The Hanover Collaboration

So another study came out about parenting. It’s times like these when our inability to talk precisely about data gets pretty painful. Too bad the study itself is pay-only. Here’s a paragraph from Slate where Emily Bazelon is trying to say that all the major media outlets are misreading it.

The source of the fuss is the latest installment of a long-running $200 million effort by the National Institute of Child Health and Human Development. Since 1991, a team of researchers has been tracking more than 1,300 children, following them from infancy through various child-care settings (home with mother, home with another relative, home with nanny, or at day care) and into elementary school. In the March/April issue of Child Development, the team asks “Are There Long-Term Effects of Early Child Care?” To answer that question, the researchers report their findings about the kids’ academic achievement and behavior through sixth grade. The study controls for a host of variables, like socioeconomic status, quality of parenting (annoyingly, this measure involves only mothers), quality of child care, and quality of the elementary-school classroom. It’s all very well-done and careful.

My problem here is “controls for”. If the study had a million kids in it, you could probably “control for” perhaps half of those things. I’m pretty sure she means “collected data on”. On the second page she expounds on how the study failed to control for the quality of the child care.

Does that distinction matter? Absolutely. You have 1300 kids, some do better on tests, some get in trouble, some went to day care, some watched baby Einstien. It’s easy to say that two variables are correlated. You just ignore all the other data. But what you would like to say is that the correlation is not an artifact of some other relationship. This requires taking all the rest of the data into account—that we “control” for those other variables.

The problem with that is it requires having enough data to show a valid relationship for every possible combination of every value of the variables being controlled for. Clearly that is impossible with a sample that small. I’d guess what they did is they fit linear relationships and then asked whether there were correlations in the residuals. I won’t even start with how dumb that is. But in their defense I should add that there is no “correct” way to summarize data. This fact has the troubling implication that no statistics can ever be exact. I guess then that it’s understandable that we don’t have a universal language by which we can describe it.

So there is only one thing to be done and that is to make the data public.

There has recently been a great deal of discussion about how to sequester carbon dioxide out of the atmosphere. I think it is a very promising way of dealing with global warming. The problem is most methods I’ve heard of (notable exception plankton) require a substantial investment of energy, e.g. electricity.

What if I told you that there was a way to sequester a ton, as in a 1 Mg, of CO2 into a ceramic pellet the size of a tiny pebble? What if I told you that the energy cost was exactly zero to accomplish this? Great huh?

Now what if there was a downside. Let’s say that these pebbles were horribly radioactive? “So what,” you may say, “dig a deep hole and throw them in.” What if they were so radioactive that all the equipment used to produce or transport them also got radioactive—not as bad, but still contaminated for the timeframe of a lifetime. Still think its a good idea?

I do.

On a related note, did you guys hear the methanol guy on science friday last week before things got stupid about the dangers of “nano-particles”. He seemed to agree with Jon that paying farmers to make too much corn and then fermenting it and distilling it into ethanol is less efficient than 1) processing fossil fuels into methanol rather than octane and 2)hydrolizing water and mixing it with CO2 to make methanol. Then he goes on to say how easy it is to convert methanol into any organic compound (including ethane).

So we can expect in the future that cheap vodka will be made from oil. People wont like that. Or instead they could make it out of “clean air and pure water”. People will go gaga over that I’m sure. I bet it will taste better too.

MIT stole Caltech’s Cannon!!

Currently girls in swimsuits are posing on it!

Work in building 18 has ground to a halt.

It is absolutely confirmed that in the last 24 hours, I have done NONE of the following things, all of which are entirely foolish:

0. Saved up money, in order to be able to split an Xbox 360 with my girlfriend as soon as humanly possible.

1. Went to camp out at the Cambridge Best Buy, in order to be at the store early enough to buy an Xbox 360.

2. Left said store at 11:30 p.m. to go to a different Best Buy because the manager had sold all his units online despite a promise not to.

3. Camped out at the Arsenal Mall Best Buy, in the freezing rain, and then stood, unprotected, in the rain for 3 hours (4 a.m. to 7 a.m.), after people got serious about queing up.

