Sunday, August 30, 2009

Further Reflections on "Let's Make a Deal"

I recently (August 25) alluded to the "Let's Make a Deal" problem (also known as the Monty Hall problem). It goes like this: There are three doors. Behind two of the doors is a worthless prize and behind one a valuable prize. The Wikipedia article linked above refers to these as goats and a car, respectively. The contestant chooses one door, and then Monty Hall, who knows what is behind each  door, opens one of the other doors and shows that there's a goat behind it. He always shows a goat, never the car. He then offers the contestant the choice of sticking with the door previously chosen, or switching to the other remaining door. The question is, should you switch? Everyone's intuition is that there's no reason to switch-- we already knew that one of the other doors had a goat, so we haven't learned anything. This is wrong. You have a 2/3 chance of winning the car by switching.

When Marilyn Vos Savant published this in Parade magazine, it created a huge uproar, with tenured professors of mathematics writing to tell her she was wrong, but she stuck to her guns. (There are a number of different ways of explaining why she's right, no one of which works for everyone. I'll post some if there's demand.)

But what recently struck me is how unusual this is among cases where our intuitions lead us astray in statistics. Typically, the mistake we make is in seeing things as non-random that are actually random. Public-policy examples include cancer clusters and the general problem of whether some policy or treatment is more effective than another. Our natural tendency is to try to find patterns in things, whether they exist or not.

What makes this case unusual is that it's one where our intuitions would be correct if something that is not random actually were random. If Monty Hall chose the second door randomly, and it had a goat, then the probability of the other doors having a car would be 1/2 for each, and there would be no reason to switch. (Of course, a third of the time he'd expose the car, which would tend to kill the suspense.) But somehow our intuitions don't take account of the fact that Monty knows what's behind the doors, even though it's no secret. I don't know of any psychological research on errors like this.

Contest: Public-policy examples of the second type of mistake. Prize: kudos only.

Tuesday, August 25, 2009

Everybody's a Critic

From an very disturbing discussion, in a New York Times blog, of an article by Jane Mayer (The New Yorker) about the adaptation of the military's Survival Evasion Resistance and Escape program to CIA interrogations:

The SERE affiliate told me that trainees often think that the interrogation portion of the program will be the most grueling, but in fact for many trainees the worst moment is when they are made to listen to taped loops of cacophonous sounds. One of the most stress-inducing tapes is a recording of babies crying inconsolably. Another is a Yoko Ono album.

The Asteroid/Swine Flu Problem

I recently saw on TV that NASA says it doesn't have enough money to track all the large asteroids that might endanger Earth. I don't know whether they do or not, but it does raise the general problem of how we deal with unlikely but potentially catastrophic events, why we're so bad at it, and whether anything can be done about it.

In my Master's program we had a case class about the first swine flu scare, the one that happened under the Ford administration. The teacher wanted to look at why the Ford administration had overestimated the probability of an outbreak, and hence blundered by launching a huge vaccination program that turned into a fiasco. One student made what I thought (and we probably all thought) was a reasonable point: what if there were only a 10% chance of a Spanish-flu-type outbreak that would kill millions? Wouldn't it still be worthwhile to start vaccinating? The teacher, a very experienced public manager, was incredulous that anyone would think of doing anything in the public sector based on only a 10% probability.

OK, so what if there is a 10% chance of an asteroid strike that will wipe out civilization?

There are three nested problems here. On the individual level, humans are bad at dealing with probabilities. (I recently read that Paul Erdos, one of the greatest mathematicians of the 20th century, never understood the Let's Make a Deal problem made famous by Marilyn Vos Savant.) At times, in fact, it seems we don't really get the whole idea of probabilities. Witness the global warming debate, in which the argument often seems to be over whether it's certain or not, with the implicit assumption being that if it's not, we should do nothing. I have never heard an environmentalist try to make the argument that we should act even if we're not sure.

