# Kevin Drum

## Statistical Zombies

| Mon Nov. 29, 2010 2:04 PM EST

Seth Michaels wants me to republish an old post setting out the top ten mistakes that infest day-to-day reporting of numerical and statistical information. Well, why not? Let's call them statistical zombies so that I can get in on the zombie craze. Here they are:

1. What’s the real income? Money comparisons over time should almost always be reported in real inflation-adjusted terms or else they're worthless. In nearly all cases, they should be reported in per capita terms as well.
2. What’s the survey error? Statistical sampling error in opinion polls is trivial compared to the error from other sources. Things such as question wording, question order, interviewer bias, and non-response rates, not to mention Bayesian reasons for suspecting that even the standard mathematical confidence interval is misleading, give most polls an accuracy of probably no more than ±15%. Example: a couple of years ago a poll asked respondents if they had voted in the last election. 72% said yes, even though the reality was that voter turnout in that election had been only 51%. Most polls and studies are careful to document the statistical sampling error, but who cares about a 3% sampling error when there might be 21 points of error from other causes?
3. Does A really cause B or might there be another explanation? If A and B are correlated, A might indeed cause B, but it’s also possible that it's just a coincidence or — even more likely — that some third source is causing both A and B. This problem is especially rampant in social science studies where virtually everything is related to everything else and even well designed multivariate analysis is extremely difficult.
4. Is it the first study? Even putting aside other errors, 95% confidence means there’s a 5% chance that the result is wrong. We only believe that smoking causes cancer because there have been hundreds of confirming studies. Always be cautious about accepting the first study on any subject.
5. Maybe it really was just a freak chance. “That can’t be a coincidence” is usually the result of not understanding how many rare things are nonetheless likely to happen once or twice in a population of 300 million. In a large country, there will always be some cities, or some groups, or some people, that are way above average for, say, cancer. The flip side of this is that something that seems dangerous might not really be. 100 kidnappings a year might seem like a lot, but in reality those are odds of one in three million. That's less likely than the odds of two people randomly picking out the same word from an encyclopedia.
6. Compared to what? A 5% rise might be good or might be bad depending on whether everything else is growing at 0% or 10%. Which is it?
7. Is there contradictory data? Two types of publication bias are involved here: researchers often don’t publish null results, and newspapers don’t bother reporting them when they are published.
8. Statistically speaking, why did the headline number go up (or down)? Did everyone’s income go up 5%, or was it just that Bill Gates’ income went up 1000%? Distribution is as important as central tendencies. Check for mean vs. median. The value of statistics is to summarize a large mass of data, but it’s important not to summarize too much.
9. Was the sample large and unbiased? For example, the original gay gene study used only about 40 people, and that was simply all the data they had. What’s worse, even if you do have a large sample it’s still difficult to ensure that it’s unbiased. Chapter 29 of Dana Milbank’s book Smashmouth is a pretty good down-and-dirty introduction to the delicate and tricky decisions that election pollsters have to make under deadline pressure to try and get accurate results.
10. Does all the data point a little too cleanly to a single cause? Life is messy. A single report can often produce masses of data and should probably be viewed with suspicion if it claims that every bit of its data can be explained by a single cause — especially if it's a cause that the researcher is already known to favor.

Plus I'll add another pet peeve that's numerical but not strictly statistical in nature: the reluctance to simply report percentages as percentages. This leads to sentences like this: "Although 40% of voters support spending reductions, nearly a third don't want defense spending cut and two in ten decline to identify any reductions at all." This is almost unreadable. Why not just provide the percentages in parallel form? I really don't think readers are so stupid that they can't handle it.

And via Seth, here's another approach to identifying bad everyday statistics, but with examples! Take a look and see if you can spot the errors.

| Mon Nov. 29, 2010 1:14 PM EST

So what do I actually think about the WikiLeaks dump of U.S. embassy cables? A bunch of contradictory things, it turns out. Here's sort of a stream of consciousness of what was going through my mind yesterday as I read about this:

