I wonder if this means that there are less crazy people in Wisconsin than elsewhere, or if it's gotten to the point where even the crazies are going, "Hey, whoa, slow down there."

Posted July 19, 2007 8:36 PM by ethan

If he was still muddling along at 50% you could have used "Stuck in the middle with you."

Posted July 19, 2007 8:46 PM by Linkmeister

I haven't been following Wisconsin politics recently, but the state has a long history of doing its own thing, and electing people who wouldn't stand a chance anywhere else, if only because of how they were labeled. A Socialist mayor, after all.

As for the crazies, maybe what's happening isn't so much Bush sinking in the polls as people's perception of Bush' craziness approaching reality. If the crazies aren't as crazy as they think he is, maybe they want to back away from him.

The cynical part of me keeps repeating, "Yeah, sure, in your dreams. Probably someone's been fudging the figures."

Posted July 19, 2007 8:55 PM by Bruce Cohen (SpeakerToManagers)

Linkmeister @ 2

Just so, with clowns to the left of you and jokers to the right. Doesn't sound strained to me at all.

Posted July 19, 2007 9:06 PM by Bruce Cohen (SpeakerToManagers)

(Is there a Bloggie award for most strained musical reference?)

Hm....

All those who remember the war
They won't forget what they've seen..
Destruction of men in their prime
whose average was 19

--Paul Hardcastle, "19"

Posted July 19, 2007 9:09 PM by Mary Dell

Hey nineteen
We can't govern together
We got nothing at all
Please give me a pardon as you slide on out

Posted July 19, 2007 9:11 PM by albatross

Bush is at 24% in New Hampshire -- with a 4.5% margin of error. He could be at 19% here, too.

Posted July 19, 2007 9:25 PM by James D. Macdonald

This seems unreal. 19% -- and we still can't get out of Iraq. We live in Bizarroworld.

Posted July 19, 2007 10:34 PM by Lizzy L

"The poll had a margin of error of plus or minus 3 percentage points."

So it could only be 16%.

Posted July 19, 2007 10:36 PM by Jonah

It's the iterated prisoner's dilemma. You interact with someone very differently if you have to worry about future interactions than if you don't have to worry about them.

Neither Bush nor Cheney is running again, nobody with any influence in this administration has any kind of near-term political future. So unpopularity almost doesn't matter. They can make the voters as unhappy as they like, and the Republican party will suffer, but they won't. They've already lost what they're going to lose. This gives them freedom to do anything that won't get them impeached or indicted.

Posted July 19, 2007 10:39 PM by albatross

Ironically, it's the a liberal act he committed to loose the crazifaciton factor - he's in favor of an amnesty program for illegal immigrants.

Posted July 19, 2007 10:58 PM by Josh Jasper

They can make the voters as unhappy as they like, and the Republican party will suffer, but they won't.

Albatross, this is undoubtedly true, but what it doesn't explain is the behavior of the Republicans in Congress. They must be hearing from their constituents about the war. Or perhaps they, like Bush, only talk to people who will support their decisions and stroke their egos.

Posted July 20, 2007 12:12 AM by Lizzy L

Lizzy L

Oh, the Repblicans are hearing about it. But they can't just turn around on a dime; they need to keep some sort of differentiation between themselves and the Democrats, or they'll have no chance at all in the next election. Who wants to vote for Republican who has exactly the same position as a Democrat? Also, they can't allow themselves to be accused of "flip-flopping" since they made such a big deal of using that epithet on the Democrats in the last two elections.

It puts them between Irag and a hard place. They can try to distance themselves from Bushy strategy, but they can't argue with all the tame generals Bush has planted in his garden, because they've made a big point of how they'll do the militarily correct thing.

Posted July 20, 2007 1:03 AM by Bruce Cohen (SpeakerToManagers)

Speaking of Bush's approval ratings, I like this handy chart.

It's truly astonishing, and, as you stare at it for a time, more and more ways in which it is astonishing keep revealing themselves.

Posted July 20, 2007 1:42 AM by ethan

ethan, that chart is amazing.

