9% response rate: http://www.people-press.org/2012/05/15/assessing-the-representativeness-of-public-opinion-surveys that states: “The percentage of households in a sample that are successfully interviewed – the response rate – has fallen dramatically. At Pew Research, the response rate of a typical telephone survey was 36% in 1997 and is just 9% today.” That’s a very small sampling that requires more guesstimates in weighting and is not as accurate as Pew would like to assert. There are numerous good articles about the problems with the polls. I’m guessing you are capable doing your own research and can easily find numerous articles.

And yes, Hickman was manipulating the polls (eg: cherry picking and presenting bad ones) for “propaganda” purposes and so forth. It’s not a surprise that this kind of behavior would fall into the category of: manipulation of polls (access a dictionary if you are still puzzled). There are several other articles that have reported on Hickman’s testimony. If you need further evidence, please do your own research.

If anyone is interested, I did link to a recent video with pollsters, Pat Caddell and Doug Schoen, supporting DonS points. See comment #88.

Here’s another fun show-and-tell on the questionable tactics of pollsters in the recent Wapo/ABC poll:

http://newsbusters.org/blogs/noel-sheppard/2012/10/01/jennifer-rubin-forces-washington-post-explain-its-own-flawed-poll

Susan (@156), for someone complaining about “snark”, you don’t seem particularly averse to dishing it out yourself (cf. “access a dictionary if you are still puzzled”). And yet, you seem rather sensitive when it comes to being on the receiving end. If you can’t stand the heat … ?

Anyhow, it still doesn’t seem like you know what you’re talking about. Words have meaning, and a 9% response rate is nowhere near the same as “9% of the population”.

Furthermore, the response rate is not the same thing as the sample size. In order to achieve their desired margin of error, Gallup needs to get a certain number of responses. If the response rate is low, that just means they need to call that many more people to get their target.

I’m not saying that their (or anybody’s) approach is criticism-proof, but you have to understand the science before you can critique the art.

And, once again, cherry-picking from among many polls (some good, some bad) is not the same as “manipulating” polls, as you keep suggesting is happening even among reputable firms.

Try not to feel insulted by my pointing all that out. Those are the facts. If you disagree, then show me where I’m in error.

No, tODD, it’s not personal. If you remember, I have gotten after you for the same types of things against another commenter. And yes, please use the dictionary if you are puzzled or google if you want more information. No snark intended. I didn’t imply you were stupid as you so often do of others, but implied you are capable. There is also my recognition of the pattern in your replies when offered information. I cannot play an ongoing game of prove it with you.

And yes, I miswrote my thoughts then I typed population. I should have written 9% of a sampling and a small sampling at that. Sampling size matters. The much larger sampling sizes are why Gallup and Rassmussen are being given more credence and many other polls have been discredited. Try not to feel provoked to further snark if I point out that your understanding of manipulation and understanding of polling falls far short of the perfection you demand of others. It would be good to look in the mirror and recognize that you are no different than the rest of us.

Actually, Susan @ 160, the sample is the pool of interviews you take, based on your theoretically random calls. Back in 1997, if you called 3,000 numbers, you could expect, on average, to have 1080 interviews (90% contact rate + 43% cooperation rate = 36% response rate). The contact rate is the percentage of phone calls where you reach an adult and the cooperation rate is the percentage of those contacts where the adult agrees to the interview). As recently as 2009, calling 3,000 numbers, you could expect to have 450 interviews (72% contact rate + 21% cooperation rate = 15% response rate). Now, in 2012, calling 3,000 numbers, the average number of interviews in this Pew survey was only 270 (62% contact rate + 15% cooperation rate = 9% response rate). It is evident that the response rate is dropping precipitously.

However, the polling sample sizes still reach whatever target the pollster sets. They just call more numbers. For example, if the target sample size is 1080, then it would take, statistically, about 12,000 phone calls to get there, whereas in 1997 it took about 3,000.

Three problems with this: 1) Expense — making 12,000 phone calls instead of 3,000 is a drag, 2) Citizen fatigue — probably one significant reason for lower response rates is that people are getting called a lot more often, since there are a lot more pollsters than there used to be and they are having to call a lot more numbers; 3) self-selection bias. This last one is the concern. At that low response rate, are 9% responding statistically representative of the entire population? Or, are they the kind who love to talk on the phone, all the time, to anyone who will call them?