On the institutional level (say, Congress), politicians and bureaucrats understand one implication of our discomfort with probabilities other than zero and one: that no one will believe a decision was a good one if it doesn't turn out right. That is, they will be blamed for investing in averting a disaster that doesn't materialize (as Ford was, though Guillain-Barre syndrome also played a role). Imagine being a bureaucrat at a hearing, trying to explain to a committee of senators why the taxpayers' money wasn't wasted even though nothing happened. The senators, meanwhile, are imagining themselves trying to explain the same thing to their constituents.

At the collective level (say, international), there are the usual collective-action problems familiar to economists and political scientists. Even though everyone may be convinced that an asteroid threat is real and imminent, they will still all prefer that someone else bear the costs. This can become a game of Chicken, and of course games of Chicken often turn out badly (the classic popular-culture examples are Rebel Without a Cause and Dr. Strangelove). The problem is further complicated by asymmetries, as the smallest players have the greatest tendency to free-ride (see OPEC, for example).

I haven't got much to say about the collective action problems, as so much has been said on international negotiations by people who have thought about them a lot more than I have. The individual-level problem, the tendency of individuals to ignore small risks, may to some extent be offset by their tendency to overestimate them when spectacular and widely reported, as with airline crashes. (There is a psychological literature on this.) Still, one hesitates to recommend that policy analysts try to induce public panic.

Can anything be done about the institutional problems? One of my usual suggestions in such cases, to take them out of the regular political process, doesn't seem very practical here-- I can't imagine how one would go about setting up an independent Catastrophes Authority, given that the next threat could be astronomical or biological or meteorological or military or...

One interesting possibility would be to eliminate second-guessing of bureaucrats and politicians by forcing voters to make the decision themselves. Unfortunately, direct democracy has a pretty dismal record in the United States, with money ending up playing a dominant role. Why it seems to work fairly well in Switzerland is an important question that I don't know the answer to. I wish I had spent more time asking when I was there.

Thursday, August 20, 2009

The R in English Literature

In the 1980s there was a British/Nigerian singer named Sade. A number of people informed me that her name was pronounced Shar-day. I didn't believe it. No one whose name is pronounced Shar-day is going to spell it Sade. Just ignore the R? Finally I figured it out: this was a British rendition of the pronunciation Shah-day. If you say Shar-day with pretty much any British accent, it comes out Shah-day.

It took me a long time, though, to understand just how widespread this British approach to transcription is, and how much confusion it creates for Americans. The next time it clicked for me was thinking about a passage that had puzzled me since childhood:

"But I thought he was a boy?"
"So did I," said Christopher Robin.
"Then you can't call him Winnie?"

"I don't."
"But you said--"
"He's Winnie-ther-Pooh. Don't you know what '
ther' means?

If you're an American, of course, you have no idea what ''ther" means. But now we see that Milne was trying to show that the word was pronounced "the" and not "thee."

Examples start to come thick and fast after a while. Why did P.G. Wodehouse, who spent a lot of time in the US, have his American characters say, "You betcher!" No American ever says "You betcher!" And neither do these; they're really saying "You betcha!" And why do people say "Er..." in books when they interrupt someone? I've never heard anyone say "Er..." But in fact they're making a sound fairly close to what would be rendered in American English as "Uh..." or "Unh..."

Some jokes don't come off as well, either, if you don't know the secret. When Alice asks why they called their teacher the tortoise if he wasn't one, the Mock Turtle says, "We called him the tortoise because he taught us, of course!" Pretty lame with a Midwestern accent. If you have a British accent, though (or a strong New England or Bronx or Southern accent), those words are nearly homophones.

Just today, I woke up from a dream in which I was playing Tevye the dairyman on stage (and also directing, if you're interested) and didn't have my costume ready. And I suddenly understood how the word "yarmulke" got its spelling. (I think the "l" is also supposed to be silent.)

This still leaves the question, why do they do it? Why don't the British just write "shah-day" and "You betcha!" Beats me. Why do they drive on the left?


