It’s Never What You Think

Silly peasants. There’s no such thing as “Woke!”:

[Ipsos] latest poll confirms that no one really knows what “woke” means, particularly when respondents aren’t provided with any accurate definitions. “Republican presidential hopefuls are vowing to wage a war on ‘woke,’” USA Today’s write-up of its survey began, “but a new USA TODAY/Ipsos Poll finds a majority of Americans are inclined to see the word as a positive attribute, not a negative one.”

Or if it does exist, it’s an unalloyed good thing!

Hear that, Republicans? Everyone loves “woke”! Well, at least 56 percent of those surveyed endorse the word when they’re told it describes someone who is “informed, educated on, and aware of social injustices.” By contrast, just 39 percent of respondents express a negative view of the word insofar as it describes someone who is “overly politically correct” and is inclined to “police others’ words.” Having deemed the Right’s obsession with “wokeness” a quixotic endeavor, USA Today bellyflops into a solipsistic reflection on how the public’s perceptions “raise questions” about the Right’s self-defeating myopia.

Of course, this is an example of two institutions that should be distrusted and veriied – media:

This is a prime example of journalism that works backward from a conclusion in pursuit of evidence to support it.

and the polling industry:

A quick perusal of the polling on the issue exposes the flaw in USA Today/Ipsos’s methods.

When respondents are not primed with erroneous definitions and are instead asked only if they would vote for a self-described “woke” candidate, as CBS/YouGov did last October, they found that 58 percent of likely voters would be less likely to pull the lever for that candidate. That same month, a Harvard-Harris poll found that 64 percent of respondents, including a majority of Democrats, blame “the increase in crime” on “woke politicians” as opposed to “other factors.” That’s, at the very least, odd if most Americans don’t understand the word or believe it only describes a heightened social consciousness.

…that should be distrusted but verified, and then almost to a mathematical certainty distrusted some more.

For those who need a primer:

Wokeness in practice is not something so quaint as speech-policing and “political correctness.” It encapsulates an alternative theory of social organization that often enters into conflict with the Constitution. It prescribes not just otherworldly speech codes but programs of reeducation for those who decline to subscribe to them. It necessitates the redistribution of economic and social goods in the pursuit of “restorative justice” for wrongs committed by generations long passed. It redefines cosmic constants like the laws of mathematics, operating on the bigoted assumption that those laws are incomprehensible to those who were born into certain identities. “Woke” does not describe a persnickety busybody who cannot abide your verbal miscues. It describes a revolutionary.

But, like the “Don’t Say Gay” bill that isnt, the “Republican War on Women” that never was, the “Threat to Abortion” that never existed in Minnesota, the “De Santis Book Ban” that isn’t and the “Blogger Registration” law that never will be, the new attempt to gaslight the right on the existence of “woke” (and critical race theory), the intended audience isn’t people who can reason or think critically.

It’s aimed at Democrat voters.

Mene Mene Tekel Polling-in

While it’s always worth casting a cold eye on polling numbers, it’s striking how poorly the Leader of the Free World is doing these days across the board. Real Clear Politics has the numbers and they aren’t pretty:

RCP Average                            12/17 – 1/11 —           41.9               53.1               -11.2
Quinnipiac                                 1/7 – 1/10 1178 RV    35                  54                  -19
Politico/Morning Consult           1/8 – 1/9 2000 RV      44                  53                  -9
Rasmussen Reports                 1/9 – 1/11 1500 LV     40                  58                  -18
Economist/YouGov                   1/8 – 1/11 1258 RV     45                 51                   -6
IBD/TIPP                                   1/5 – 1/8 1308 A         44                 45                   -1
Reuters/Ipsos                            1/5 – 1/6 1000 A         45                 51                   -6
Economist/YouGov                    1/2 – 1/4 1201 RV      43                 51                   -8

It’s one thing if Rasmussen, a pollster generally favorable to the GOP, posts ugly numbers for Joe Biden. What’s truly eye-opening are the results from Quinnipiac, a poll historically gentle with the portsiders. If their numbers are accurate, Biden’s disapproval rating is a whopping -19, and that’s from a poll conducted after the recent Pelosi-produced passion play of 1/6.

So what are the larger implications? A few thoughts:

  • If you’re Joe Manchin or Kyrsten Sinema, do you fear the political wrath of the leader of your party? Apostasy has rarely been safer.
  • If you’re Xi or Putin, do you concern yourself with anything Joe Biden says? Or, for that matter, with anything he does? I doubt the citizenry of Ukraine or Taiwan is particularly sanguine at the moment.
  • If you’re Mitch McConnell, do you sleep well? Never better, I would guess, and certainly better than your average Ukranian.
  • If you’re Stacey Abrams, do you see any value in hanging with Grandpa Joe?

While there’s certainly entertainment value in watching Uncle Joe moving kinda slow, we’re not headed for Petticoat Junction at the moment.

Motivations

So the pre-election polling saying the electorate was going toward Biden in landslide lots was wrong?

Who could have possibly figured that out?

Oh, yeah – anyone that’s been reading this blog for the past ten years.

There are three possible explanations:

1. Evolution! – The pollster’s craft hasn’t caught up with the “new normal”, in a society where people legitimately fear being “canceled”, losing jobs, social standing and being targeted for violence because of their beliefs.