• This isn't a whistleblower case. In fact, surprisingly little official lying of any kind has been revealed so far. Rather, it's an action aimed very generally at weakening American influence and exposing American intentions.
• To the extent that this is done by, say, Australians and Germans, there's nothing unpatriotic or even wrong about this. If foreign nationals who oppose American hegemony get a chance to lob a stink bomb at America, why shouldn't they?
• At the same time, the American citizen who leaked this stuff — Pfc Bradley Manning, apparently — should do serious time as long as the government can convince a jury of his peers that they've nailed the right guy. No government in the world can, or should, tolerate this kind of massive security breach from one of its own.
• Governments have lots of legitimate reasons for wanting to keep communications confidential. This is not some kind of weird pathology exclusive to nation states, either. You keep secrets. I keep secrets. Companies keep secrets, families keep secrets, labor unions keep secrets, nonprofits keep secrets, and your neighbors keep secrets. There's always the risk of this stuff going too far, and there's always the risk of your secrets getting spilled. But this is all part of the human condition, not a sign of depravity in the State Department.
• In general, this kind of indiscriminate data dump is a bad thing. This particular dump, for example, could conceivably hurt chances of ratifying the START treaty, strain relations with the UN, strain relations with Russia, make an attack on Iran a little bit likelier, and reduce even the meager leverage we currently have over Hamid Karzai. More broadly, it could hobble American efforts at replacing saber rattling with genuine diplomacy, which is really in nobody's interest.
• And yet....even secrets that are justifiable in the short term can often end up being toxic in the long term. Routine secrecy quickly becomes a crutch, as it plainly has in the United States, and an occasional informational enema like this can have a salutary effect. Governments might have a legitimate need for secrecy, but they should also be keenly aware that there's a risk to doing business like this. I wouldn't want this kind of mass disclosure to become a regular occurrence, and I do think the leaker should pay for his crime, but at the same time I can't honestly say that I'm entirely sorry this happened.

I suppose this is all very unsatisfactory, but those were the thoughts buzzing through my brain yesterday. They still are. Perhaps they'll gel into something more definite as I give it more thought and the ramifications become more clear.

## The Future of the Euro

| Mon Nov. 29, 2010 12:18 PM EST

With financial contagion from Greece and Ireland starting to spread to Portugal and Spain, the eventual breakup of the euro is becoming at least conceivable. But is it likely? Barry Eichengreen is probably the guy who's thought the hardest about this, and back in 2007 he concluded that it was extremely unlikely regardless of circumstances. The folks over at Five Books recently asked him for his latest thinking on the subject:

Rarely does an academic have the privilege of a real-time test of his hypothesis. With the benefit of that test, I would now say that I was both right and wrong. I was right in that, yes, if the Greek government were to announce tomorrow that it had decided to reintroduce the drachma, it would precipitate the mother of all financial crises. Everyone would know that its intention was to depreciate the new drachma, so in the first minute everyone would rush to get their money out of the country, out of its banks, and out of its bond market. The result would be the biggest bank run and financial crisis the world has ever seen. This danger is a formidable deterrent to even contemplating going down this road. So I think the argument I made in 2007, that attempting to exit the euro area would be the equivalent of burning down your own house in order to find a way out, was exactly right.

In what way were you wrong?

I was wrong in that, as Paul Krugman observed earlier this year, if the house is burning down anyway, then the normal advice not to play with matches loses much of its force. If there’s a run on your banking system anyway, then the deterrent to action no longer applies. If there’s a run on your banking system and you have to close down your banks and financial markets anyway, you may want to take that opportunity to reintroduce your own currency. I still don’t think that things will be allowed to get to this point, but I no longer attach a zero probability to a country’s exiting the euro — just a close to zero probability. Never say never, but I still believe that the euro is an example of a path-dependent historical process that is unlikely to be reversed.

I think he's probably right. But I'm also not entirely sure of that. So far there's been no sign of a serious run on any euro-area banking system, but there have been signs of a "bank walk," for lack of a better term. And that could lead to a run on one bank which, in turn, could lead to a run on so many banks that the European monetary authorities can't stop it. Megan McArdle's warning that pundits "have started seeing Creditanstalt everywhere" is a good one, but this is still a scary process we're going through. I'm not sure exactly what nonzero probability Eichengreen would attach to a euro breakup, but even five or ten percent is pretty serious. Buckle up.

As always, Paul Krugman has more. So does Tyler Cowen, who is pessimistic. "In a nutshell," he says, "we're watching the most pitched, highest-stakes, most determined battle between politics and finance which has been staged. I am expecting finance to win."

## Bombing Iran, Part 2

| Mon Nov. 29, 2010 1:21 AM EST

David Frum asks a question (in a series of 140-character chunks):

What do we learn from Wikileaks re Iran?

