Friday, September 21, 2012

Is polling worthless?

It's counter intuitive that Barack Obama isn't behind by 30% or more.

His abysmal record of lies, distortions, broken campaign promises, indifference, cluelessness... the disaster he's been to this country's foreign and domestic policies... he should be getting crushed.

But none of the polls indicate that.  So, we're told that Obama is "ahead in all the swing states" and that Romney's campaign has been "off the rails."

Bull.

This has been arguably the worst week ever for Obama.

He and his people lied through their tooth about the situation in North Africa.

He refused to admit to Hispanics that he lied to get their support for his election.

The economy continued to head south.

More of us were murdered in the Middle East... and this clown continues to lead?

Not.
Possible.

And that's because I'm thinking, more and more, that it's not true.

Take here in Washington State.

On August 1, Survey USA released a poll that indicated Inslee was out-polling RINO Rob in this state by 3.  I even wrote about it.

But when you look at the cross tabs of the poll, it turns out that democrats were over sampled.... by 9 percent.

Party Affiliation

Republic Democrat Independ

 ...27% .........36% .......35%


If you oversample democrats by 9%... how likely is it that the democrat will lead the poll by 3 or more?  And isn't a better question that even with a 9% over-sampling, why doesn't Inslee's have an even BIGGER lead in this poll?

This seems to be a pattern repeated around the country.  This is a microcosm of just one poll... but it seems to be indicative of polls around the country.

graph2
Democratic sampling advantages in recent polls.
If it is... then what is the value of a poll that doesn't accurately reflect the electorate?
Blogger keithbacker at Battleground Watch sums up the graph this way:
The real take-away which I have mentioned the times I blog national polls is that many of those national polls are HORRIBLE for Obama, namely the ABC/Washington Post and CBS/New York Times polls where you have large Democrat over-samplings but rather small leads for Obama. This means if Obama doesn’t meet or beat his stellar 2008 turnout advantage he’s in for a drubbing on election day.

These over-samplings serve a few purposes but mainly drive down enthusiasm for Republicans while assisting the Obama campaign with “bandwagon” supporters who simply like being on the winning team (they’re real and they count).
If they have a value, it's that we can draw inferences from the lack of a massive lead given a massive oversampling.


I've learned a lesson here: when polls go up, dig deeper than the surface numbers to make a conclusion as to what they mean.

5 comments:

Martin Hash said...

You are the most sophisticated politico I know so I have to assume you're following some Conservative "talking points" on this issue because, dude, Obama's got this thing going away. Every election, the powers-that-be make the race look close, probably for selfish reasons, but Romney has no more chance than McCain had 4 years ago.

I support Obama but I'm not making this post based on partisanship - it's just the way elections go.

K.J. Hinton said...

You assume wrong.

I do not EVER do "talking points," and it's offensive that, after all these years of us going back and forth that you'd believe that I don't absolutely believe EVERYTHING I write.

No one pays me to do this blog. I have hammered those in the GOP I oppose.

I pro-offer an explanation for the inexplicable, and your response, instead of disputing the facts I've presented devolves into partisan eye covering and exclamation that it's just wrong...

Not HOW it's wrong. Not that my facts are wrong, you understand; but that the facts are meaningless to my conclusion, a position you took as part of a presentation where, for example , you would lead us to believe that oversampling in polling means nothing.

All the way back to Truman-Dewey, the first presidential polling done by telephone, the damage of oversampling is clear.

We have bone fide cases of it here, and we're supposed to ignore that as a factor in making conclusions?

You claim you're not making this post out of partisanship, but then fail to provide any proof of anything you write. As an attorney, if I'm not mistaken, you know that kind of thing wouldn't fly in court, and it doesn't fly here.

I started my post by showing where the president has been a complete disaster. This guy makes me long for the Carter era.

And when presidents are a complete disaster, the other side can run a barn-headed ape and win. By every measure of every past presidential election, this guy (Obama) has no chance. That he is even to ahead in most polls with large democrat oversampling cannot be a factor that anyone, including you, can just summarily toss out the window.

In fact, martin, I ask you: what would the polls shw with a GOP oversampling of the same size? Do you REALLY think the outcomes they show would be the same?

You're a better person than that.

I have never once looked at a political situation and made a claim or a prediction based on the "R" after someone's name. And I am not doing so now.

It appears to me that your response is based entirely on what you hope to be true... and not what the facts tell us. And further, your support of someone is certainly NOT just how elections go.

My analysis is based on what I believe to be true. No talking points and no partisanship determines my conclusions. I don't respond to fringe-left idiocy like Brancaccio's annual hate-Benton column today with anything but facts and what I know... Lou's bogus claims of non-partisanship notwithstanding.

I'm certainly willing to learn and adjust my conclusions. By all means, feel free to provide me with a viable alternative involving math and statistics to show me where I'm wrong.

Jack said...

Martin's gonna be one unhappy camper when Zero Obozo gets thrown out on his incompetent butt.Martin will be in shock for months,maybe even years.

Martin Hash said...

Okay, you convinced me - that post was all yours... But, guy, I'm not on the boat. Romney looks like a loser for a lot of reasons, and Obama has been almost mythical. Even though I usually abide by your experience on local issues, I'm thinking I may have a better read nationally. No insult intended. We'll just wait and see how it turns out - that's what democracy is all about.

Jack said...

"Romney looks like loser" and after 3 1/2 years Zero Obozo "doesn't "look like a loser"??Whatizzit that you're smoking lately, Martin? It must be something really hallucenogenic.The nation is in an abdolute mess and the moron at the wheel is none other than Zero Obozo. so you're saying that Romney gets the blame? you're just consumed by your own "wishful thinking", that's all.