That is simultaneously possible, and not mutually eclusive

2. Incompetence! – The pollsters absorbed the lessons of 2016, where they actually did a little better than they did this year…

…and learned nothing.

3. Never Ascribe to Incompetence What Can Be Chalked Up To Malice – I’m going to present three facts and a conjecture:

Fact 1 – On December 1, 2016, representatives of the New York Times and Washington Post newsrooms went on WNYC radio’s “On the Media“ program (syndicated on NPR) and said, In as many words, that was time to change the rules of journalism. It was time to move past “passing the facts on to people and letting them make up their own minds” to “Denormalizing Donald Trump“.

Fact 2 – in 1986, a UCLA psychology professor, Dr. Mehrabian, showed the existence of a “bandwagon effect“; when polls showed that a candidate had no chance of winning, “swing“ voters tended to stay home or vote for someone else.

Fact 3 – for the past 30 years, the Star Tribune “Minnesota Poll” has had a fairly clear pattern; the closer a race ended up being, the more wildly distorted pre-election polling numbers were. For example, they showed Tim Pawlenty, Norm Coleman and Tom Emmer getting blown out just before the election. All three races ended up being famously close. On the other hand, they tend to report blowouts pretty accurately; they had Amy Klobuchar and Kurt Bills pretty much dead on.

Conjecture: It’s not an “accident”, or a learning error, that polling predicting a landslide up until election day was completely wrong.

Thoughts?

Praetorian Guards Gonna Guard Praetorially

Twin Cities media are giving wall-to-wall coverage to a poll showing thunderous agreement with the DFL administration shutting down most commerce in the state, indefinitely, at the whim of any “science” they officially sanction.

But – does anhone see any problems with this poll?

It’s from Public Policy Polling, And it was paid for by…

…Wait for it…

… Our old friends at the Alliance for a Better Minnesota!

So of course it oversamples Democrats and women.

And of course the Twin Cities media runs it like it’s evidence of a mandate.

The Twin Cities Media Won’t Cover This…

…I don’t suspect – but I will:

The governor’s race is inside “Statistical Noise” levels.   Ditto Housley and Smith.   (Klobuchar is shown nine up over Newberger, which is closer than other polls as well).

Last week, an NBC/Marist poll claimed Walz and Smith were pulling away – this being the same poll that showed Tim Pawlenty and Lori Swanson winning the gubernatorial primaries in landslides.   Was that a media/Democrat attempt to “bandwagon” Republicans into staying home?

Maybe, maybe not.

But don’t get bandwagoned anyway.

The Bandwagon

NBC/Marist poll shows DFL Senate Candidates leading by biblical margins in the upcoming election.

The glum picture for the GOP comes as both parties point to the state’s four competitive House races as bellwether contests in the race for control of the lower chamber. The poll finds that 53 percent of likely voters prefer a Congress controlled by Democrats after the November midterm elections, while 41 percent prefer Republicans.

Unmentioned: this same showed Lori Swanson winning the DFL primary fairly easily, and Tim Pawlenty clobbering Jeff Johnson. They were off 28 points in that matchup.

But then the point of these October polls isn’t accuracy. It’s “bandwagoning“.

What Is Best In Life?

In the TV series MASH, there was an episode featuring a statistician – an Army officer who predicted how many men would be killed or wounded given the parameters of an upcoming battle.    To the statistician character, it was all about numbers – “just business, nothing personal”, to invoke a line from a different seventies production.  To surgeon Hawkeye Pierce, the character who had to try to patch together the actual men behind the numbers, is was in fact personal.

At the end of the episode, losing his temper at the statistician, after showing the geek through the operating room, Pierce yells “the thing I hate about you isn’t that you’re good at your job.  I hate you for liking it so much”.

I have a similar reaction to people who try to boil all human behavior down into numbers, statistics and analytical models.

If blogs existed 50-60 years ago, a story like this would be accompanied by a photo like this. Good thing this is 2016, right?

Now, before you launch into some misguided jape about conservatives hating science, remember – part of my day job is, well, boiling down human behavior into numbers, stats and patterns.  A bigger part, at least for me, is finding the qualitative answer behind the numbers.

But I digress.  Among the many joys of this past election – the potential for a safe SCOTUS, a solid cabinet, no Hillary, no leasing of US foreign policy to the Saudis and Qataris – was the complete collapse of analytics in predicting (and, via our media, shaping) this past election.

The ana­lyt­ic­al mod­els for both sides poin­ted to a Clin­ton vic­tory, al­beit not a run­away. The Clin­ton cam­paign and su­per PACs had sev­er­al of the most highly re­garded polling firms in the Demo­crat­ic Party, yet in the places that ended up mat­ter­ing, very little if any polling was done. So while 2016 wasn’t a vic­tory for tra­di­tion­al polling, it cer­tainly took a lot of the luster from ana­lyt­ics. In the end, big data mattered very little.

While tinkering with stats can be fun, I’ve long loathed notion that all of human behavior can be boiled down into numbers.   And I’ll admit, the schadenfreud when the geeks fail to do so is glorious.

Shot In The Dark: Today’s Corruption News, Six Years Ago

In the aftermath of the 2010 election, I noted that the Star/Tribune “Minnesota” and Humphrey Institute polls were consistently, statistically, not only erroneous, but in a very suspiciously consistent way; in their polling, especially their election-eve polls, they always showed Republicans doing much much worse than they ended up doing – and this correlation was even stronger in races that ended up being close.