This National Journal poll was just released: http://www.nationaljournal.com/daily/obama-romney-tied-among-likely-voters-20121002

Among registered voters, Obama wins by 5, and among likely voters, they are tied at 47 (this is why I think Gallup will move into a near tie when they switch to likely voters). But, of more interest, they talk about their electorate demographics:

In estimating the turnout on Nov. 6, the poll projects an electorate that is 74 percent white, 11 percent African-American, and 8 percent Latino. The likely-voter party splits are 36 percent Democratic, 29 percent Republican, and 30 percent independent. The estimates are similar to the 2008 turnout, when, according to CNN exit polling, 74 percent of voters were white, 13 percent black, and 9 percent Latino, with Democratic turnout at 39 percent, Republicans at 32 percent, and independents at 29 percent.

Though the report says the demographics are similar, reducing African-Americans by 2% essentially shaves 2% off Obama’s total, though since non-whites obviously replace them, I’m not sure it matters that much. I still don’t believe this electorate will be D+7, though. I figure D+3.

@DonS

Yes, that’s what I thought I was trying to point out about sampling sizes and response rates though I did not go into the depth you did in self-selection, etc. It’s similar to why studies of small groups of patients are discounted or rejected in medical studies. It’s not large enough to be representative of the population of patients. etc.

With Pew’s 9% response rate (eg: 900) of a small sampling and then weighting them – it inherently invites much more room for errors in weighting and larger deviations. And why it is less reliable or discounted when compared to Gallup or Rassmusen’s much larger sampling numbers with corresponding much larger response numbers (eg: 9000). It would be 9% = 900 responses vs. 9% = 9000 responses.

Hence the link in comment #158. It was an example of a bad poll and lame excuses being offered for gross disregard of industry standards/practices. 161 responses from 8 different states? And they offer those highly questionable results? Yikes. It would be better to trash the poll than be caught and discredited as they were. It’s a useless poll. I can only imagine the smoke, fire, and brimstone coming from Pat Caddell when he read that poll.

Susan @ 163: But it’s not that the low response rate causes smaller sampling sizes. It’s that the low response rate requires a lot more calls to maintain statistically valid sampling sizes. The reported polls still sample about the same number of people — they’re just having to make a lot more calls. And dealing with the concern that the people willing to talk with them are not representative — normal people no longer participate in polls. This problem affects all pollsters, including Rasmussen and Gallup.

Yes, the Washington Post-ABC News poll you referenced was a reporting travesty. The poll itself was legitimate, and showed a national race within the margin of error, with a reasonable sample size. But, for some reason, they chose to extract a subsample from that national poll to measure the race just in select swing states. The Post reported the subsample as the big story, rather than the main poll, above the fold on page A1! It took a conservative columnist to uncover the chicanery involved, that the subsample was only 161 respondents, with an 8% margin of error and an 80% confidence level, instead of the normal 95%. Pure garbage! The reason for the low confidence level was that the poll was not demographically designed to be regionally sliced like that.

The good news for Republicans is that polls almost always dramatically move in their direction in October as the pollsters scramble to tighten up their samples and voter screens so they don’t look like morons.

@DonS #162

In your assessment of the NJ poll, are you taking into account the reports that show that not only is voter registration down, but that voter registration for Democrats has fallen while voter registration for Republicans has risen? Polls are also showing the R ticket is higher among Indies than the D ticket. Just curious.

Susan @ 165: Democratic fervor was at a peak in 2008. They wanted Bush out and Obama was their savior. A lot of independents and Ron Paul Republicans voted for Obama as well, for a host of reasons. The electorate hadn’t been D+7 for a presidential election since at least the 1980’s, when there were still a lot of Southern Democrats who voted Republican for president. There is no way the current Democratic electorate is that fired up, but the Republican electorate is. That’s why we are seeing the dramatic turn in voter registration numbers and in early absentee ballot responses. So, yes, that is why there is no way I’m seeing D+7 this year.

@ DonS – thanks for your reply to my #165. Definitely not D+7. Are you still thinking D+3 (see your #162)? Seems generous?

I may be wrong, or we may be talking past one another, but it is my understanding that of the numerous polls, all have different numbers of respondents based upon different sampling sizes and do not all poll the same kinds of voters (eg: registered, likely, etc). The only two polls that have large #respondents and sampling sizes were Gallup and Rasmussen. If I understand correctly, the respondents they are counting are also likely voters. If so, those 2 polls would be the best to watch.

I’m certainly no expert in statistics and sampling, but I do remember from my university classes in stats how important a sampling size/#respondents in relation to the population you are studying have to be in order to be valid. A sample size of 900 respondents vs. 2 million voter turnout (eg: one state) does not invite my confidence. Especially with the problems you have already pointed out and the yet to be mentioned large number of people who only have cell phones instead of landlines. This election also seems to have a very different feel to it than others, in the sense of the extreme polarization and intimidation that seems prevalent in the public square. Add to that the way the polls are being manipulated by the pollsters, campaigns, and news media. For better or worse, the polls seem less than reliable this election cycle than I ever remember in the past and I don’t see much point in taking them seriously. Do you notice changes from past elections?