1) Many more govts than you might think back a US military strike.

2) It's now public knowledge that Iran and North Korea are exchanging deadly military technology.

3) Whole world can see that US has gone every extra mile to reach a negotiated outcome with Iran

4) Nevertheless Iran has pursued nuclear ambitions all-out.

Seems to me the combined effect of this information would be to make US military action more politically acceptable....Both inside US and outside.

Discuss.

Also for discussion: all governments have a legitimate need for a certain amount of secrecy. In particular, embassy officials need to be able to report candidly to their superiors about what's going on in their sphere of responsibility. So what's the most likely consequence of the WikiLeaks document dump? That governments around the world realize the error of their ways and become more open about their dealings with the rest of the world? Or that governments around the world — and in particular the United States government — clamp down hard on classified information and restrict its distribution even more than they have in the past?

## Bombing Iran

| Sun Nov. 28, 2010 6:52 PM EST

I have a question. Several people are suggesting that the most interesting/damaging part of the WikiLeaks embassy cable dump is the revelation that Saudi Arabia’s King Abdullah has repeatedly asked the United States to bomb Iran. Dave Schuler comments:

It isn’t just King Abdullah—the rulers of Jordan, Bahrain, and Abu Dhabi have apparently all made similar requests and leaders in Egypt have stopped just short of that.

IMO were Israel or the United States to eliminate Iran’s nuclear development capability by force Iran’s neighbors would make some outraged noises while being secretly relieved. These leaks have removed the possibility of cloaking their hostility with such a figleaf from the Sunni regimes of the Middle East.

Here's my question: is this really news? I thought it was common knowledge that most of the Gulf states felt this way. It's obviously true that "common knowledge" isn't the same thing as a bunch of diplomatic cables that confirms this stuff to the world, but still. Everyone seems to have known this already, presumably including the Iranian leadership.

Am I off base here? It seems like I've heard this so many times that it didn't even register as something newsworthy to me. But maybe I'm missing something here.

POSTSCRIPT: Perhaps I'm just being too America-centric. Issandr El Amrani at The Arabist acknowledges that many of the cables "just confirm certain widely held theories," but nonetheless thinks the diplomatic damage will be huge:

There is so much information flowing around about US policy — and often, a good deal of transparency — that a smart observer with good contacts can get a good idea of what's happening. Not so in the Arab world, and the contents of the conversations Arab leader are having with their patron state are not out in the Arab public domain or easily guessable, as anyone who reads the meaningless press statements of government press agencies will tell you. Cablegate is in important record from the Arab perspective, perhaps more than from the US one.

So: more important to the Arab world than to us. Maybe so. In fact, probably so.

## Yet More American Secrets No Longer Secret

| Sun Nov. 28, 2010 6:22 PM EST

The WikiLeaks release of 251,000 U.S. embassy cables is, I'm told, a "diplomatic crisis." Hillary Clinton is running point on damage control. Foreign allies are on high alert. Life will never be the same. So what's in those cables? Here's the Guardian's bullet list:

Grave fears in Washington and London over the security of Pakistan's nuclear weapons programme.... Suspicions of corruption in the Afghan government....The extraordinarily close relationship between Vladimir Putin, the Russian prime minister, and Silvio Berlusconi, the Italian prime minister....Allegations that Russia and its intelligence agencies are using mafia bosses to carry out criminal operations....Inappropriate remarks by a member of the British royal family.

Say it ain't so! Corruption in the Afghan government? Silvio Berlusconi is a douche? Prince Andrew said something naughty? Pass the smelling salts. But maybe the New York Times has more. Here's their list:

A dangerous standoff with Pakistan over nuclear fuel....Gaming out an eventual collapse of North Korea....Bargaining to empty the Guantánamo Bay prison....Suspicions of corruption in the Afghan government....Mixed records against terrorism....Arms deliveries to militants....Clashes with Europe over human rights.

Oh, and the Saudis would like us to bomb Iran, thankyouverymuch. Yawn. Even Der Spiegel, which excels at finding the most sensational spin possible on this kind of thing, was pretty much stumped:

What, though, do the thousands of documents prove? Do they really show a US which has the world on a leash? Are Washington's embassies still self-contained power centers in their host countries?

In sum, probably not. In the major crisis regions, an image emerges of a superpower that can no longer truly be certain of its allies — like in Pakistan, where the Americans are consumed by fear that the unstable nuclear power could become precisely the place where terrorists obtain dangerous nuclear material.