I also pointed out actual research indicating that a “bandwagon effect” had been identified in political polling; that negative polling about one’s candidate tended to make that candidate’s supporters stay home from the polls.

At the time, I noted that it was possible the media – operating in their capacity as Democrat operatives with bylines – might not be doing it on purpose to drive down Republican turnout in close elections – but if they were, it’s hard to think of what they’d be doing differently.

I needn’t have hedged; when I suspect the media of some pro-Democrat perfidy, I’m rarely disappointed.

John Podesta’s emails, hacked by Wikileaks, show that the Democrats, working through their network of sympathetic pundits, journalists and pollsters, have been doing exactly what I suspected they were;  getting pollsters to jiggle the samping to underpoll Republicans and overpoll Democrats.

“Suspicion of Democrat perfidy is all but certainly proof, and is almost always correct”.  It might be a Berg’s Law soon.

 

Survey

SCENE:  Mitch BERG is at work in his home office.  His phone rings.  

BERG:  Hello?

POLLSTER:  Hello.  I’m Kandi, a pollster working on a combined study commissioned by Harvard University, Northeastern University, the Trace and the Guardian, four organizations dedicated to disarming Americans by any means, fair or foul.  If you have a few moments to spare, I’d like to ask you some questions about gun ownership.

BERG:  Go ahead.

POLLSTER:   How many guns do you own?

BERG:  How many guns am I going to admit I own to an anonymyous rep for  four organizations that are dedicated to ensuring that Americans are disarmed, docile sheep,?

POLLSTER:   That’s correct!

BERG:  None!  Guns are scary!

POLLSTER:  So that’s no guns, then?

BERG:  As far as you know.

POLLSTER:  Wow. It’s amazing how the number of gun owners is dropping, according to our Fact-Based Research ®.

BERG:  It is, isn’t it?  Absolutely astounding.

SURVEY:  We’re also finding three percent of American adults own 50% of the guns!

BERG:  Huh.  I’m also gonna guess 3% of American adults own 50% of the iPhone 7s, and roughly .000001 of all Americans own 90% of all newspapers.

SURVEY:  No comment!

BERG:  Naturally.

SURVEY:  Now, if you did buy a gun, why would you buy one?  Are you a hunter, a target shooter, or would you buy a gun due to fear?

BERG:  If I did have a gun, which I don’t, as I already told you, it’d be for self-defense.

POLLSTER:  OK.  “Fear”.

BERG:  No, self-defense.  A prudent response to the vicissitudes of human nature.

POLLSTER:  Right.  Fear.

BERG:  Nope.  A rational, prudent assessment of and response to life’s actual risks, based on data, ability and experience.

POLLSTER:  Right.  We call that “fear”.  It’s just a category.

BERG:  Naturally.  Hey, someone’s calling…

POLLSTER:  I didn’t hear a click…

(But BERG has already hung up the phone).

And SCENE.

 

Empirical Data

Joe Doakes from Como Park emails:

6,500 Minnesota State Fair goers answered a survey.  They demand:

 Higher gas taxes

Higher sales taxes

More gun control

More college subsidies

More sick leave

Right-to-Die

Student privacy

 They hate:

 Restrooms designated by sex

Talking on cell phones

Non-partisan elections

Legislators setting their own salaries

 There are 5 million Minnesotans who did NOT vote in this survey.  Do you think their wishes will be considered?  Or is the fix already in?

 Joe Doakes

With Minnesota bureaucrats, the fix is always in.

Rescued From Pre-Thanksgiving Doc Dump

It’s only the SCSU Poll – a poll we’ve pretty well shredded in the past for its systematic bias toward the DFL.

But even that can’t varnish the fact that it’s a whole new campaign for the DFL in Minnesota.  Approval ratings of everyone but Amy Klobuchar are in the toilet.  How in the toilet?:

Minnesotans came down in the middle on Franken and Dayton, two Democrats who will stand for re-election next year. Franken got a 51-degree rating, while Dayton was at 49.

Respondents are feeling cooler toward Obama, rating him at 46 this year, down from 54 in SCSU’s 2012 survey.

And that’s just the warm-fuzzy poll.  When you get into job approval, it’s even dodgier for the DFL incumbents:

Reflecting national polls, the president’s job performance ratings also dropped from last year and returned to 2010 levels. This year, 38 percent of Minnesotans rated him positively, compared to 47 percent in 2012.

For Dayton, less than half the respondents (44 percent) gave him positive marks this year, while a slight majority (52 percent) rated him negatively.

Franken had a low approval score of 39 percent, while 57 percent approved of Klobuchar’s performance.

The media will, of course, do their best to rehabilitate the DFL, Dayton and Franken over the next 11 months.  But they’ve got their work cut out for them.

Dip

Governor Messinger’s Daytons’ approval ratings – the media says – have dipped.

Survey USA shows the Governor at a 47 percent approval rating:

Significantly, however, the September approval rating is down 9 percentage points since May of 2012. Back then, when Dayton was still grappling with a Republican controlled Legislature, the governor’s approval rating was 56 percent in the SurveyUSA poll.

The dip was particularly glaring among Democrats:

According to the SurveyUSA poll, Dayton went from an 80 percent approval rating among self-identified Democrats in May of last year and February of this year to 67 percent approval rating. The pollster said that segment of the poll had a plus or minus 6.7 percentage point margin of sampling error.