4. Became beligerent when “camo man” tried to cut in line despite have arrived at only 6:00 a.m. Helped mop of geeky college kids and one VERY pissed off mom to rally and get aforemented individual ejected from the line, whilst simulateously avoiding a brawl.

5. Received the second to last system available.

6. Waited In Best Buy for 2 hours for the computer systems to come online so the sale could be processed.

7. Passed out upon returning home due to lack of sleep and exposure to the elements.

8. Woke up at 5:30 and been thoroughly unsure if that was a.m. or p.m.

We had a brief discussion earlier regarding the effects of exercising in cold weather on one’s pulmonary function. The basic concern is that exercise in very cold weather damages ones respiratory system, leading to impaired oxygen uptake or even “causing” exercse induced asthma (EIA).

Via personal communication, Steve Gaskill, a professor of Health and Human Performance, confirms many of my suspicions that these concerns are largely unfounded. (Disclaimer: Steve actually had this conversation with Audrey, my fiance, who relayed it to me. What follows is my summary of what he said. Any innacuracies are entirely due to me, or Audrey.)

Short answer: Cold weather exercise is not dangerous to your lungs or airways, at essentially any temperature. Cold weather exercise (contrary to popular belief) does not cause exercise induced asthma (EIA).

Long answer: First, damaging actual lung tissue with cold, dry air is essentially a physiological impossibility. If it were cold enough that you couldn’t warm the air enough before it hit your lung tissue, you’d have much bigger problems than damaged lung tissue. And in any case, even if Axel Tiechmann lost a significant percentage of his lung surface area, it would have no effect on his oxygen uptake. Our lung surface area far exceeds our ability to use it to absorb oxygen.

Exercise in cold, dry weather does irritate the tissue lining your bronchial tubes. This causes mast cells in this tissue to break or burst, releasing histamines. The histamines cause your bronchial tubes to constrict, resulting in wheezing and coughing. This is the “post-race hack” we all know. Why doesn’t this happen until after the race? Well, while you’re racing, you’re producing a lot of adrenaline (and other chemicals) that happens to bind with histamines, blocking their effect. Once the race is over, and your adrenaline levels go down, you still have all sorts of histamines floating around and bingo! Wheezing and hacking.

This physiological response I’ve described (breaking mast cells; releasing histamines; constricting bronchial tubes) is clinically indistinguishable from asthma. When we get post race hack, we are literally having a minor asthma attack.

As Tom alluded to in one of his comments, asthma and EIA are exceedingly difficult to diagnose. This is because, as I’ve noted, anyone can experience these symptoms under the right circumstances. What happens is that some people have lots of mast cells. Or their mast cells are really fragile. These people have a predisposition towards more frequent and more severe asthma attacks. This predisposition exists prior to them ever exercising outside in cold dry weather! That last part was important, so I’ll say it again:

Some people, due to natural (or inherited) variations in physiology are prone to experiencing EIA. Exercise in cold, dry air will unmask this tendency.

This means that the unusually high rates of EIA among, say, xc skiers is not unusual. It’s the activity that is unusual. If we took a large sample of people who lived only in the tropics, and forced them to exercise in cold dry air, we would see EIA at the same “elevated” rates we see in xc skiers.

But what about repeatedly irritating all that bronchial tissue and mast cells? Couldn’t that make them more likely to burst and release histamines in the future?

Not likely. Apparently, bronchial tissue is essentially the most regenerative tissue in the human body. Any cells you kill by racing in 5F weather will have grown back and be good as new within like 8-12 hours. (Brayton has pointed out that several people noticed after the Fairbanks races that the post-race hack stuck around longer than usual; specifically that going hard would immediately trigger hacking as much as a week later. I have no explanation for this.)

What does happen is that during the period where you’ve irritated your bronchial tissue, there probably is an increased risk of infection (i.e. upper respiratory colds), but after your tissue has regenerated that risk disappears. He was not entirely sure how big the increased risk of infection actually is.

