Cross sectional reasoning

I want to follow up on a post by Bad Astronomy that has been discussed by both Mark and Noah Smith. The post comments on the results of the 2010 census on employment rates:

I highlighted one in particular: Astronomy and Astrophysics. Note that it has a 0% unemployment rate; in other words, last year everyone who majored in these fields got a job! Now, I find myself being a tad skeptical about this, but if there’s some weird thing going on with this survey, I can at least make the broad assumption that the relative job numbers are probably OK. Majoring in astronomy is still a good idea, and will strengthen your chances of getting a job after college.

I want to take this in a different direction. What this metric shows is that, if you were lucky enough to have majored in Astronomy in 2005 then you were very likely to be employed in 2010. It says nothing about what will happen to somebody who enters the program in 2011 and whether they will be employed in 2015.

See, I was actually a physics major in the mid 1990's, in a school with a large astrophysics group. I knew a lot of these students and even took classes with them. Do you know what they mostly ended up as: High School Teachers. Plus a few academics. At the time there was a terrible job placement rate in physics and we were all depressed by the poor employment outcomes. Using the tool, I see a 4.5% unemployment rate for physics, which does make me wonder how many astrophysicists are counted in this group.

But, in general, past performance is no guarantee of future employment. A depressed job market could easily have led to full employment years later, long after only the most dedicated students remained. I've seen this phenomenon in a lot of fields -- people go where the markets signal but, in education, the signals are lagging indicators.

So maybe we are seeing the unemployment ghettos of the future?

Another "It's too late tonight to do it justice..."

Noah Smith has an excellent post on the employment picture for science major with important implications for our ongoing mathematics education debate:

No, a science major does not guarantee you a job...

I'm not knowledgeable enough to sufficiently mock this* #

DNA Testing Could Help Choose Your Kid's Sport





* Yeah, it's a split infinitive. What are you, the grammar police?

# Now with no words omitted from the title.

Libertarians in Space

Herman Cain gave a speech to a group of Young Republicans and the subject of the space shuttle came up:

In his speech, Cain praised President John F. Kennedy as a "great leader" for inspiring a national effort to put a man on the moon, a goal achieved when astronaut Neil Armstong stepped onto the moon's surface in 1969.

"He didn't say, 'We might.' He didn't say, 'Let's take a poll,'" Cain said. "He said, 'We will.' And we did. Only for this president to move us back by canceling a major part of our space program."

Cain also criticized Obama for using Russian technology to ferry astronauts and cargo to the International Space Station.

"I can tell you that as president of the United States, we are not going to bum a ride to outer space with Russia," Cain said to loud applause. "We're going to regain our rightful place in terms of technology, space technology."

I don't know what the reaction of the crowd was (the reporting wasn't that detailed) but I'd imagine it was friendly. You can usually get a warm response from a Republican crowd by coming out in favor of manned space exploration which is, when you think about, strange as hell.

If you set out to genetically engineer a program that libertarians ought to object to, you'd probably come up with something like the manned space program. A massive government initiative, tremendously expensive, with no real role for individual initiative. Compared to infrastructure projects the benefits to business are limited. You could even argue that the government's presence in the field crowds out private development.

(Much has been made of the rise of private space firms, but barring a really big and unexpected technological advance, their role is going to be limited to either unmanned missions or human flights in low earth orbit for the foreseeable future.)

There have been efforts in libertarian-leaning organs (The Wall Street Journal, Reason, John Tierney's NYT columns) trying to argue that interplanetary exploration can be done on the cheap. These usually rely heavily on the blatant low-balling of Robert Zubrin* (Tierney, a science writer who has no grasp of science, made a particularly ripe mark), but even if we were to accept these numbers, it's still difficult to reconcile this kind of government program with libertarian values.


* On a related note, check out this other example of Zubrin's estimation skills.

Tex Avery animates David Frum's sudden move to the center

At about ninety seconds in.

Pre-blogging Jack Shafer

I've got a couple of posts coming up on related subjects so I thought I'd get this out of the way in advance.

Here's Jack Shafer explaining how his intense antipathy toward plagiarism is all about the readers (via Salmon):

The plagiarist defrauds readers by leading them to believe that he has come by the facts of his story first-hand–that he vouches for the accuracy of the facts and interpretations under his byline. But this is not the case. Generally, the plagiarist doesn’t know whether the copy he’s lifted has gotten the story right because he hasn’t really investigated the topic. (If he had, he could write the story himself.) In such cases he must attribute the material he borrows so that at the very least the reader can hold somebody accountable for the facts in a story.

Putting aside the fact that Shafer has never gotten that worked up about colleagues' inaccuracy (you'll notice he didn't jump on this story about his friend Gregg Easterbrook), there's an interesting game here of rhetorical Three-card Monte. The cards in this game are the three types of plagiarism: theft of ideas and interpretations; theft of facts and data; and theft of wording.

In the world of research, the first two are considered the most serious. Stealing the hypothesis of another paper or presenting someone else's data as your own is about the worst thing you can do. Lifting a passage of someone else's writing is frowned upon but prose style does not drive impact factors.