Certainly, I could be wrong, but I’m not convinced Obama has any real lead. If our nation is center right (as we’ve been told) and the silent majority turn out to vote, then I would guess that Romney would more likely win than Obama after the last four years of economic chaos and high deficits. And now that CNN appears to be sticking to it’s guns reporting the debacle in Benghazi and that forces the other MSM to report on it, Obama’s hubris may have met Nemesis.

DonS (@162), you’re doing that thing that several here have done and holding out one poll over the others. Why? Is it because, as it appears, that it’s a poll that shows a result you want to see?

Because, really, why are you touting the only recent poll in the Real Clear Politics collection that shows a tie?

That’s right, even Rasmussen is now showing Obama up by 1 point. That’s a shift of 3 percentage points in his favor since their previous poll 2 weeks ago. And, you know, Republicans seem to trust Rasmussen.

Rasmussen’s poll has a MoE of 3.0 points. Conversely, the National Journal poll has a 3.7 point MoE according to this page (N=1005), but RCP labels it as a fairly whopping 4.5-point MoE. Not sure if there’s some subsampling not listed in the previous link that’s used for the RCP analysis.

Meanwhile, obviously-not-liberal media outlets also list Obama with a lead larger than their MoE, such as Fox News (Obama +5) and Washington Times (Obama +9). Okay, even I don’t think Obama will win by anything close to those spreads, come November, but still.

Look at the RCP trend. Obama’s led for nearly the entire year, minus one very small blip. His numbers have climbed in recent weeks, even. The only reason the race has tightened at all recently is because Romney’s number took a decent jump very recently (fewer uncertain or third-party voters, I guess).

And yet, it would take a whole lot of work (or, you know, cherry-picking) to see any sort of Romney victory in the poll trends.

Susan (@163) said:

With Pew’s 9% response rate (eg: 900) of a small sampling and then weighting them – it inherently invites much more room for errors in weighting and larger deviations.

Which is funny, because the tied National Journal poll Don cites had only 1005 respondants. I guess that also invites room for error, huh?

You also said (@167):

The only two polls that have large #respondents and sampling sizes were Gallup and Rasmussen.

Sort of. Gallup’s N=3050 (only RV, though, so far), the largest by far (at least, on the RCP average), which is why their MoE is the lowest (2.0 points). Next largest would be Pew (N=2268 LV, MoE=2.4). Then Quinnipiac (N=1912 LV, MoE=2.2). And then Rasmussen (N=1500 LV, MoE=3.0).

Of course, Gallup (9/25-10/1) has Obama +6. Pew (9/12-9/16) last had Obama +8. Quinnipiac (9/25-30) has Obama +4. And now Rasmussen (9/29-10/1) has Obama +1.

Still think we should only trust the polls with large numbers of respondants?

didn’t the infamous dewey victory have a pretty big N?

not sure, but it’s problem was the majority of respondents (maybe all respondents) had telephones….and they tended to be republican, ha!

so a big N size is no guarantee if the sampling “strategy” (if that’s what you call it….you know, how many of this group, how many of that group, etc. and how well your statisticians or the person putting together the makeup of the random sample distribution and how accurately that reflects the “demographics” of the actual voters) seems to be at least equally critical, if not more so?

the calculation and mathematical theory is the science and the design of the sample population is the “art” of these polls…and like other art, the beauty is in the eye of the beholder…or rather, political hack 🙂

cheers!

Susan @ 167: “Are you still thinking D+3 (see your #162)? Seems generous?” — Except in extraordinary circumstances, nationally, Democrats tend to outnumber Republicans. In 2004, exit polls showed equality between the numbers of R’s and D’s, which was pretty extraordinary. Conservatively speaking, it is prudent to assume that the electorate will include more D’s than R’s. D+3 is a pretty typical electorate in recent presidential elections. But, we shall see. The hard thing about polling is figuring out who is going to show up at the polls, because the screening questions, no matter how well crafted, always engender more “yes” answers than the number actually voting. It is embarrassing to admit that you are not likely to vote.

Typical “snapshot” national polls, like the National Journal poll that came out today, have samples in the order of 1,000 – 1,500 or so adults, so that their likely voter subsample is in the 700-1000 range. If the sample is properly constructed, that is a sufficient number to result in a 4% or so error margin with a 95% confidence level. Rasmussen and Gallup are both tracking polls, with new nightly surveys being added and the oldest one falling off each day. Rasmussen polls about 1,000 per day, so their total sample size is 3,000. Gallup rolls its tracking poll over 7 days, so the sample size is relatively large. Consequently, their error margins are much smaller. I think they are the best to watch, generally speaking, and of course they both have very good track records.