Posted July 20, 2007 1:47 AM by Linkmeister

Just eyeballing the chart, some of the pollsters seem consistently higher than average, some consistenly lower, which suggests systemic polling errors. Are those errors equally distributed? That is, are some of the biases for Bush, and some against?

(And all we can really tell is the difference from the average: we can't tell if anyone is really biased against Bush; just that they give lower than average results more often than chance would explain. Or vice versa.)

Posted July 20, 2007 2:53 AM by Dave Bell

That poll also showed 46% approving of Bush's handling of the war on terror, and 43% disapproving. I don't even know what that means. How could an overwhelming majority not approve of his performance, and a plurality approve of his handling of the war on terror?

Posted July 20, 2007 4:12 AM by Sean Sakamoto

the whole crazy factor neglects the slow thinking popularity followers. These are the people who follow someone because they are popular or successful but take somewhat longer to realize that popstar X is no longer cool than the average popularity follower.

So what's going on is that these people are being laughed into realizing "hey the president sucks"

I expect the craziness factor may in fact be somewhere in the twenties but the chances of the crazies actually matching Bush's policies and profile with their fantasies seems very negligible.

Posted July 20, 2007 4:41 AM by bryan

Dave Bell @ #16 - Just eyeballing the chart, some of the pollsters seem consistently higher than average, some consistenly lower, which suggests systemic polling errors. Are those errors equally distributed? That is, are some of the biases for Bush, and some against?

I would take the opportunity to brush off some of my stats skills but unfortunately I'm just about to go to a Beer Festival and drinking and maths don't mix*. The other way is to look at the methodology of each poll - are they asking the same question, are they asking the same people every time or different groups, cohort size etc. - questions you ought to ask every time you compare polls.

* or rather drinking and maths mix too well.

Posted July 20, 2007 5:58 AM by Neil Willcox

Damn it - I made comment number 19 without making any clever references.

Posted July 20, 2007 6:01 AM by Neil Willcox

Jay Leno had a sad but true joke about this:

Bush's poll numbers are so low, the only thing he's still above is the law.

Posted July 20, 2007 6:03 AM by L. Raymond

Sean @ 17:
How could an overwhelming majority not approve of his performance, and a plurality approve of his handling of the war on terror?

Not enough tax breaks, of course.

Also, nowadays even crazies know that WoT and Iraq are different things... WoT good ("ragheads in prison!"), Iraq bad ("We should have been in Teheran YEARS ago!").

Posted July 20, 2007 6:13 AM by GiacomoL

Anyone care to expand on what amazes them about that chart?

All I'm seeing is a classic "how to lie with statistics" feature--the bottom quarter of the chart is cut off to make Bush look even less popular than he is, and if I extrapolate by eye, his approval rating should be at zero by September 2008.

Posted July 20, 2007 6:29 AM by Nancy Lebovitz

Dave Bell (16), it does nothing to discourage my longstanding suspicion of Gallup Polls.

Posted July 20, 2007 6:30 AM by Teresa Nielsen Hayden

More food for thought: this chart from the same site, showing educational level versus percentage of vote for Bush in 2004. Looks like his doofus impersonation really paid off.

Posted July 20, 2007 7:01 AM by Bruce Cohen (SpeakerToManagers)

Nancy Lebowitz @ 23

All I'm seeing is a classic "how to lie with statistics" feature--the bottom quarter of the chart is cut off to make Bush look even less popular than he is

That was the second thing I noticed; the first thing was that the period of the small cycles in the interpolated curve is about the size of the averaging window, which tells me that the "real" period of the swings is probably shorter, like about one news cycle. That and the (at least by eye) apparently constant variation of each poll from the average tells me that the polls are not really telling us what people feel about Bush at any given moment; more like what the pollsters want to tell us about what they think people should be feeing about Bush, averaged over whatever the current fuss is about.

But what's really interesting to me is not the absolute value of the approval rating at any given time, but the rate at which the curve changes. Note that after each major upward jump (9/11, the capture of Saddam, etc.), the curve decays rapidly downward, and that in the first 3 major such events, the decay is faster than than the one before. That indicates to me that most people's approval of Bush is based mostly on artificial spikes of feeling good about something he said (he was "presidential" on 9/11) or an event that he co-opted to represent his own actions (Saddam's capture). All that supports my own belief that his approval is almost completely based on PR, that his actions themselves are generally viewed, at least in retrospect, as neutral or negative. Maybe you can't really fool all the people all the time after all.