And naturally, the Strib poll – which has a long history of comical inaccuracy – differs: 

The Star Tribune’s polling over the past two years do not track with the SurveyUSA polling. The latest Minnesota Poll, which had a margin of sampling error ofplu sor minus 3.5 percentage points, was taken in June and found Dayton’s approval rating was 57 percent. That was up from 45 percent in February of this year and 53 percent in September of last year in Star Tribune polling numbers.

In the Strib’s poll, Democrats approved of Dayton overwhelmingly – which would appear to be a major difference in the two polls.

That, and SUSA having a track record of relative accuracy, or at least not being utterly comical.

Poll-arity

One outcome is certain tomorrow – the pollsters will finish last.

Give the pollsters of the 2012 cycle some credit – they’ve managed to straddle the fence, predicting a solid electoral victory for Barack Obama…and potentially a major popular vote win for Mitt Romney.

The top line of most of the recent polls has been easy enough to read.  The Real Clear Politics national average represents a statistical tie as Obama leads by 0.7% but the sheer numbers of polls showing slight edges to Obama in key states has the conventional wisdom pegging the President at somewhere around 290 to 303 electoral votes.  A step drop from 2008 but a large win by comparison to the recent histories of 2004 or 2000.

Yet the crosstabs of almost every pollster suggests a far different outcome as Mitt Romney holds a lead among unaffiliated/independent voters.  And the margins are anything but slight.  Romney leads independent voters by 7% with Fox News’ polling. By 9% with Rasmussen Reports.  12% according to two separate polls by NPR and the New York Times.  16% by Monmouth’s numbers.  And a jaw-dropping 24% by CNN.

The lead isn’t universal – Gallup has Obama up 1% among indies with Politico having a similar result…after deciding they would qualify more indies as Republicans following Romney’s 10% lead just two weeks earlier.  The trendline is obvious.  The question is how much does it matter to win independents?

Conventional wisdom in politics is like conventional wisdom about everything else – it’s right up until the point it’s wrong.  Whereas independent voters have been prized possessions in past elections, suddenly the value of these voters has been called into question:

It’s true that independents are a diverse group. But that’s mostly because the large majority of independents are independents in name only. Research by political scientists on the American electorate has consistently found that the large majority of self-identified independents are “closet partisans” who think and vote much like other partisans. Independent Democrats and independent Republicans have little in common. Moreover, independents with no party preference have a lower rate of turnout than those who lean toward a party and typically make up less than 10% of the electorate. Finally, independents don’t necessarily determine the outcomes of presidential elections; in fact, in all three closely contested presidential elections since 1972, the candidate backed by most independent voters lost.

Let’s look at that last statement in greater detail.

On the surface, it’s 100% correct.  Jerry Ford, John Kerry and George W. Bush all won the independent voter demographic and all three lost the popular vote (although not the election in all three cases).  Bush won indies by 2% and lost by 0.5% in an electorate that was 4% more Democrat than Republican.  Kerry won indies by 2% as well but lost by 3% in a tied partisan affiliation election.  And Ford, amidst a massive movement of Republicans to Independents post-Watergate, won that block by 4%…the largest margin for a losing candidate and done in an electorate with a 15% Democratic advantage.

The trendline here is simple as well – a narrow advantage among independent voters guarantees nothing other than perhaps a close election.  But compare Romney’s margin among indies to past performances.  Obama won indies by 7%.  Clinton won indies, despite an independent candidate on the ballot, by 8% in 1996 and 6% in 1992.  Bush Sr. won by 14% in 1988 and Reagan by 28% and 25% respectively in his two races.

Can Romney win independents and still lose the election?  Of course.  But only if a few other conditions arise.  The electorate has to be strongly Democrat.  Many pollsters are using D+8ish models ala 2008 even as 825,000 voters in eight key battleground states dropped their Democrat registration.  Or Romney could lose a key chunk of Republicans to offset his gains among indies.  That too seems unlikely as Democrats have held voter identification advantages every year since 1972 except in 2002 & 2004 – and the largest Republican advantage was 1% in ’02.

Some have argued that Romney’s lead among independents is simply a reflection of dissatisfied Republicans having left the party but whom will still vote conservatively.  It’s not a bad theory and it’s supported by some evidence.  Gallup has Republicans at 28% and Independents at 38%.  Pew has Republicans at 25% and Independents at 36%.  Yet neither Gallup or Pew reflect such a shift in their presidential polling.  Gallup has Obama up 1% among indies, as previously stated, and Pew has Romney up only 3%.  If Republicans just dropped the ‘R’ from their ID, someone forgot to tell them.

The end result isn’t actually about who wins on Tuesday.  Regardless of the  outcome, most of the pollsters have made a series of startling errors.  Either they’ve completely whiffed on properly defining party IDs within whatever likely voter model they’re using or they can’t accurately identify independent voters as a demographic.  Simply put, the numbers don’t match.  Obama can’t win if he loses the largest party ID block by high single or low double digits.  Conversely, Romney can’t lose if he wins independents by those kinds of margins.

The question in doubt tomorrow isn’t whether the pollsters erred but on which end of the spectrum.  We’ll find out for sure on Tuesday.  The pollsters will have to find out how they went wrong starting on Wednesday.

ADDENDUM:  Over at Mr. Dilettante’s, D pithily surmises the conundrum of the 2012 polls:

One thing will be decided this time — either polling is broken, or the time-honored tradition of reporting and observation is obsolete. It’s a fascinating question to resolve.

