What works in healthcare?

Short answer: almost nothing.

I’m being flip about a very serious subject, but at the same time I am in fact serious. Modern medicine probably works a lot less well than you probably think it does.

How could I say such a thing when you look at graphs like this?

 Maternal mortality over time Our World in Data Global Health https://ourworldindata.org/health-meta

Maternal mortality over time
Our World in Data Global Health https://ourworldindata.org/health-meta

I say it because that is precisely what the evidence shows. If you look at Cochrane, the world’s preeminent aggregator of medical statistics, it is hard not to come away a bit disappointed. Effect sizes [usually the difference between the test group and the control group scaled to standard deviation units] tend to be rather modest.

Here are a few examples:

You can amuse yourself by finding your own examples. There are a few things that genuinely work well. But even for things like MMR, the evidence isn’t as good as you might think.

I suspect what is going on is that medicine works, just barely, on average. You get things like the long slow decline of maternal mortality from the confluence of lots and lots of little things added together. If you look at anything else, heart disease, or cancer, you will see the same pattern. Vaccines are an exception. Disease rates for things with effective vaccines just drop off immediately.

 Polio dropped right off

Polio dropped right off

 Heart disease is on a long, slow decline

Heart disease is on a long, slow decline

Which brings us to my motivation for bringing this up at all. Random C. Analysis just published an updated argument that healthcare spending in the United States isn’t badly out of line with the rest of the world, once you take into account how much richer we are than just about everyone else. We have more, so we spend more. According to RCA’s data, when income goes up 1x, healthcare spending goes up 1.6x.

My contribution to this is to suggest the reason for this is we purchase more and more healthcare is precisely because it doesn’t work very well. At lot of modern medicine treats symptoms better than causes, because we don’t understand the causes very well. You buy as much symptom relief as you can afford. Even when treatments are genuinely curative, the success rates are often low. For example, the controversial statistic number needed to treat attempts to quantify how many patients need to be treated in order to produce one cure.

This number is often quite large, often on the order of 50 to 100. Even for a really good treatment with an NNT of 10, 9 times out of 10 it isn’t better than the alternative it is being compared to. That is a lot of wasted time, effort, and money.

We just don’t know how to predict the 1 time it works, so we treat everybody and hope for the best.

You can find speculation like this from Goldman Sachs that you can’t make enough money curing disease as compared to offering palliative care. I’m sure there really are businessmen who would gladly milk you for everything you are worth, but I don’t worry about in reality because no one understands how to cure most of the things that ail us. It isn’t like miracle cures are being withheld, or even that research is being directed away from cures. We don’t know enough to do that.

If we really wanted to limit costs for healthcare, it might be possible if we ruthlessly limited access to only things that really worked. We would get 80% of the benefit for 20% of the cost. But people would be pissed. The other 20% of the benefit does actually work, sometimes. I don’t think this is possible, or even really a good idea. Any real solution will involve technology we don’t currently possess.

LinkFest 2018-07-16

 On the left, what everyone thinks machine learning is. On the right, what is actually is.

On the left, what everyone thinks machine learning is.
On the right, what is actually is.

Ways to think about machine learning

I've been a skeptic about artificial intelligence in general, and a critic of the ways the actual technology has been hyped. This is a pretty reasonable take from someone who is willing to invest a lot of money in machine learning. Machine learning is another kind of automation. We've been seeing big things come out of automation for 100 years, it makes modern life possible, but it is easy to lose perspective.

 

Why the Future of Machine Learning is Tiny

An example of what machine learning can mean in practice.

 

Snapping Spaghetti

Applied mechanics of fracture with slo-mo video! Why does a piece of spaghetti break into three or more pieces when bent? Now you can find out!

 
 Manufacturing output per capita, colored by what percent of the economy manufacturing is

Manufacturing output per capita, colored by what percent of the economy manufacturing is

 Manufacturing output divided by employment in manufacturing, Canada and Taiwan were missing the employment estimate

Manufacturing output divided by employment in manufacturing, Canada and Taiwan were missing the employment estimate

Global manufacturing scorecard: How the US compares to 18 other nations

Manufacturing stats are a subject of interest to me. I don't find much of interest in the Brookings manufacturing scorecard, which is just their subjective rating of various things. Rather, I plotted the manufacturing output for each country per capita, and per person employed in manufacturing, a kind of crude productivity number.

I think the *really* interesting thing here is how much Switzerland sticks out. The parts of the economy in Switzerland I am most familiar with are chemical precursors for pharmaceuticals and medical devices, which are both high value sectors.

 

When Evidence Says No, But Doctors Say Yes

This is a great article on how hard it is to find clear evidence that common therapies work, and how hard it is to disseminate that knowledge once we have it.

 

Israeli space probe to land on Moon in 2019

I was going to say this isn't surprising from a country that also made their own nuclear weapons, and then I saw the money for it came from a South African businessman. Israel and South Africa *probably* cooperated on nuclear weapons too.