For journalists, the situation is exactly reversed. Reusing another writer's phrases is clearly considered the worst kind of plagiarism, perhaps the worst journalistic crime period. Stealing facts (such as using other people's reporting to cover an event) is seldom even mentioned except in the most flagrant of cases. As for appropriating ideas, the practice is so common as to almost be standard. Even those most modern of journalistic concepts, memes, are almost always based largely on the plagiarism of hypotheses and arguments.

Now, here's the part where the cards really start to move. Shafer's criticism only applies to the first two types of plagiarism, the two types he doesn't object to. (aren't the first two nested in the third? Sometimes, but Shafer apparently doesn't have a problem with borrowing and paraphrasing, it must be the not-paraphrasing part that bothers him). If Shafer really wants to convince us that borrowing without paraphrasing is more than journalist-on-journalist crime, he'll have to do better than that.

Health Care Costs

We have known for a while that the United States spends a lot on health care. What is interesting is that the direction seems to be moving in the wrong way both in terms of life expectancy and cost:

You can see that not only is the United States the outlier when it comes to spending, but we are moving in the wrong direction: we are becoming more of a spending outlier, and we are drifting down from the average life expectancy into the lower group (currently surpassing only Turkey, Hungary, Mexico, Poland, and Czech Republic).

and

The other thing you see is that our life expectancy gain was the absolute lowest of the whole group (and we weren’t starting from a particularly high level, as you can see in the previous chart).

Ordinarily, you would think there should be convergence across countries. Since other countries spend less and live longer, you would think that we would learn from them—global competition, you know. But instead we’re moving the wrong way on both dimensions.

The article and neat charts are worth looking at in their entirety.

Now, it is true that there can be a lot of reasons for low life expectancy and high medical costs. It could be that the environment in the United States makes us much more accident prone, for example, requiring both higher spending and more fatalities.

But, in general, it is uncomfortable when the most important metric of health care outcomes (all cause mortality) is so uncorrelated with cost. This suggests the possibility of productivity improvements. I read a lot of the Incidental Economist, who try to explain these issues. But I admit that I tend to come away confused.

The major comparison is often Canada. It is a bad reference on a lot of levels (as they have their own issues). But they have similar culture, ethnic diversity, large geography, heavy use of cars, high levels of obesity and yet they are improving on both metrics (from a lower level of cost and higher life expectancy at baseline).

Why is health care the one area that we aren't willing to look at how other countries have been successful and try to steal ideas?

Old story, different hemisphere

More good work from American Public Media's Marketplace:

Rob Schmitz: If you walk down the street in Shanghai, it's hard to miss the advertisements for college placement agencies. In big bold letters, they promise -- some even guarantee -- your child's admission to an American university. The price: $5,000, $6,000, sometimes $7,000.

Jiang Xueqin is a well-known education reformer in China. He also heads the International division at Peking University High School -- one of China's best high schools. He has followed college placement agencies for years. He says this is how many of them deliver on their admission promises: they falsify students application materials.

Jiang Xueqin: There's a lot of pressure on the agencies to write the application essays, to fake transcripts, to fake recommendation letters. This is just general business practice in China to falsify a lot of documentation.

A report by consulting firm Zinch China seems to confirm this. Zinch advises American colleges and universities on recruiting Chinese students. The firm interviewed agents and admissions consultants, as well as more than 200 Beijing students headed to U.S. schools. Zinch estimates 90 percent of these students submitted false recommendation letters; 70 percent had other people write their personal essays, and half of them submitted forged high school transcripts.

Werner Herzog, LA, and structural employment in hot jazz

The director had an interesting interview on KCRW's none-too-imaginatively named show on the business of Hollywood. The part that particularly caught ear occurred about nineteen minutes in:

"We [Herzog and his wife] have to move into the city with the most substance in the United States, cultural substance."
"And that wasn't New York..."
"New York is more on the receiving end, it's more on the consuming end. ... Los Angeles, this is a place where things get done."

This jibes with something I've noticed since moving to LA and falling in (through no fault of my own) with the hot jazz scene in town. There are some extraordinary musicians around here, the kind who develop international followings and get glowing notices in the New York Times.

LA has a wealth of musical talent, but if those musicians want any wealth for themselves they have to head east. Because, though you might argue that the West Coast has the best musicians, the East Coast unquestionably has the best fans, at least when it comes to ponying up. In NYC these musician actually make good money; in LA you can often find them working for kind words and Pez.

If you're looking for creative energy, I'm with Herzog -- you can't do better than LA. If you're looking for a market for all that creativity, you might want to try the other coast.

But some of my best friends are administrators

As I've said before, most principals and superintendents I've dealt with have been dedicated and highly professional educators, so why do I keep picking on them? While it's possible that I do have a small problem with authority, my motives here are mostly pure. I'm trying to correct a cherished but dangerous piece of misinformation embraced by movement reformers.

For all their admirable traits, we have to acknowledge the world administrators live in: they are generally at least one step removed from the classroom and regular contact with students; their jobs are highly paid and highly political with badly aligned incentives and considerable temptations to abuse the system. Movement reformers like Jonathan Chait have convinced themselves that a major key (perhaps the key) to improving education is removing checks on administrators' power and upping rewards for manipulating the system. I'm just pointing out that some administrators need less power and fewer temptations.