 

It’s Entirely Possible…

…that Barack Obama wins this election (more at noon).

But as the situation among reputable polls shifts ever more to Romney, and as more and more hitherto “likely Obama” states flip to “Leans Obama” and “Toss-up” and even “Lean Romney”, it’s interesting to watch Nate Silver doubling and tripling down on his prediction; he’s still giving The Light Worker a 75% chance of winning.

I’m not a statistician – but I can read and reason, and I’ve been dinging on Silver’s polling, methodology and predictions for a couple of years now.  My beef – and I’d suspect the beef of any rational person who isn’t one of the incurious low-information voters at which Silver’s polling is aimed – is that he calculates his results based on weighting existing polls based on some proprietary secret sauce known only to him.

Is the “sauce” valid?  I don’t know – nobody does, really – but as I showed in the 2010 Minnesota Governor’s race, it involved giving exaggerated weight to polls like the absurd “Minnesota Poll”, the so-bad-it’s-out-of-business Humphrey Institute poll and the frankly left-leaning PPP poll, while systematically shorting polls like Survey USA and Rasmussen.

Is Silver right?  Even if I could check his math, I probably couldn’t check his math, if you catch my drift.  Maybe Obama still is a near-sure thing, even after this past three weeks; maybe the Dems and Silver know something we don’t (like how many dead people will be voting).  We won’t really know until next week.

But while there will be many things about a Romney win that I’ll applaud, one of the big ones for me, personally, will be dancing – rhetorically, of course – on Nate Silver’s professional grave.

Shot In The Dark: Today’s News, Two Years Ago

Nate Silver at the NYTimes has been widely respected for his ability as a statistician.

His reputation, though, seems to stem largely from his facility at what amount to rhetorical parlor tricks (he once earned a bit of a living counting cards at poker, and he made a name for himself with baseball stats), and his calling of the vast majority of the 2008 election slate correctly (with the help of an epochal wave election and lots of access to Obama campaign internal polling), leading to his hiring at the NYTimes in time for the 2010 race.

Silver’s method at the NYTimes involves…:

  • Taking regional polls – from polling services as well as media polls – and…
  • “weighting” them according to some special sauce known only to Nate Silver, Registered Statistical Genius.

Now, I wrote about Silver’s method two years ago, when he spent much of the race predicting Mark Dayton would win by six points (with an eight-point margin of error).  As I pointed out, Silver’s “methodology” involved giving a fairly absurd amount of weight to  polls like the long-discredited Star Tribune “Minnesota” Poll and the since-discontinued Hubert H. Humphrey Institute poll (for whose demise I sincerely hope I deserve some credit, having spent a good part of the fall of 2010 showing what a piece of pro-DFL propaganda it has always been).  During the middle of th 2010 race, Silver gave the absurdly inaccurate-in-the-DFL’s-favor (especially in close elections) HHH and Minnesota Polls immense weight, while undervaluing the generally-accurate Rasmussen polls and, to a lesser extent, Survey USA.

I said Silver’s methodology was “garbage in, garbage out” – he uses bad data, and gets bad results.  I was being charitable, of course; his methodology, untransparent and proprietary as it is, processes bad data into worse conclusions.

That was in 2010.

Today?  NRO’s Josh Jordan reaches the same conclusion:

While many in the media (and Silver himself) openly mock the idea of Republicans’ “unskewing polls” (and I am not a fan of unskewedpolls.com by any means), Silver’s weighting method is just a more subtle way of doing just that. I outlined yesterday why Ohio is closer than the polls seem to indicate by looking at the full results of the polls as opposed to only the topline head-to-head numbers. Romney is up by well over eight points among independents in an average of current Ohio polls, the overall sample of those same polls is more Democratic than the 2008 electorate was, and Obama’s two best recent polls are among the oldest.

But look at some of the weights applied to the individual polls in Silver’s model. The most current Public Policy Polling survey, released Saturday, has Obama up only one point, 49–48. That poll is given a weighting under Silver’s model of .95201. The PPP poll taken last weekend had Obama up five, 51–46. This poll is a week older but has a weighting of 1.15569.

So it wasn’t just Minnesota!

And remember – PPP polls, while leaning a little left, are not generally flagrantly inaccurate in the sense that the Strib is and the HHH was.

And it’s not a fluke…:

The NBC/Marist Ohio poll conducted twelve days ago has a higher weighting attached to it (1.31395) than eight of the nine polls taken since. The poll from twelve days ago also, coincidentally enough, is Obama’s best recent poll in Ohio, because of a Democratic party-identification advantage of eleven points. By contrast, the Rasmussen poll from eight days later, which has a larger sample size, more recent field dates, but has an even party-identification split between Democrats and Republicans, has a weighting of .88826, lower than any other poll taken in the last nine days.

Jordan reaches a conclusion that even I didn’t:

This is the type of analysis that walks a very thin line between forecasting and cheerleading. When you weight a poll based on what you think of the pollster and the results and not based on what is actually inside the poll (party sampling, changes in favorability, job approval, etc), it can make for forecasts that mirror what you hope will happen rather than what’s most likely to happen.

Well, you can – if your goal isn’t so much to measure the nation’s zeitgeist (and report on it) but affect the election.

Which has, of course, been my contention all along.

Gallup-ing Towards The Finish

They don’t call it a horse-race for nothing.