I think pollsters are having difficulty modeling the turnout for this election. Intensity is high, which makes things less typical. The 2010 midterms were obviously a 180 from the 2008 election, and so the question is whether 2012 will represent a turn back toward the Democrats or continue the reaction to Democratic overreaching during Obama’s term. Cell phone polling is new and unproven. The record low response rates are also an issue. It remains to be seen how well the pollsters are doing this cycle, but there are definitely good reasons not to take the course of many in the media in declaring the race to be over 😉

tODD @ 168: Please read my prior post again. Do you know why I cited the National Journal poll? It wasn’t for its results — there are plenty of polls (almost all of the recent national ones polling likely voters) showing the race to be within the margin of error. Hint — read the part about the demographics that I specifically quoted, referencing the 2008 exit polls. Ring a bell in connection with our prior discussion? That’s why I mentioned it — not to “cherrypick” polls.

By the way, you can definitely throw out that Washington Times poll, tODD @168. John Zogby has been almost entirely discredited in the polling community. Why this particular newspaper is still paying him to do polling is a mystery.

@tODD

Yes and yes. And, yes on size, but one has to look at the methodology too. Both matter. The RCP average doesn’t distinguish the quality of methodology or differences in the standard of deviation when it averages the polls. And, no, we don’t trust polls. They give a snapshot of a group of people at a particular time based upon the answers these people give and the differences in the way questions are worded matters. Polls can be thrown out because of the way questions are worded. Everything in a poll matters and too many of the snapshots are poor to mediocre at best.

@dust

You are right. Methology matters greatly. So does size. It’s my understanding and as you pointed out, it’s called an art because of the difficulties in putting all the variables together so that it gives an accurate reflection of the population being studied.

@DonS

Thanks. I didn’t know there were that many more registered Ds vs Rs. You made good and helpful points. It has been my understanding that the large increase in the number of Independents has changed the dynamics and made it harder to predict voting patterns, too. It is sad that some progressives in the MSM are proposing that the race is over. It’s dishonest partisanship and unprofessional. Much can change between now and election day.

Some more polling perspective from Jay Cost: http://www.weeklystandard.com/blogs/morning-jay-race-has-just-begun_653404.html

Again, the message is do not wring your hands at every poll, good or bad. There is a lot of time, and most of Romney’s money, still to go.

@DonS

Good article. I hope he and you are right. I am concerned about Obama’s early campaign expenditures where he was successful in setting false narratives (eg: war on women). The heavy early expenditures seems to be a unique factor in this election as compared to past elections. My question/concern is whether, this late in the game, Romney can expose the tall tales and shake people loose from the fictions that O has tried to set into stone. Or is that moot for the October model?

More on how Gallup constructed its likely voter models in 2008 here: http://www.gallup.com/poll/111268/how-gallups-likely-voter-models-work.aspx

Note that, for the so-called “traditional” likely voter model they use a combination of a series of questions and estimates based on prior election turnouts. In 2008, they also reported “expanded” likely voter results, which only relied on the questions without modeling, because they were concerned about reported extraordinary voter interest that year. Their “traditional” model predicted an 11 point Obama win, while their “expanded” model predicted a 13 point win. Obama won by 7.

I predict no “expanded” Gallup model this year, when they start reporting likely voters.

my take on polls now is there are 2 very different varieties:

1. there are the ones for public consumption.

those are the ones in the news. they are made by a host of firms, and the results are all over the place. very convenient for a candidates “propaganda” purpose or spin machine. pick the poll that tells the “narrative” you want out on the street and voila, you can point to something with some credibility (at least in the minds of John and Jane Doe) to back your case, and it comes with built in “plausible denia-ability” of any poll manipulation on the part of your campaign. Hey, you’re just reporting the results from an independent polling firm, yahoo! In these cases, everybody except John and Jane, understands the game and plays along with a wink and a nod from our media to boot…probably generating some polls themselves that fall in this category and profiteering from them as well?

2. the second are the real polls.

these ones come closet to telling the truth about the political landscape. these are the ones the candidates use to make critical money decisions about where to visit, what ads to run, what to say in stump speeches, etc. Always, just follow the money for the real story. it’s just my hunch, but you know they have to be paying big bucks for polls with accurate info on the real “pulse” of the electorate. why? well, they don’t have infinite funds or infinite time or infinite other resources so good, highly analytic and accurate polls are worth their weight in gold? they can provide clues to what direction to go to minimize their weaknesses and maximize their strengths. but the public is never going to see them…they are confidential and for the eyes of the candidate and their staff only. sort of like that wizard behind the curtain?

well, that’s my story and am sticking to it 🙂

cheers!