Posted July 20, 2007 7:19 AM by Bruce Cohen (SpeakerToManagers)

Nancy @23:
the bottom quarter of the chart is cut off to make Bush look even less popular than he is

That chart is probably automatically generated, and the program cut it off at the nearest "5" (see how it includes 95 even though the highest value is around 92, but doesn't reach 100). Malice might be in the eye of the beholder here :)

Posted July 20, 2007 7:43 AM by GiacomoL

I always put the value of CF rather lower - 20 per cent, on the principle that assuming crazy is normally distributed, there's a 20 per cent probability of 80 per cent crazy.

But this is challenging my thinking. Question - is it that the crazies are coming around, or that the president is marching along the crazy curve?

Posted July 20, 2007 8:07 AM by Alex

Bruce Cohen @ 3: Vermont not only elected a socialist mayor (for Burlington), but then sent him on to the house and then the senate... (Sanders)

Posted July 20, 2007 8:23 AM by Mike G.

To steal a phrase from Clinton, "It's the immigration, stupid."

It's not that the right-wing crazies are becoming any less crazy, it's that Bush finally did something to piss them off. I think it's hard to overestimate the extent to which the right-wing crazies were scandalized by the slightest suggestion that Bush might support granting any sort of legal status to illegal immigrants. These people's biggest fear is that the brown people (whether they're Muslim or Hispanic) are going to come over here and out-breed us, turning America into a third-world country, murdering us in our beds and polluting our precious bodily fluids.

This also explains the discrepancy with his approval ratings on the war on terror, it's just that their fear of brown people over there in the middle east is trumped by their fear of the brown people right on our borders.

Posted July 20, 2007 8:32 AM by Chris W

Alex @ #28:

Given the stated margin of error, the real value (whatever "real" means in cases like this) might be as high as 22%.
Which is still markedly lower than KFM's CF value, but safely higher than yours.

Posted July 20, 2007 8:34 AM by Paul A.

I always go to the poll-of-polls chart at "Mystery Pollster" which includes a calculated-by-a-statistician trend line.

Posted July 20, 2007 8:37 AM by Matt

Josh @ #11, he pissed off some liberals even more by his support of an expanded guestworker program. The current program is a nightmare; see the Southern Poverty Law Center's report.

Posted July 20, 2007 9:11 AM by Lila

Nancy @ 23, I personally figure that his approval ratings should be in negative numbers long before then.

Posted July 20, 2007 9:45 AM by Dave Weingart

Chris #30:

I think the immigration debate was the last straw for a lot of conservatives, but that their support has been decaying for quite a while.

One thing the chart doesn't show as much as I would have expected (though you can see the effect) is Katrina. That struck me as the thing that snapped a lot of people out of their "in danger, must follow the alpha male" spell. In some sense, a visible disaster here made disaster out of sight in Iraq more plausible, and some of the weird media gymnastics on Katrina, made it a lot more plausible that the media might be playing down some of the downside in Iraq.

Posted July 20, 2007 10:18 AM by albatross

Speaking of Katrina: FEMA puts thousands of people into poisonous trailers; refuses to test for formaldehyde because then they'd have to do something about it.

Posted July 20, 2007 10:50 AM by Lila

#27 ::: GiacomoL

OK, the chart wasn't manipulated in order to make Bush look even less popular--the system is automated to make *all* results look more dramatic.

Posted July 20, 2007 11:11 AM by Nancy Lebovitz

Nancy at 23: if I extrapolate by eye, his approval rating should be at zero by September 2008.

Which, depending on who the Republican nominee for President is, will make the 2008 election even more interesting... Pass the popcorn.

Posted July 20, 2007 11:56 AM by Lizzy L

Lila: thanks for the guestworker report pointer! It's going in my online books listings tonight.