I haven't taken a cheap shot at an administrator for a while

And superintendent Eric Ely certainly has one coming. He's certainly not the main villain in this amazing story from This American Life, but he clearly did his share to allow this psycho* to terrorize a number of people and endanger students, and while that character ended up in prison (probably for the rest of his life), Ely landed in another cushy superintendent's job.






* think this is hyperbole? Listen to the episode.

Speaking of curriculum -- Armenian edition

From NPR's The World:

Armenia’s public schools started mandatory chess classes for every second, third and fourth grader in the former Soviet republic this year. Twice a week, seven- to ten-year-old Armenians are getting a half hour of instruction in chess basics, with the goal of being able to play a competent game by the end of fourth grade.

One of those schools is Public School 81 where Grigor Martikian is drilling 20 second graders on how to move the bishop. He positions a bishop into the corner of a large model board in front of the class.

The students follow along on chess mats on their desks.

Edouard Aroustian and Seta Kevorkian are both seven and have learned a new checkmate move. When I ask them if they know about Armenia’s national chess team, they nod and smile.

That team was the pride of Armenia this past summer, when it won the World Chess Team Championship in Ningbo, China. Armenians treat chess champions like star athletes. Chess is one of the most popular games here and there are 32 grandmasters in a population of about 3 million.

If only I could talk them into trying octagonal games...

At least we aren't giving them antidepressants

From NPR:

Babies have been crying and spitting up since time immemorial. But these days many parents ask: Isn't there a drug for that?

"Parents come in often demanding medication," says Eric Hassall, a pediatric gastroenterologist at Sutter Pacific Medical Foundation in San Francisco.

Prescriptions for acid-suppressing medicines for infants have increased dramatically. Hassall says some parents have picked up on the idea that heavily advertised medicines for reflux in adults can help fussy babies who spit up a lot.

He documented a 16-fold increase in prescriptions of one proton pump inhibitor, or PPI, Prevacid, which comes in a child-friendly formulation. A Food and Drug Administration review also found an 11-fold increase in number of new prescriptions dispensed between 2002 and 2009.

These medicines aren't approved for infants with reflux, or GERD. Still, some doctors have been prescribing them off-label anyway. Doctors generally agree this practice is OK when babies really need the medicine, such as when they're spitting up so much that they're not gaining weight.

One of the risks of living is dying

From the incidental economist:

I can hear the howls of protest already. But here’s the example I always go to: the number one killer of children in the US is car accidents. But we don’t ever consider stopping driving. I know that every time I put my kids in a car, I’m significantly increasing the chance that they could die. But I (and pretty much all of you) believe that the benefit to our lives from cars outweighs the increased risk of death in our children. Let me put it another way. We all accept that it’s worth a number of children dying so that we can all get around more easily.

One of the great challenges that we face as a society is how to balance risks and benefits. I am becoming increasingly convinced that people are simply poor at making these trade-offs. This is especially true given that the risk of death is 100%. In a real sense we all end up dying. The goal, instead, seems to be to make the time that we have as good as possible. A theory of the joint maximization of lifespan and happiness seems to be the best way to go.

Given that, I think Dr Carroll's point is quite sound: we take risks all of the time in order to make life worth living. The trick is to quantify which risks are worthwhile, conditional on the absolute level of risk.

The worst example of curriculum dead wood?

One of the first things that hit me when I started teaching high school math was how much material there was to cover. There was no slack, no real time to slow down when students were having trouble. The most annoying part, though, was the number of topics that could easily have been cut, thus giving the students the time to master the important skills and concepts.

The example that really stuck with me was synthetic division, a more concise but less intuitive way of performing polynomial long division. Both of these topics are pretty much useless in daily life but polynomial long division does, at least, give the student some insight into the relationship between polynomials and familiar base-ten numbers. Synthetic division has no such value; it's just a faster but less interesting way of doing something you'll never have to do.

I started asking hardcore math people -- mathematicians, statisticians, physicists, rocket scientists*-- if they'd ever used synthetic division. By an overwhelming margin, the answer I got was "what's synthetic division?" Not only did they not need it; it made so little impression that they forgot ever learning it.

Which bring us to this passage from a recent Dana Goldstein post (discussed earlier):

The problem, according to [David] Coleman, is that American curriculum standards have traditionally been written by committees whose members advocate for their pet pedagogical theories, such as traditional vs. reform math.

Except, of course, that's not what happened here. As was the case with so many topics in mathematics, synthetic division remained in the curriculum because no one who knew what was going on had bothered to look that closely. Coleman has a clever narrative, but it doesn't fit the facts all that well.

Now I have a request for all the math geeks in the audience (and given that you're reading a blog called Observational Epidemiology...). Since we need to pare down the curriculum, what you choose to cut? Specifically, what mathematical topics that you learned in school can future generations do without?


* Literal rocket scientists -- JPL's just down the road.

Also posted in a slightly different form at Education and Statistics.