It seems more plausible that race temperature cutoffs are there to protect race organizers from frostbite/hypothermia and racers from the same.

You see statistics like these all the time. The number of opposite-sex partners for men and women being drastically different. This particular work uses the median but often you hear it as the mean.

Can anyone point out the problem?

Hookups count for both teams, there are the same number of boys as girls. The means should be equal. If these variables behaved, the medians would be equal too.

Let’s pretend that the source of this isn’t self-reporting.

This suggests that the distribution of hookups for girls is substantially more skewed than that of boys. In the case of the mean, there must be a very small number of girls nearly single-handedly holding the true mean away from the estimated mean. Let’s call them sluts, for lack of a better term. In the case of median, you don’t really know except that there are many more hook-ups for the girls above the median than the boys above the median.

On the other team, I know some of you might think of yourselves as “playas”. I’ve got news for you. You are coming nowhere near the sort of numbers the sluts are putting up. You probably have no game at all. You probably just have a knack for finding the sluts.

Look at this one. Over 20% of men have had more than 15 partners while only less than 10% of women have had so many. At some point those curves need to cross and when they do, the sluts will have a lot of ground to make up.

The need to make up all the ground lost by the other >90% of “good girls”. Just to make up the ground lost in the 9th decile they have to have at least 15 additioanl partners each, plus whatever they boys in the 10th decile are putting out, plus all the ground lost by everyone else. It’s got to be over 70 partners. So when it said that only 10% of women have had more than 15 partners, it was a pretty glaring omittion that they averaged 70 partners apiece. Good scientists would have noted that. The only explaination is that those datapoints just didn’t show up.

What was ‘n’? 61,147! If you put them far enough into the tails to be missed by a survey that large, there are just too few of them. They would have to have a hundreds of thousands of partners each. There just isn’t enough time in the day for that, is there?

Here’s an alternative explaination, the missing hookups are by men younger than 44 and women older than 44. I could imagine that I guess with all those cougars on the prowl.

Despite the fact that evolution is just a theory. The FDA has finally started trying to cut down on reckless use of antibiotics.

Synopsis: Meat farmers have been lacing their feed with antibiotics. Most people think it is to prevent local outbreaks due to dirty and cramped conditions. That is certainly true. When you pack them in as tight as in poultry conditions, and you never clean up the shit, there is a serious outbreak risk. But that isn’t in fact the main reason. The main reason is that it stimulates growth and we do not really know the mechanism. Rest assured, it is being looked into.

A step in the right direction to be sure but I don’t think we are going to see a broad prohibition on non-theraputic use any time soon. What do you think this is, Europe? You want us to hose down the pens once a day? Do we have to move the chickens before we spray? [No video! Luckily this guy saved it]

Who is in favor of antibiotic feed? Farming, big lobby.
Who is against it? Scientists, Doctors not enough of a lobby

Where does big pharma come in? They get whichever they want. You might think they want to sell more drugs to farmers. But then again they want to still be able to sell viable drugs to humans.

The funny thing is is that big pharma isn’t very interested in antibiotics right now. Most of the IP is public and for whatever reason, very little new work is being done by the major players. Source

I find that troubling. We seem to be able to keep up with the evolution of the flu, why are these bacteria so much harder? I’m not entirely clear on the biology, could someone fill me in here? Is it because they have an easier time outside of a host? In that case, I’m tempted to think there would be less pressure to adapt to the drugs (so long as we don’t pour them in the ground water).

Here’s the thing that I only just realized. You know how that antibacterial soap of yours kills 99.9% of the bacteria on whatever it is? That is what it kills the first time. The surviving .1% get to reproduce. I think that the primary goal of soap is just to move the bacteria to a place where it will be harder to get into my body. For dishes/skin, I think regular use of antibacterial soap might be counter-productive. In my digestive tract it is well known that an antibacterial agent can be very harmful in fact.

On a surfaces though, particularly rough or porous ones, I’m less sure. I think a genocide policy probably works best there. We just need to be serious about the agent. Ammonia? Bleach? Alcohol? Acetone? Phenol? Think of the frat basement super-bug that you just carried home on your shoe.