As a rule in polling, outliers tend to get ignored.  Or you can choose to believe that Bush won Hawaii in 2004, Alf Landon won a 1936 landslide, or that Clinton v. Dole was a nail-biter.

But it becomes harder to ignore an outlier when it’s A) close to the election and B) one of the oldest and most respected polling outfits in the nation.  Thus as the media enters Campaign 2012’s home stretch, the narrative of a nip-and-tuck contest looks decidedly jeopardized by Gallup showing Mitt Romney with a 7% lead – and such an outcome apparently has to be challenged:

With a record of correctly predicting all but three of the 19 presidential races stretching back to 1936, Gallup is one of the most prestigious names in the business and its outlier status has other polling experts scratching their heads.

“They’re just so out of kilter at the moment,” said Simon Jackman, a Stanford University political science professor and author of a book on polling. “Either they’re doing something really wacky or the other 18 pollsters out there are colluding, or something.”

The caveats to Gallup’s polling (as with any pollster) are well-versed.  But to find an answer as to why Gallup posts a major Romney lead while the Real Clear Politics average of pollsters shows essentially a tie has nothing to do with credibility or collusion.  It has everything to do with turnout.

Take the recent IBD/TIPP poll as Gallup’s doppleganger with Obama leading by 5.7%.  Democrats are outsample Republicans by 7%.  The UConn Courant showing Obama up 3%?  The sample shows Democrats with an 8-point advantage.  Gallup plays their cards close to the vest, not showing the partisan affiliation of their likely voter model.  But their registered voter breakdown still shows a Romney lead, albeit of a modest 3% and is likely based on their party affiliation polls showing Democrats up 4 points.

Gallup says it determines its “likely voters” by asking whether they have voted in the past, if they know where their polling place is located, and other similar questions. The formula has been tweaked this year to take into account the increasing prevalence of early voting.

Gallup’s Newport pointed out that the firm’s likely-voter formula has more accurately predicted the election results than its wider poll of all registered voters going back to the 1990s and, in fact, the likely voter prediction tended to slightly favor Democratic candidates.

The idea of a single pollster being simply a part of a larger trendline is accurate, even if most media outlets tend to overlook that fact to trumpet their own poll to the exclusion of competitors and thus create news rather than report it.  Yet even if we exclude Gallup’s results, the trendlines have to be concerning for Obama’s camp.  Despite wielding turnout margins better than what propelled him into office four years ago, many polls show Barack Obama at best narrowly ahead – and more commonly tied or behind.

Gallup might be overstating Romney’s support, although the pollster’s worst estimations of support were in the 5-6 point range and happened in 1936 and 1948.  In the modern era, if anything Gallup has consistently overestimated Democratic support at the polls, giving Obama 2% more, Kerry 0.7% more and Clinton 2.8% and 5.7% more in his campaigns.  Which may mean that despite a 7% lead causing headaches among the media, Mitt Romney may…hold for dramatic effect…lead by more.

Chanting Points Memo: And We’re Back To The Fine Print

The left and media (PTR) was skipping and gamboling about like happy little meerkats yesterday; a new PPP poll showed – as PPP polls tend to do – nothing but good news for Minnesota Democrats.

In an automated phone survey of 937 likely MInnesota voters, they found…:

PPP’s newest poll on Minnesota’s amendment to ban gay marriage finds it running slightly behind, with 46% of voters planning to support it and 49% opposed.

That represents a 4 point shift compared to a month ago when it led for passage 48-47.

The poll claims that the major movement has been among indies and women.

“The marriage amendment in Minnesota continues to look like a toss up,” said Dean Debnam, President of Public Policy Polling. “Voters are very closely divided on the

issue.”

Well, we more or less knew that.

Then, they addressed the other Constitutional Amendment:

When we polled on it in June it was leading for passage by a 58/34 margin. By September that had tightened to a 56/39 advantage. And now it’s leading only 51/43. Democrats are now even more opposed to the voter ID amendment (23/71) than they are to the one on

marriage. And although independents continue to support it their 52/41 favor for it is down a good deal from 62/33 a month ago. This fight may end up a lot closer than people

initially expected.

Or it may not.

We’ll come back to that one.

They also put the DFL up substantially on a “generic legislative ballot”, which would be big news if voters voted for a generic legislature.  They don’t, of course.

As always, the devil is in the turnout model:

Here it is, buried deep in a set of crosstabs:

That’s D+9.  Not as far out as the D+13 we got from the Strib a while back, but it still higher than 2008.

That’s especially interesting compared to this other bit of crosstabbery:

So Democrats outnumber Republicans 38/29, but conservatives outnumber Dems 37/34?

At any rate – the polling services continue to put out (if you look hard enough for them) polls with turnout models that, when you ask them, they are are legitimately what they’re encountering out there…

…but do not in any way pass the sniff test.

And the media?

Well, they just shovel it on out there.  It’s just the topline number that really matters.  Right?

Chanting Points Memo: Puff

Who’s a better actor; Scarlett Johannson or Donald Rumsfeld?

Now, if you’re a sophisticated polling operation like the Star/Tribune “Minnesota Poll” – a Mason-Dixon joint – it’s an easy question; Ms. Johannson will likely outpoll the former Secretary of Defense.

Of course, you may respond “but measuring a Secretary of Defense in terms of acting skill is meaningless!”

