Today's Word Is "Outlier."

Here's to you, dumb-ass reporters.

Yeah, that's right.

Today's dumbassery comes to us thanks to the USA Today/Gallup poll, and more specifically the coverage thereof.  In a stunning twist, so the poll says, John McCain now has a 54-44% lead over Barack Obama.  That ten-point lead is among those defined by Gallup as "likely" voters.  Among registered voters, they claim McCain leads Obama 50-46%.

Wow!

Wait.  Hold on, dumb-ass reporters.  Let's not go sucking McCain's old dick just yet.

You see, even with the best sampling frames, even if you could achieve the Holy Grail of sampling (Simple Random Sampling, said with heads reverently bowed), then, if you have a 3% margin of error, as Gallup says they do, then you'd still get a fucked-up result, on average and with respect to true population values, in three polls out of every one hundred taken.  See?  That's because of something called random sampling error.  A sample, of course, doesn't measure what everyone in the population thinks.  It can only estimate those unknown population parameters.  Done correctly, samples are a very powerful way of estimating population parameters from a (relatively) few number of people.  No, really, it's quite amazing.  However, because of the way combinations work, you will get some inaccurate sample results, just due to random sampling error.  If you took every possible sample of size 1,022 out of the population of millions of registered voters (which would be impossible--there are way more combinations there than there are protons and neutrons in the universe), eventually, you'd get a sample of 1,022 people who are all Republicans.  It's unlikely, but it's bound to happen.  Even more unlikely, but still bound to happen, you could get a sample of 1,022 people who were all named "McCain."  Assuming, of course, that there are at least 1,022 registered voters who have that last name.  The point is, given enough fucking iterations, you would eventually draw every sample that is possible.  Most of the time--indeed, the vast majority of the time, those samples will have statistics that are pretty close to the population parameters they're esitmating; that is, they will be representative samples.  But every once in a while--and you have no fucking idea when this will happen--you'll get a nonrepresentative sample.  Oops.  When that happens, the polls are considered outliers.  But you can only make that determination, for an SRS, if you have lots of other polling data.

Head hurting yet? 

Just bear with me.

But Gallup, nor any other nationwide polling firm, never conducts simple random samples (it's not that they wouldn't want to--it's just that it's impossible).  Instead, they construct sampling frames.  And this article doesn't tell us just how they constructed that frame.  Did they include more Republicans?  How did they weight or oversample different groups?  Without that information, it's hard to accurately assess whether or not this poll is simply an outlier, or if the sampling frame itself is fucked.  Or if there's some combination of those two things going on.

However, what we can do is look at this:

Hmmm.  Notice anything?

So, yeah.  I think we might have an outlier here.

Not, of course, that any of the news outlets are reporting that.

Oh, fuck no.  It's a post-convention/Sarah Palin bounce!

You fucks.  You don't get a seventeen-percentage-point swing (just in the Gallup poll) in a week.  Not without something extraordinary happening--and I'm not talking about that coma-inducing GOP convention, either.  How about you, I don't know, look for a little context for your goddamn stories?  Or would that be biased?

And you people thought all I was good for was swearing and pretty pictures. 

Comments

[BuggyQ]

Jude, I get what you're saying (probably better than most, being an amateur statistician in a professional statistician's job), but they don't.

But even if they did, they don't want to. They want to be able to say McCain got a convention bounce. They want it sooooooo bad.

I was talking to my boss yesterday about all this, and she was still trying to figure out why people like Cokie Roberts--who of all people ought to understand that a Republican administration would be BADBADBAD for America--still seem to be cheerleading for John McCain. I admit, I have trouble with understanding it myself. I think it's 4 parts cognitive dissonance (I know he's bad, but he's a war hero maverick!), 3 parts the beltway-cocktail-weenie effect (OMG! The popular girl actually TALKED to me!), and 3 parts self-loathing (fully justified, if I may say so).

On the other hand, Jon Stewart seems to have gotten his head out of McCain's ass enough to point out the massive flip-flopping McCain has done since 2000. I wonder if Jon has finally gotten past his genuine fondness for McCain enough to see that the guy will say and do and fuck fucking anything to become president.

Finally, I apologize for my language of late--watching the RNC apparently shorted out something in my brain, and now the words "John McCain" and "Republican" automatically come out with "fucking" attached.

[Paddy]

A The Fuck Men. Plus, your swearing makes me tingle.

[Elspeth Ravenwind]

Ohhh, Jude!!! The intelligence, the passion, the swearing and the flippin da bird pic above - swoon!!! :)
Angry DFH Liberal pRon!!!
More!?
Elspeth

[Rocky]

You know why the media is making it a "McCain-Palin" convention bounce?

Because a close election will result in bigger ratings. No one (except for the die-hard team fans) wants to watch a landslide, whether it's the Super Bowl, the World Series or a presidential election.

So they have to stir things up. They have to influence the viewer's perceptions and make it more dramatic than it really is.