Jon has been excited about Freakonomics and has mentioned it to me a number of times. I didn’t look into it because I am lukewarm on economists in general. But today I happened across a snipet of an interesting article about car safety seats that turns out to be by the same guy.

Livett seems like the sort of guy I’m into… You know, a “Show me the data!” sort of a guy. “54% with respect to what?”

But I don’t think it is likely that a NYT bestseller would have the meat and potatoes that I would want on these subjects. I’m guessing that they probably don’t throw around named functions and distributions that readers would have to go look up on Mathworld.

I read the intro on Amazon and while you would try to make the intro pretty easy even if the book were going to be the brutal slog I am hoping for, it didn’t assuage my concern about it being watery. I’m going to find one of those buildings that has the information on dead trees and skim it for figures.

But then I dug up his homepage which has been very effective at helping me get nothing done today. A lot of neat stuff in there.

I think I want to be a freakonomist in my spare time. But first I need to win some major credibility other field so people don’t realize that I’m a crackpot. You know, like Feynman.

Remember the stats fairy? He seems to have visited Slate.

He does a good job explaining why Rector and Johnson are hacks, but never quite comes out all the way to admitting that it wasn’t a terribly good paper. While the other paper that they linked is only a bit better, it looks like he has had grad students looking at virginity pledges for long enough to be more credible than I initially assessed.

Suppose we want to answer the following question. “Do abstinence pledges reduce teen pregnancy or STDs?” Let’s take a survey and find out if the answer is yes or no.

Hey look! kids who take an abstinence pledge are more likely not to have sex for 12-18 months later than their non-pledging peers. Well at least that’s what they self-report. But c’mon, why would you lie about breaking a pledge. Let’s trust them.

Wait a second, what if the only kids who take the pledge planned on not having sex anyway? Really all our survey data show is correlation. So lets look at some other factors. If pledgers abstained more across the board, then we might be more inclined to think that pledging makes a difference. That way we remove the shared cause issue. There would still remain the “Does abstainance cause pledging?” question, but that could only be resolved by controlled experiment.

So what else correlates with pledging? In order 1) Asianness (shrug) 2(tie)) religiousity 2(tie)) not having a paramour (more variance than religion) 3) being unpopular 4) parents dissaproving of sex

Funny, the things that correlate with no sex are pretty similar 1) no lover, 2)asian 3)religious 4)parental 5)unpopular. I am not going to pick my way through the fog of multivariate regression. The paper couldn’t do a good job of it, neither will I. It is imaginable that the pledge has some remaining effect. But 18 months? We all know that is too much to possibly be true.

Heritage has two studies simultaneously attacking and defending the work. Here is the NYT? piece on the issue.

The team needs to do “a lot of work” on its paper, said David Landry, a senior research associate at the Alan Guttmacher Institute in New York. He said in an interview that it was “a glaring error” to use the result of a statistical test at a 0.10 level of significance when journals generally use a lower and more rigorous level of 0.05.

Dramatic Interlude
NYT reader: OMFG! They didn’t use the .05 level?! That’s not scholarly… Actually what does that mean?
poof
Stats Fairy: That means that there is more than a 5% chance that he erroniously rejected the hypothesis that there was no effect. In other words there was more than 5% but less than 10% odds that the variation was by mere chance.
NYTR: To say that, wouldn’t you need to know the odds ahead of time?
SF: Well, kinda. You need to know the odds of getting certain results. Luckily someone long, long ago did some math and made tables. Sometimes things turn out just like in the tables. For instance coin flips…
NYTR: I can handle coin flips. Is this like coin flips?
SF: Well, no, but we statisticians can handle coin flips too. You see if I flip a coin a very large number of times…
NYTR: You just said it istn’t like coin flips.
SF: Do you like garlic knots? Let’s to go to Ramuntos.
poof
NYTR: Hey, come back! I have another question.
poof
NYTR: What is so special about .05 and .1? Why can’t they just tell me the odds straight away.
SF: No, no. Once you’ve rejected a hypothesis, it is completely rejected; it is utterly gone forever. It is not rejected until you get to .05 (or .1 if you live in a backwards country like Uzbekistan) and then blam. You see .05 is a very special number. Look at your hand. How many fingers are there?
NYTR: Where are you going with this?
SF: Right, there are 5. That is why we use 5% significance testing.
NYTR: You’ve got to be joking.
SF: God wouldn’t have given you five fingers if it weren’t the right level of significance.