Tuesday, August 18, 2009

Obama's Cairo speech: The missing page

Sheet of paper found on the floor of the auditorium at Cairo University:

"We hear a great deal these days, from those who want to justify terror, about a 'war against Islam,' supposedly led by the United States. I find this puzzling. When the Muslims of Bosnia were under attack from Christians, who saved them? It was not, certainly, the jihad crowd, who may have had a good time killing people but had no discernible effect on the outcome. The war was stopped, and with it the killing of Bosnian Muslims, by the intervention, including military, of the United States. And a few years later, when NATO went to war for the first and only time in its history, it was against a Christian country, to stop them attacking Muslims.

"Why did we do this ? Because we like Muslims more than Christians? No. We did it because we believe that human rights apply to everyone, no matter what their religion is.

"So let us dismiss the notion that there is a war against Islam. But while there is no war against Islam, in many places Muslims are under attack. In Iraq, for example, tens of thousands of noncombatants have been killed. But by whom? Not by Americans-- by Iraqis! Tens of thousands of Muslims have been killed by Muslims. Indeed, the allegedly anti-Muslim invasion of Iraq was aimed at overthrowing the man responsible for the deaths of more Muslims than anyone in history.

"Or consider another conflict, one that has been going on for over a decade. This conflict has resulted in over a million Muslims being displaced, hundreds of thousands being killed, and hundreds of thousands of Muslim women being raped. Hundreds of thousands of Muslim women being raped! Yet in the Muslim world apparently no one cares. Americans care, Europeans care, but the Muslim world has been silent, if not worse. Apparently the attitude in the Muslim world is, 'Displacing, killing, raping Muslims is perfectly fine, as long as we do it.' This attitude is very difficult for Americans to understand."


Monday, August 17, 2009

EZ Health Policy Analysis II

This doesn't even require long division, just Google.


World rank of United States in life expectancy: (CIA World Factbook, 2009 est. Includes only independent countries.)

36


World rank of the United States in infant mortality, 1 lowest: (see above)

36


World rank of the United States in health spending as a percentage of GDP, 2005: (Statistical Abstract of the United States)

1

Friday, August 14, 2009

Dr. Frant's EZ Health Policy Analysis

One of my mentors, Richard Zeckhauser, says that one of the most powerful tools in policy analysis is long division. Let's see if he's right.

1. Approximate cost of the health insurance plans now under consideration in Congress:

$1 trillion over 10 years

2. Average annual cost of (1)

$100 billion

3. U.S. GDP in 2008 (U.S. Bureau of Economic Analysis)

$14.441 trillion

4. Average annual cost of plans as a percentage of U.S. GDP in 2008

0.7%

5. Total U.S. expenditures on health care in 2008 (projection, U.S. Center for Medicare and Medicaid Services)

$2.4 trillion

6. Average annual cost of plans as a percentage of 2008 U.S. expenditures on health care

4.1%

7. Expenditures by pharmaceutical companies on drug promotion, 2004 (PLoS Medicine 5(1) (2008), Gagnon, M.A. et al.)

$57.5 billion

8. Average annual cost of plans as a multiple of 2004 drug promotion spending by drug companies

1.74 times


Yes, that's right. The staggering bill for Congressional proposals on health care is 4% of what we're spending now on health care, or less than twice what the pharmaceutical industry spends on drug promotion. If we had more current figures for the denominators, the percentages would be even smaller.

Welcome

This blog will be mainly about things related to public policy and politics, though it will include other topics that I find interesting. Please be aware that I find many things interesting. I will try, though, to limit postings to things I think will be of wide interest; if you want to know what I had for breakfast I am happy to tell you, but please ask by e-mail.

I am calling this blog "Claritas" because I think that clarity is in short supply in discussions of public policy (and, indeed, most things), and because this is an area where I think I can make a contribution. If I fail, please tell me so I can try again.