[kay osentowski]

When did electing the president of the United States come down to a gimmick and a ploy. When did winning at any price or cost to the American people become so tainted and so disparate that a man who has not offered one policy on how he is going to suddenly "change" things picks a small town would be zealot with no experience and shoves her in front of millions to suddenly believe? The only thing that is going to make people believe this is the media. They will not ask the probing questions. And when they do accidently ask a question that has substance, the party of gimmicks and ploys doesn't answer it. The interviewer never ever calls them on it. They just go to the next question. The media will make this ticket a believable one. Nothing McCain or Palin can do will make people believe in this farce. Only the media can make it happen. How incredibly sad.

i'm stealing that pic. i won't say why. tis a surprise.

[pansypoo]

poo. im anonomous again.

[Bud White]

reject the sexism used against Palin:

Dear friend,

I received your email with the attachment about Sarah Palin. While this report might be scary and threatening to some, to me it sounds like the same the kind of trash written about Hillary Clinton, Bill Clinton and even Barack Obama. However, since the Clintons are no longer in the race we can leave them out. But what I wonder is why it’s okay to say that Palin is a fundamentalist Christian but not okay to ask why Obama spent twenty years in the pews of Jeremiah Wright’s intolerant church? And why is it okay to allege that Palin wants to destroy Alaska’s wildlife but not okay to ask if Obama plans to support Hamas instead of Israel? And people can can accuse Palin of racism based on rumor and hearsay but not hold Obama accountable for playing the race card, quoting Malcolm X and accusing good people, particularly the Clintons, of racism. While I may not agree with Palin’s politics, I am unmoved by the smear campaign launched against her. Please do not send me any more political emails. While this e-mail goes to great lengths to appear to be politically correct and to be doing me a favor, it’s actually savaging someone’s reputation based on shoddy evidence or outright lies. After forty years of voting for the Democratic candidate for president, I want you to know that these sexist and hateful lies against Sarah Palin you forwarded go against the reason I am a Democrat. Therefore, under no circumstance will I be voting for Obama.

Best Regards,

[PaminBB]

I think the fact that only _2%_ are undecided in the USA/Gallup poll indicates that it must be flowed. No way are there only 2% undecided.

[Elspeth Ravenwind]

Um, dearest moronic 'concern troll' Bud Light -
If you want to restrict female Americans from having the right to choose to carry out a pregnancy or not, or if you want to criminalize birth control methods, you just stay in your mom's basement in your rag-tag 1980's spiderman underoos and rancid cheetohs dust w/the ruined Victoria's Secret catalog. kthxbai!
Elspeth

[Mike]

Monday, September 08
Race Poll Results Spread
Virginia SurveyUSA McCain 49, Obama 47 McCain +2
Michigan PPP (D) Obama 47, McCain 46 Obama +1
National CNN McCain 48, Obama 48 Tie
National USA Today/Gallup McCain 54, Obama 44 McCain +10
National Rasmussen Tracking McCain 48, Obama 47 McCain +1
National Hotline/FD Tracking McCain 44, Obama 44 Tie
National Gallup Tracking McCain 49, Obama 44 McCain +5
National CNN Obama 45, McCain 45, Nader 3, Barr 3 Tie

http://www.realclearpolitics.com/epolls/2008/latestpolls/index.html

So, where is Obama leading? It looks like Democrats have managed to snatch defeat from the jaws of victory with a loser candidate once again.

[ServingPatriot]

Jude,

yes... you made my head hurt. And I r an enguneer...

nevertheless, I think the polls are really all in the crapper this year. Waaaayyyy to many un-modelled unknowns (e.g., first black prez, first woman VP, really old dude, new voters,etc, etc, etc). No way these polling folks are capturing all of this in their processes; to do so would require an election to model and we've never had one like this one.

As you so eloquently observe, this will not stop the MSM chasing the poll baubles because close race = horse race = ratings = ad $$$. And in the end, isn't that all it is about for the MSM and their respective tv, radio or print Ad Delivery Vehicles?

So, I say simply FYYFFs! and go back to work trying to GOTV for the Dems. The race is by no means beyond stealable and we're gonna need everyone this year.

SP

PS - and for Pamin... yep, I DO think that 2% undecided is about right... at least in the current polling methods... been that way for at least 2 election cycles and will be that way until more millenial/Yers make it into the voting ranks. :)

[eRobin]

SP is right - keep your heads down and work on GOTV for Obama and the local Dems. You can find your local campaign office at Obama's site. If you live in a safely blue state, find out how to do remote phone banking. Stop reading polls.

[mdh]

For a second there I thought I was reading Sadly, No!, and I think that's a compliment.

"We asked 1,022 residents of Darian, CT....."

[Daniel Z.]

First and foremost, what we need to remember is that you don't need to win the popular vote to win the Presidency. I highly expect Barack Obama to lose red states by larger margins than normal Democrats lose them by. I also expect him to squeek by in several of the swing states. It is highly possible for Obama to lose the popular vote by millions and still win the election.

Seeing how intelligent Obama was in campaigning for the nomination, it is safe to assume that he will be campaigning intelligently in the general election as well.

And to those Hillary who want to vote Palin to "make history", if they win you may very well find that the rights that Hillary would stand up for will BE HISTORY after McCain nominates 1-4 Supreme Court justices.