Seems like we need to do a quick credibility check. So what do we have for papers by Bearman? What do we have for Rector and Johnson? Landry seems potentially credible except for the fact that he deliberately mislead the reporter.

No surprises there considering the articles. We’ve got acedemic hacks versus political hacks. Rather than defend either, I think I am going to cut this post off short.

There is some peculiar usage in English with respect to (WRT) health issues. When your body is infected by a parisite you might say “I have crabs.” That usage seems completely natural. Then the verb, “to have,” is extended to conditions also. “I have laryngitis” or “I have athlete’s foot.” or even “I have a pain in my knee” This seems like a natural extention of the previous case. Considering the fact that “I have a problem” is standard usage and “A problem presents itself to me.” is an akward conjugation, I think we should stick with it. I can’t think of anything better.

But this is rapidly starts breaking down when we start describing less binary conditions. While you could have a case of variable severity, you either have crabs or you don’t. But consider, “I have high blood pressure.” Ordinarily, that would mean that my blood pressure is abnormally high right? But since we use “have” with conditions, that sentence carries with it a lot of baggage. That baggage is potentially very snaggy in some situations. It makes “high blood pressure” into a medical condition. I very much preffer “My blood pressure is high.” But come on, higher than what? How about “My blood pressure is typically %110 of normal.”

Still don’t see the problem? How about syndromes? We use the same language to describe the having the flu as having a syndrome. It gives a syndrome—just collection of symptoms (“symptom” is even to strong of a word)—a realness as something that you can physically have. I don’t think the problem here is with syndromes themselves or the science surrounding them. I think it is great that people are documenting the sort of states that are correlated under certain circumstances. I think the real problem here is simply language usage. How’s about this for a start “I exhibit uncredible witness syndrome.” Or “My state is largely consistent with uncredible witness syndrome.”

Now we are getting close to the truth of the matter. We know very little about mental health or about any relationship that goes through the brain. We’ve got a vague feel for what goes on where and a few regulatory chemicals and a tiny bit about how neurons work but that’s really all. So we document all we can and give things helpful names. The crucial mistake is pretending that we understand it because we can refer to it by name (personally, I blame Aristotle). For instance, when someone has some sort of mental problem, i.e. they exhibit unwanted mental abnormalities, what happens is they try a lot of different drugs. If a drug has a positive affect on the abnormality, they now know what the problem is. It is the problem that that drug treats! I know someone who is often abnormally paranoid and depressed. She has been called many things (among them schitsophrenic) solely because she responded to certain medications.

But this isn’t very diagnostic. Consider the fact that everyone responds pretty well to cocaine for instance. The doctors have realized this since so many diverse things are being treated with practically unabashed amphetimines. Amphetimines are perhaps going to be a recurring theme here.

Let me get back to the thread from which I’ve strayed. Let’s not pretend we understand what is going on. To that end, let’s use language that is more acurate. Rather than saying “The boy has ADHD” perhaps we should say that “The boy responds well to amphetimines.” But doesn’t pretty much everyone respond well to amphetimines? Well, yes, they pretty much do.

¹ As an aside, the only worse verb that is worse than “to have” is “to be.” It is too broad a usage. “Tom is crazy.” “That girl is overweight.” “That whore is HIV positive” (I always say “has HIV” seeing how it is a virus that you carry around). I remember in Spanish class learning about whether to use ser or estar with gordo(fat). I thought it was kind of neat that Spanish with less than half of the verbs of English would have two words for “to be” one describing atributes like “I am American”(ser) and another to describe conditions “I am hungry” (estar). It turns out that you can use either for fatness. Using Google hits to see which is more popular tells that it depends on conjugation and gender. Of course, context matters too. Teammates seeing Ulrich in the spring surely would say “Jan Esta gordo!” while if asked about Cosmo’s girth, I would say “Cosmo es gordo!” For loco (crazy) with regard to people, again you can use either but estar dominates usage.