Andrew Sullivan actually thought the cables showed that the State Department is on the ball: "Overall, I have to say that this brief glimpse into how the government actually works is actually reassuring. The cable extracts are often sharp, smart, candid and penetrating. Who knew the US government had so many talented diplomats?"

Now, I did leave a few things out. There's some interesting stuff about the Chinese Google hack, some frank and candid military assessments of British troops in Afghanistan, a bit of Israeli bluster about bombing Iran (though it's nearly identical to Jeffrey Goldberg's very public piece in the Atlantic a couple of months ago), and confirmation that embassy officials often try to spy on people in their host countries. And the Turkish government is probably going to be pretty pissed at us for a while. Still, this is hardly sensational stuff. In fact, what's really struck me so far is how little our diplomats talk out of school in private cables.

Maybe there's more to come on this,1 but so far I just don't see these leaks causing an epic amount of damage. Obviously feelings will be bruised by the blunt language in some of the cables — though if Spiegel's excerpts are typical, the language is only slightly blunter than your run-of-the-mill anonymous carping — and foreign officials might be reluctant for a while to share confidences with American diplomats. And just as obviously, the United States would really prefer that its confidential cables remain confidential. Hillary Clinton will indeed have her hands full for a while. But honestly, there's hardly anything here that I haven't already read on the front pages of multiple newspapers. Titillating, but not much more.

1In fact, Blake Hounshell tweets: "Anyone who thinks this batch of WikiLeaks docs is not interesting clearly isn't reading them." Marc Lynch agrees: "Guardian + NYT undersold them." So maybe there really is more to come. The three English-language accounts, however, just don't make this stuff seem especially earthshaking.

## Paying the Piper

| Sun Nov. 28, 2010 1:47 AM EST

Bruce Bartlett:

A prime reason why we have a budget deficit problem in this country is because Republicans almost universally believe in a nonsensical idea called starve the beast (STB). By this theory, the one and only thing they need to do to be fiscally responsible is to cut taxes. They need not lift a finger to cut spending because it will magically come down, just as a child will reduce her spending if her allowance is cut — the precise analogy used by Ronald Reagan to defend this doctrine in a Feb. 5, 1981, address to the nation.

Bruce goes on to look at the empirical evidence — namely that spending went down after the Clinton tax increases and up after the Bush tax cuts — and concludes that STB is a "crackpot theory." True! But what makes it even more crackpotty is that basic economic principles, of the kind that Republicans are endlessly lecturing the rest of us about, predict the same thing. If you raise taxes to pay for government programs, you're essentially making them expensive. Conversely, if you cut taxes, you're making government spending cheaper. So what does Econ 101 say happens when you reduce the price of something? Answer: demand for it goes up.

Cutting taxes makes government spending less expensive for taxpayers, which makes them want more of it. And politicians, obliging creatures that they are, are eager to give the people what they want. Result: lots of spending and lots of deficits.

If you want to reduce spending, the best way to do it is to raise taxes so that registered voters actually have to pay for the services they get. I don't have a cute name for this theory, but it's true nonetheless. Even for Republicans.

## Our Boring Future

| Sun Nov. 28, 2010 1:15 AM EST

Nassim Nicholas Taleb thinks the nation state is fated to disappear over the next couple of decades, to be replaced by "city-states and statelings" that rely on a gold standard and can manage their finances properly. Matt Yglesias is skeptical:

Maybe so. And yet it seems to me that people have been predicting the nation-state’s demise for a long time and it seems like a very robust structure. If anything the trend I see toward greater adherence to a strict interpretation of what a nation-state is supposed to be. Belgium splitting in into two properly “national” states seems much more plausible than Los Angeles emerging as a quasi-sovereign entity.

Yeah, I don't think LA is quite destined for national greatness yet. Ditto for the idea that our current recession spells some kind of permanent change in "consumerism" and spending habits. I know this kind of thing sounds cool, but it's really unlikely that even a big global recession is going to fundamentally change either the course of human history or the current state of the art in human nature.

Our problems today may loom large, but they're also quite solvable. True, the solutions involve a fair amount of nonheroic drudgery, and that's not very much fun to write about, but it's a whole lot more likely to improve actual human lives. Noses to the grindstone, folks.