(If anyone know of other online books, on whatever subjects, that should get wider attention, I invite you to visit http://onlinebooks.library.upenn.edu/suggest.html )

Posted July 20, 2007 11:58 AM by John Mark Ockerbloom

These people's biggest fear is that the brown people (whether they're Muslim or Hispanic) are going to come over here and out-breed us, turning America into a third-world country, murdering us in our beds and polluting our precious bodily fluids.

At which nefarious plot they've already succeeded in England, according to the cruise-goers this journalist went among at great risk to life, limb, and sanity...

I am standing waist-deep in the Pacific Ocean, both chilling and burning, indulging in the polite chit-chat beloved by vacationing Americans. A sweet elderly lady from Los Angeles is sitting on the rocks nearby, telling me dreamily about her son. "Is he your only child?" I ask. "Yes," she says. "Do you have a child back in England?" she asks. No, I say. Her face darkens. "You'd better start," she says. "The Muslims are breeding. Soon, they'll have the whole of Europe."

Pretty much everyone on that ship was telling the journalist to move to America since England was already irrevocably browned.

Brrrr. Better him on that ship than me.

Posted July 20, 2007 12:22 PM by Nicole J. LeBoeuf-Little

#40:

God, what a beautiful portrayal of an echo chamber. I kept wondering how many more people were sensible, but didn't say anything because they didn't want to get shouted down. And how many had self-selected out of this group, from a single trip or from knowing ahead of time what the experience would be like. But spout the party line in this place, however crazy, and you could expect social acceptance and support. Buck the party line and you could expect pressure to shut up and argument from all sides.

There are lots of generally applicable lessons here. Don't join an echo chamber. Find news sources, conversations, and books that don't share all your starting assumptions. Try to entertain the hypothesis that any big idea of yours is false, and see how it works out.

Frex, try it as an exercise with the war in Iraq. Start with the hypothesis that it's really going way better than it looks, that the media are painting it as a disaster when it's really a glowing success. Then see whether you can honestly falsify the hypothesis with available evidence. Note the multiplicity of sources that all look bad. Note the unrest on the border with Turkey, the massive numbers of Iraqi professionals who've fled, the many different mortality reports including the Lancet study, the international coverage of things in the war, the observable events like the need to extend tours of duty and "surge" and big explosions that get covered everywhere.

Doing that is enormously valuable. Not doing it makes it easy for you to get caught in a feedback loop with others who share all your ideas, and become immune to contrary evidence. But your life may be more comfortable, and you may fit in better, if you don't do it, because it sometimes really p-sses people off when you contradict their everybody-knows-this ideas.

Posted July 20, 2007 1:21 PM by albatross

What strikes me as bizarre about that cruise is that it was sponsored by the magazine (Nat'l Review) that was once a flagship for William Buckley's conservatism. His version wasn't off-the-charts-insane and blatantly racist, which his movement seems to have devolved into now.

Posted July 20, 2007 1:58 PM by Linkmeister

#42: I didn't have the heart to do more than skim that article, but one of the pp I did read described a participant making a gesture that suggested that Buckley was getting soft in the head for not buying into the paranoia.

Posted July 20, 2007 3:03 PM by Stefan Jones

An interesting note w.r.t. the original topic of the discussion: The BBC is reporting that Bush has signed an executive order banning torture of detainees. (I guess that would be the torture that we weren't doing, which was all attrocity stories spread by the liberal media, and which the terrorists deserved anyway.)

Posted July 20, 2007 3:20 PM by albatross

Alex @28: I'm afraid I don't quite follow; are you suggesting that the mean level of craziness is 50%? and one s.d. is about 30% craziness? It's been far too long since I actually had to do any stats...

Posted July 20, 2007 4:26 PM by Jakob

There's a simple, cheap air filter for formaldehye, called "houseplants" -- studies show the common types like spider plant and the one I can't remember the name of that someone at MITRE had going around and around in their office (literally, there was a suspension system for it on the ceiling and the leaves had grown on stems that had gone around the office -twice-!) take more than 90% of formaldehyde of out the air in a mere 24 hours....

Posted July 20, 2007 5:07 PM by Paula Lieberman

Lila #36: When I read that in this morning's paper (buried on page 3), my first thought was that the Shrub maladministration has fallen below even my expectations. These people, using the term loosely, combine incompetence and evil in a starkly new way.