So, its looking increasingly likely that we’re going to be drilling for oil in ANWR, especially after yesterday’s senate vote. I was listening to On Point on NPR last night , and the matter was being debated by various parties. You can listen here

The guy who works as an oil consultant was extolling the virtues of openning ANWR to drilling, claiming that it would reduce the amount of oil we’d have to import by 3%, which would increase our energy security.

I’ll go out on a limb in saying that energy security and decreased reliance on the Middle East are both good things, but 3%! Is that really worth the effort and the risk to the environment. I’m thinking probably not.

I’ve got a way better idea, how about we spend all that mental, physical, and monetary energy on coming up with better hybrid cars and better fuel cells. I have this crazy notion that we should make floating hydrogen farms that would be, essentially, big solar panel arrays. You could use the sun to generate electricity, and then use the electricity to electrolyze sea water into oxygen and hydrogen. Of course, until we can come up with a good way to safely and efficiently store hydrogen in a non-gaseous form (easily reversible addition/elimination to a polymer or other material), there’s no way we could pull it off. Whoever solves that problem is going to Stockholm, hands down.

I consider myself to be an experimentalist which was defined to me as someone who thinks that all knowledge is available through experiment only. This implies that any question for which an experiment couldn’t be designed, is not meaningful. The best example came from my freshman physics professor as a response to a meaningless quantum question: “How many angels can dance on the head of a pin?”

Is social science lost since it is impossible (or at least impossibly difficult) to perform meaningful experiments on many of its subjects?

Let’s say that I have a hypothesis tester into which I can feed in an idea and it will spit back out a number telling how likely the hypothesis. What if I had another machine that could come up with every possible hypothesis? I could plug these two machines into one another and after a while (let’s not worry about how long just yet), I would have a big probability distribution over hypotheses. I would consider this distribution to be all knowledge that exists. By doing all kinds of integrals on this distribution, I could answer any answerable question.

Of course there are an (uncountably) infinite number of hypotheses even for a tiny question like, “How long is Cosmo’s wang?” But we can always partition the hypothesis space into a finite number of regions to fix this. Then we just to build a machine that will enumerate these hypotheses. There would be a granularity problem that could potentially overlook good hypotheses. Under certain conditions one could iteratively hone the partitions to be finer in regions where we are more interested. Under other conditions, we might simply need to sample for a long, long time.

Now consider the hypothesis tester. Let’s say that we programmed it with everything that has ever happened. There will be things that haven’t happened yet that would shed light on on our hypotheses. So to be perfectly accurate, we would have to wait until the end of time and then program the machine.

To be more practical let’s approximate it with all information up until the time that I run the machine. In a lot of cases, I would expect that to be sufficient to get a decently accurate answer. For instance, I would expect certain social questions like “Income vs Education” questions to be in very good shape since natural variations have provided plenty of data. On the other hand questions about bonding of hypothetical molecules could be less accurate than we might need. In these cases, we could use the results to guide experiments that maximize the new information.

Of course, in practice we only really need to give it the “relevant information” once we’ve determined what that all is. That will doubtlessly speed things up.

You may think that I’m being silly but I took a class where we made limited versions of such machines.

Back to Social Science

At the heart of most problems is a big pile of correlated data. How do we deal with that? First we need to define the hypothesis space. Take all of the variables from the data and our hypothesis space consists of all possible relationships between these variables. These relationships are limited to

1. A directly affects B,
2. B directly affects A or
3. there is no direct affect.

We can also add to the hypothesis space any number of mystery variables for which we have no data but to which we can freely assign relations. For instance, one could be the hidden cause of several inter-correlated variables.