And if you say that, then you’re already too smart to buy the latest Minnesota poll.

———-

The final Star/Tribune Minnesota poll of this long, nauseating week was released this morning – and it has two conclusions:

After two years of budget battles, vetoes and the longest state shutdown in Minnesota history, DFL Gov. Mark Dayton is winning the popularity battle with the GOP-controlled Legislature, a Star Tribune Minnesota Poll has found.

A slim majority of 53 percent of likely voters say they approve of Dayton’s job performance, while 31 percent disapprove. Another 16 percent say they are undecided.

For the majority leaders of the Legislature, the poll found 51 percent disapprove of the job they are doing. Another 21 percent approve and 24 percent are undecided.

This poll really needs two responses.

Damnation With Faint Praise

The poll notes that Governor Dayton scored a 53 percent approval rating.

Now, on its face, that’s not a good number.  Under 50%, says conventional wisdom, is trouble – and Dayton is a governor that’s done virtually nothing in two years but make odd, slurred pronouncements before scuttling away into his office under the cover of a fawning media.

But as always, you have to look below “the face”.  Of course, the poll has the same absurd, worse-than-Watergate-level turnout model – Democrats 41, Republicans 28 – as all the other polls this week.

However, for some reason the Strib doesn’t favor us with the full range of crosstabs; while reporter Jim Ragsdale notes the Governor’s geographic strengths and weaknesses, at no point in the article is the approval of Republicans or Democrats explained.

Is this an accident?  Or would showing it reveal something about the poll the Strib doesn’t want us to see?

Because if we assume Democrats are over-polled by 3%, and that they were the vast majority of the “approve” numbers, then Dayton’s approval drops down to right around 50% – and the “disapprove” numbers jump into the high thirties if we assume most Republicans disapprove of Dayton’s job.

Of course, it’s all guesswork until the Strib releases those crosstabs.

Damnation By Packing Peanuts

Of course, the numbers on the legislature are just plain nonsense

For the majority leaders of the Legislature, certain trouble spots stand out: Only 24 percent of voters in the metro suburbs outside of Hennepin and Ramsey — which include strong GOP areas — approve of their leadership, while more than half disapprove. In southwestern Minnesota, 57 percent disapprove. They had slightly stronger showings in Rochester, where 30 percent approve, and in the northwest corner of the state at 33 percent.

This is just a stupid thing to measure.

Grading a legislature, as an entire body, all together, is like asking what a football fan thinks of the NFC Central Division.  You will get a dog’s breakfast of opinions, or no opinion – because the division is not the focus (except for bracketing playoffs).

Put another way?  Nobody is going to vote for “The Legislature” this fall.   They will vote for or against candidates.  I will be voting for Rick Karschnia for State Senate and Dan Lipp for House; not “for the legislature” or even “for the House GOP caucus”.  And I”m a wonk! 
Indeed, this next paragraph sums up the absurdity of the whole question:

One startling figure is that the GOP-controlled Legislature only broke even among Republican voters: 31 percent approve, 32 percent disapprove and 37 percent are undecided. Undecided numbers are higher throughout the Legislature’s poll, suggesting many voters do not have a clear opinion on the topic.

Well, dug.

I disapproved of the Legislature’s job!   The GOP caucuses gave away too much money in 2011 and caved on the stadium last year without getting anything useful in return!  I give them a “D”.

Does that mean I’m going to support Mark Dayton?  Not if you held a gun to my head (not to give the SEIU any ideas).

It’s a meaningless number.

So Why Run A Poll With A Meaningless Number?

That one’s easy; it gives the Strib a nice tidy number – 53-21 approval ratings! – to toss in front of people who don’t pay much attention to what the numbers actually mean.  These voters – the “Low-Information Voters”, people who retain headlines from the Mainstream Media and believe things that appear in the Strib – are the target for this sort of polling, or as it’s known in the world of logic, a “non-sequitur”.  The Strib is comparing apples and axles, just like comparing Donald Rumsfeld with Scarlett Johannsen.

While the individual numbers might be valid (they’re not – remember, the turnout model is absurd), at least in terms of math used to generate the numbers we see – but even if both numbers were in fact dead-nut accurate, what the Strib has done is created a phony horserace.

And why would the Strib splash a phony, non-sequitur comparison on its front page when it only serves to show Mark Dayton with a decisive (if phony) lead…

…oh, wait.

Never mind.

More Monday.

Chanting Points Memo: The Rigger’s Dilemma

It’s my contention that the Star/Tribune “Minnesota” poll is, and has been for two and a half decades, less a “public opinion” poll and more an instrument of DFL propaganda.  I’ve supported that contention with a raft of circumstantial evidence; proof that the Minnesota Poll underestimates GOP turnout – especially in races that are perceived to be close; it showed Mark Dayton with an absurdly huge lead over Tom Emmer, and Al Franken with a four point lead over Norm Coleman, while guessing the Klobuchar/Kennedy race fairly accurately.

It’s my contention that this is to leverage the “Bandwagon Effect” – to discourage Republicans and conservatives from going to the poll.

But this year’s race presents a dilemma for the editors who – I’m being half-hyperbolic here [1] – plan the results of these polls.  On the one hand, you have the Voter ID initiative which is likely to win in a blowout.  To skew that poll enough to encourage Democrats and opponents of the amendment, the Strib would need to skew the poll to an absurd extent.  As in, assume conditions that are the same as in 1976, after Watergate.