Posted July 20, 2007 7:38 PM by Fragano Ledgister

Dave Bell @ 16:

I can't say anything specific about US polls, but in general opinion polling is a bit of an art. You might think it's as simple as asking people "what do you think of X" or "who did you vote for last time", but it tends to be a lot more complicated than that. First, you have to make sure your sample is reasonably representative. This is non-trivial, but it's actually one of the easier parts of the process. The you have to somehow account for the fact that people will lie to you - or, to be more charitable, misstate or misremember their opinions. This has been a notorious problem in UK political polling, in that many more people vote Conservative than will admit to it. Polling companies try to identify these confounding factors and introduce parameters to compensate for them - but different companies do this differently, and it can be difficult to tell whose method is superior. And if that isn't bad enough, these confounding factors change with time. Is it better to frequently update the model, to take account of shifting public responses, or would it be better to maintain a consistent model so that valid year-on-year comparisons can be made?

The consensus among political polling geeks is that every polling organisation has its own systematic bias, and the trend in the polls is much more informative than the absolute levels of support for particular parties or leaders. This is borne out in the example we're dealing with just now: there are systematic differences between each poll, but each one shows the same trends.

The bottom line is, don't get hung up on the absolute percentage points: pay attention to the long-term trends.

Posted July 20, 2007 7:51 PM by Iain Coleman

Paula @ 46: I'm guessing that's pothos, which is a wonderful indoor plant -- almost impossible to kill, survives in near-darkness, and will happily grow as long as you'll let them.

It's nice to know that houseplants are good for something in addition to looking pretty. I knew there was a reason I liked them. :-)

Posted July 20, 2007 8:45 PM by Kevin Riggle

#40, Nicole - that's a really statistically significant polling population there - Maudie Littlehampton on vacation.

Of course, we've already had a female Prime Minister and are likely to have another, so we must be doomed.

Ah, but what do I know? I only live here.

(Maudie Littlehampton - a venomous Home Counties lady wot lunches from daily cartoons of yore, whose politics might be defined as regarding Margaret Thatcher as a dangerous liberal and wogs begin at Dover and Watford)

Posted July 21, 2007 8:43 AM by Martyn Taylor

Giacomo #27: It may be that the y-axis limits are automatic, but equally likely that they were selected by a human to maximize detail and contrast in the curve. I do a lot of graphing in my work (desktop publishing in an investment bank) and there is a large contingent of bankers who abhor white space in their charts, even if it would make understanding them "more intuitive." (Things like maintaining a 0-100% range on a y-axis, for example.)

Posted July 21, 2007 9:26 AM by Bruce Adelsohn

Paula @ #46, it's unfortunate that FEMA didn't take the simple expedient of issuing spider plants along with the trailers.

Posted July 21, 2007 12:36 PM by Lila

42

There's a wonderful copy circulating of Buckley's 1964 or so defence of States Rights in the National Review (you know, against that nasty Mr Martin King and those slickers in Washington with their Voting Rights Act and their Civil Rights Act).

Buckley has always played fast and loose on both sides: sounding reasonable and intellectual to the greater audience, and then speaking in the coded tongues of some of his extreme right backers. David Niewert at Orcinus is very good on how these right wing pundits manage this so adroitly, so that the Mainstream Media seldom if ever calls them on it (Trent Lott was only brought down by diligent, diligent work by Jonathan Micah Marshall and Talking Points Memo) whilst they still talk to the 'base'.

Generally I think these polls show that Bush will make that pop at Iran: he has nothing to lose, and he believes history will judge him right.

War with Iran, on top of our other miseries...

Posted July 21, 2007 2:57 PM by Valuethinker

Paula; while certainly useful information, I wonder a little at the point of bringing it up? J8ust as a useful piece of trivia? That this may explain why some families aren't as effected (Which FEMA seems to now be using as its next excuse for stalling - "We need to do more research to find out...") ? That people living in government issue trailers should just go buy houseplants to make up for government negligence? (And how much disposable income do they have?) That the government should issue them plants rather than doing a proper clean up? Or am I missing something?

Posted July 25, 2007 2:02 PM by Lenora Rose