Now all that is left to do is nail down mathematically rigorous definitions for “affects” and how exacly we compute likelihood and we will be able to get science-style answers. Furthermore, neither of these are terribly difficult either. For instance “affects” could mean “changes the probability of” and “likelihood” could be how well we can “predict” 1/e’th of the data if the model is based on the other (1-1/e)’th.

Do we have any social scientists in the readership? You want the science-style answers? There are computer programs that can do this. You can just go get one.

¹ I secretly know that such machines would still have some problems breaking symmetry. For instance given data on weather and wet sidewalks only (i.e. we didn’t have any sprinkler, or dog-piss data), the two hypothesis,

1. when the sidewalk is about to get wet, it affects the weather and
2. rain affects sidewalk wetness

will be equally likely. Then the researcher will need go outside with a bucket of water to determine which is correct. In general, these symmetries will either have obvious answers or they will point to the most telling experiments. Should those experiments prove impossible, I am prepared to live with multiple hypotheses. In fact, I have no choice.

Just a heads up everyone. The point of this blog is for me to be a become a better writer, not to be a journalistic entity, so I’m going to rewrite posts without notice or justification.

I’ve got a major crush on economics. There are some haltingly brilliant thinkers in the field of economics. One quirk of serious economists is that they refuse to make “normative” judgments. I take that to mean the refuse to say whether something is right or wrong, good or bad. They can only comment on efficient, or inefficient. Let me give you an example.

Last week Becker-Posner (Becker and Posner, trust me, are both really smart) addressed the issue of Medicare on their blog. They had some really interesting ideas. Why are deductibles in medicine so low? How elastic is healthcare? Can we make patients be less wasteful just by making them pay a bit more? We should try to encourage people to give up more expensive treatments (surgery), in favor of less expensive treatments (ie pills, or death).

I also thought that it was interesting that they both supported doing away with MediCare (old people coverage.) Why should rich old people get free medical insurance? If they don’t plan their medical coverage well enough, then let them pay for it until they’re poor enough to be covered by MedicAid (which covers poor people.)

Really interesting, guys. They’ve clearly thought about this stuff! This is where I think things get a little out of hand. Well, Posner flies off the handle, Becker seems to keep it in check (I suspect this is only because he didn’t reply to comments.)

Posner slips in the first one kind of quietly:

A person who has no assets lacks a compelling reason to buy medical insurance; he will be able to obtain medical treatment free of charge, as a charity patient.

Ok, maybe you’re right, Posner. Maybe poor people are free-riding the system because they can, not because they have to, but I’m so not certain.

Then:

If people value additional years of elderly life at more than the cost of the extension, the cost may be worthwhile, though it doesn’t follow that it should be subsidized.

Ok, now he’s saying that if you’re dying, and you’re too poor to pay for the treatment to keep you alive, then tough. There’s no reason for society to pick up your tab. That’s getting a little creepy.

I didn’t read the comments, but some people must have called him out on some of these things.

(Side note, every week they reply to the comments, and every reply starts out “Some of the comments brought up some really good points. I don’t have the time to address them all, but here’s a few thoughts…” Someone should tell them that we get the picture, and we appreciate that they write at all.)

Anyways, they called him out, “Posner, you can’t compare health insurance to car insurance!” and he gets a little defensive.

The analogy of medical to auto insurance was criticized in some of the comments [for no good reason] ... you cannot deny medical treatment to someone who refuses to buy health insurance … but you can punish him for not buying it

Woa. He’s serious about mandatory health insurance. If you can afford it, you had better buy it. Otherwise you’re no better than an uninsured driver (asshole free-riders.) I argued that it wasn’t a fair comparison, because you can avoid paying for car insurance by not owning a car, but no one seemed to notice or care.

What about people who can’t afford medical insurance?

It is true that some, and probably many, people will not be able to afford health insurance, and I agree that, as a practical matter, they cannot be denied treatment just because they can’t pay for it.

A practical matter? Jesus. Never mind that the thought of denying poor people healthcare simply because their poor is morally abhorritious. You know what, I don’t want these guys in charge of health care anymore.

Summers “girls are all stupid” speech transcript here.
I know it’s long but nobody is allowed to say anything until you’re done.