On the other hand, you have a Senate race between Amy Klobuchar and Kurt BIlls that is widely perceived to be a pretty safe race for the incumbent.   Skewing the sample too far to the left would make the results look completely implausible.

The answer, if you’re the Strib?  You see it in this weeks’ Minnesota Polls; this is their old buddy Jim Klobuchar’s daughter we’re talking about here!  Of course they’ll do what it takes to make her re-election as epic as possible – why, everyone on Editor’s Row remembers Amy when she was just this tall, dagnabbit!

Beyond that – and more germane to the propaganda organ – they know that the voters the DFL needs are the “low-information” voters. The ones that rarely get past the headline, much less the lead – forget about looking at partisan breakdowns.  The ones that still believe the Strib is anything but DFL shills, or don’t care either way.

The Strib is showing a 57-28 lead for Klobuchar, with 15 percent either undecided or voting for someone else.  As we’ve been showing every day this week, this is based on a sample that includes 41% Democrat/28% GOP turnout.

Now, if we assume it’s more like 38/34 – which is more in line with Rasmussen’s figures, which have been traditionally vastly more accurate – and multiply the changes by the support each candidate gets within their party…

…well, that’s bad for Bills, if you believe the Strib.  While 90-odd percent of Dems say they’ll vote for A-Klo, the Strib claims only 2/3 of Republicans will vote for BIlls.  While the nomination battle was a bruising one in the GOP, and left a lot of bad blood, Republicans are much more suck-it-up-and-support-our-guy than that.  This strikes me as dubious – the “stink test” is crying “BS!” – but I’ve got no hard evidence to the contrary just yet.  Absent that, let’s run with the Strib’s numbers.

So if we subtract 3% from the Democrat split, multiplied by 90%, we get a net loss of 2.7% for Klobuchar, taking her down to 54.3%.

Adding 6 to Bills – times the 66% support in the party, naturally – leads to a four point rise, to 32.

54-32 still isn’t close.  But it’s not the 2:1 humiliation…

…that, I contend, the Strib wants Republicans to believe is coming, on top of all the other “bad news” they’ve brought us this week.

So what does all this mean?

More tomorrow.

Continue reading

Chanting Points Memo: Tie Manufacturing Is Way Up!

The Star Tribune “Minnesota Poll” has declared Minnesota tied on the Obamacare issue:

About 46 percent of the state’s likely voters say they support keeping the Affordable Care Act, whose main tenets were largely upheld by the U.S. Supreme Court this summer, while 47 percent believe the law should go.

As always when dealing with any polls – especially polls with records of dishonesty as long as the Strib’s -http://www.startribune.com/local/171271561.html.  The Strib’s, in case you’ve forgotten,  is 41% Democrat, 28% Republican – which, as I showed yesterday, is more Democrat-leaning than the 1976 post-Watergate, post-Nixon-pardon election, the post-war nadir of GOP fortunes.

As a bipartisan sampling – liberal Hamline poli-sci professor Dave Schultz and I – both agree,  the sample is more like 38% DFL, 34% GOP.   Since 80% of DFLers (according to the Minnesota Poll) support the amendment, that means you deduct 80% of 3 points – 2.4 – from the “support” column.  Likewise, 90% of the six additional points of Republicans – 5.4% – support repeal.

With those assumptions – a more realistic turnout model and those levels of support – the “repeal” case is really more like 52-44.

It looks like the MInnesota Poll is building up to the Senate Race.

Chanting Points Memo: That D+13 Split

As I wait for the latest “Minnesota Poll” to release its results for the Senate race, I’ve been turning the poll’s D+13 (their sample of respondents was 41% Democrat and 28% Republican) number around in my head.

After all, as the Strib tells us, “Brad Coker, managing director of Mason-Dixon Polling & Research, which conducted the poll for the Star Tribune, said those numbers are consistent with what he has seen over the years”.

But as we noted yesterday, the 2008 election – an epic Democrat win – was D+6 or so.  The 2010 election had turnout of D+2, roughly, and turned out to be a GOP rout nationwide and in the MN Legislature.

So what about the worst election in the past 50 years for the GOP – the post-Watergate presidential election of 1976?    Where the GOP got shredded in DC and in Saint Paul, sending the MNGOP running to their “Independent Republican” label?

I can’t find the partisan split – but does it seem unreasonable that in a year when Jimmy Carter beat Gerald Fold by 12 points in Minnesota that the partisan split was, maybe in the neighborhood of D+12?

In other words, maybe somewhere around the D+13 number the Strib would have you believe today?

The Great Poll Scam: A Blast From The Past

As we look at the abusive travesty that is the Minnesota Poll – in this case, the ludicrously skewed, 3:2 pro-DFL partisan breakdown in this weekend’s polling on the two Constitutional Amendments – let’s take a trip back through history.

Frank Newport, the president of the broadly-respected Gallup Polls, savaged the Minnesota Poll in the immediate aftermath of the 2010 midterm and gubernatorial elections.

Let’s make sure we’re clear on this.  Pollsters attacking pollsters in public is a little like magicians publicly heckling other magicians.

Is there any evidence the Strib has polished up their methodology?

No.  Indeed, the two polls released over the weekend on the Marriage and Voter ID amendments show quite the opposite – or that some copy editor took a poll of Ramsey County voters and mislabeled it a poll of Minnesotans.