## Hard Truths on Afghanistan

| Sat Nov. 27, 2010 1:40 PM EST

I don't remember where I first saw this, but Ahmed Rashid's two-hour interview with Afghan president Hamid Karzai, someone he's known for 26 years, is essential reading:

Afghan president Hamid Karzai is a changed man. His worldview now is decidedly anti-Western....By the end of our talk, it was quite clear to me that his views on global events, on the future course of NATO’s military surge in southern Afghanistan, and on nation building efforts throughout his country have undergone a sea change. His single overriding aim now is making peace with the Taliban and ending the war—and he is convinced it will help resolve all the other problems he faces, such as corruption, bad governance, and the lack of an administration.

....He no longer supports the war on terrorism as defined by Washington and says that the current military surge in the south by the United States and its NATO allies is unhelpful....Karzai also maintains that there is a political alternative to NATO: much more of the onus could be placed on countries in the region—especially Iran and Pakistan—to end the war and help reach a settlement with the Taliban. Senior Western and Afghan officials in Kabul say Iran has stepped up its support to the Taliban in western Afghanistan in recent months, possibly as a bargaining chip for future talks on a peace settlement. For its part, Pakistan, where the entire leadership of the Taliban is based, wants a leading part in any talks that NATO or Karzai may have with the Taliban. Yet Karzai told me that in the last six months neither Iran nor Pakistan has provided any substantive support to facilitate peacemaking.

There have been several reports recently suggesting that Karzai has given up on the U.S., followed by other reports that, no, he really hasn't. But Rashid's interview, which is the deepest and clearest that I've seen, seems to confirm that the earlier reports were the real deal. Karzai really has definitively wearied of the U.S. presence and really would like us to leave.

He hasn't insisted, of course, because his government would almost certainly collapse within weeks or months if NATO weren't around to prop it up. Beyond that obvious reality, there's also an odd strain of delusion here that I'm surprised Rashid didn't follow up on: namely Karzai's contention that Iran and Pakistan should help end the war and reach a settlement with the Taliban. That may be true, but as Karzai himself points out, neither country appears to have any serious motivation to do so. Apparently he thinks Iran and Pakistan could somehow take NATO's place, even though he acknowledges that neither has been helpful, and neither really shows any signs of being helpful in the future.

From the U.S. point of view, of course, they key thing isn't whether Karzai is tired or delusional or getting bad advice. What really matters is that over the past year he's apparently come to the firm conclusion that a continued U.S. presence is unhelpful. This pretty plainly makes our military efforts in Afghanistan pointless. As Gen. Petraeus and his counterinsurgency gurus continually tell us, political support is crucial to eventual success. If we don't have it — and it's now about as clear as it can be that we don't — then all the Lisbon conferences in the world won't produce a plan for victory. It's about time for Barack Obama to start leveling with the American public about this.

## The Volt and You

| Fri Nov. 26, 2010 5:54 PM EST

The EPA has released its official mileage ratings for the Chevy Volt, and Dodd is unimpressed:

The woefully limited 35-mile range on battery and mere 37 MPG on gas leaves one wondering what all that hype was really about.

I don't actually care about the Volt all that much, but this is a really common reaction and seems completely misguided to me. No car is designed to appeal to every single person, and the Volt is no exception. It's designed mostly to appeal to a specific kind of driver: someone who does the great bulk of their driving around town, maybe 20 or 30 miles a day at most, but occasionally needs to drive further and doesn't want to buy a second car just for those occasions. There are lots of people like that, and for them the Volt is great. They'll spend 98% of their time running solely on battery power and recharging at night when rates are low, and 2% of their time getting 37 mpg — which is actually pretty damn good. There are a few hybrids that do slightly better and one hybrid (the Prius) that does a lot better, and that's about it.

If you commute a hundred miles a day, the Volt isn't for you. If you're a traveling salesman, it's not for you. If you need to haul around a Boy Scout troop, it's not for you. If you need lots of towing capacity, it's not for you. But that's not a problem. It's not supposed to be for you. It's for people who drive ten miles to work each day, run some errands on the weekend, and drive out to grandma's house once a month. Those folks are going to get pretty awesome fuel efficiency, and they're going to get it with just one car. What's not to like?1

1Answer: the price tag. That's really the car's only serious Achilles' heel. Even after the government rebate, it'll run you around \$33,000 for a car that would cost less than \$20,000 with a standard engine. Until the price of the car comes down, it's going to be a tough sell for anyone who's not dazzled by its eco-friendliness.