 
 

Thou Shalt Not Wirehead: Religion vs Gratification

This is pretty good. I think I mostly agree, except I am also very interested in whether religion is *true*. Religion can be pretty helpful in encouraging behaviors that help you in this world, for example, the prosperity Gospel is pretty popular because it actually works out that way. If you give up drinking, gambling, and whoring, usually your life materially improves. But sometimes religion can make you do things that are the opposite of helpful in this world. For example, the Xhosa.

 

Welcome to the Party, Pal

A reflection on how the political coalitions in the United States came to be.

 

Does Free Trade Bring Lower Prices?

Dani Rodrik reminds us that we have to describe the world as it is when we make economic projections, not a model of it.

 

Donald Trump tells us truths we don’t want to hear

Matthew Paris argues that Donald Trump acts like an Emperor, and you shouldn't be surprised by that.

 

The Fear of White Power

What is the value of political correctness to a minority in society? And is its cost?

 

Shortwave Trading | Part III | Fourth Chicago Site, East Coast, Patent, Regulation, and Farmer Kevin Mystery

High volume traders are rolling their own radio networks to get a leg up on the competition.

 

Traditional Euro-bloc: what it is, how it was built, why it can't be built anymore

The perfect counter-point to my post on modern urban development. We can't just build things because we like how they look, we have to care about money, and how neighborhoods evolve, and what will actually work for the people who live there.

Linkfest 2018-06-11

I meant to get this one out last Friday. Ah well.


How we can learn from the history of protectionism

It is easy to find lots of economists who are down on protectionism, but the evidence turns out to be rather mixed on exactly what its effects are. There are countries that have done poorly with this policy, and countries that have done very well indeed.

It’s Time to Think for Yourself on Free Trade

Dani Rodrik is an interesting and thoughtful economist. This example from his article is illuminating:

In some sense we all know this. Consider another thought experiment: Suppose Harry and John own two companies that compete with each other. How do you feel about each of the following four cases?
  1. Harry works really hard, saves and invests a lot, comes up with new techniques, and outcompetes John, resulting in John and his employees losing their jobs.
  2. Harry gets a competitive edge over John by finding a cheaper supplier in Germany.
  3. Harry drives John out of business by outsourcing to a supplier in Bangladesh, which employs workers in 12-hour shifts and under extremely hazardous conditions.
  4. Harry “imports” Bangladeshi workers under temporary contracts and puts them to work under conditions that violate domestic labor, environmental, and safety laws.
From a purely economic standpoint, these scenarios are what economists call “isomorphic” — they are formally indistinguishable because each creates losers as well as winners in the process of expanding the economic pie in the national economy. (That is, Harry’s gains are larger than John’s losses.)

For economists to call these four situations in some sense identical is probably important for analysis, but it probably also warps the mind to do that too regularly.

JASP

 JASP is an open-source project supported by the University of Amsterdam.

JASP is an open-source project supported by the University of Amsterdam.

I haven't used JASP myself, but I saw people talking about it on Twitter. I will give it a try, and perhaps report back. I am entirely in favor of easy to use stats tools.

Burying Your Father and “Return of the Jedi” (1983)

This was a fascinating reflection on fatherhood, spurred by the climax of Return of the Jedi.

Is Global Equality the Enemy of National Equality?

I like Dani Rodrik's work, but sometimes I also think he's nuts. This is a good example of why. I think really bad things would happen if we tried to implement this suggestion of globally free labor movement.

Tolkien 101: The Animated Tolkien Movies

A roundup of the animated Tolkien adaptions over the years. The author has a whole series on this subject. Ooh, and one on Conan!

Even Dead, The Expanded Universe Is Better Than Disney Star Wars. And That's A Good Thing

I have said my piece on Disney's decision to reboot the Star Wars universe, but in the time since, I have found the new novels pretty lackluster. There was some crap in the old EU, but the crap to good stuff ratio seems poor in the new canon. Thankfully, the animated series are making up for the deficit.

The Lifespan of a Lie

A retrospective of the famous Stanford Prison Experiment, that is a case study of the failures of social science that led to the replication crisis. The first person, and the last person, Philip Zimbardo lied to was himself.

Normalizing Trade Relations With China Was a Mistake

A perennial theme here at the blog: are we sure we really knew what we were doing?

Linkfest 2017-09-09

Posting has been light of late, my home PC needs a new power supply. The replacement should be here by Tuesday.

When Correlation Is Not Causation, But Something Much More Screwy

UCLA sociologist Gabriel Rossman explains how easy it is to fool yourself with the way you collect your data.

Toyota’s Research Institute head says full autonomous driving is “not even close”

I'm a bit of a skeptic about how easy it really is to completely automate driving.

The Tater Tot Is American Ingenuity at Its Finest

The Tater Tot was made out of french fry waste products.

Moving the Finish Line: The Goal Gradient Hypothesis

This is a fancy term for the idea that the closeness of a goal can influence our motivation. This is the idea Uber uses to get drivers to work longer, and how video games are made more addictive to play. Something that doesn't get discussed here is risk. For example, a big difference between the cited example of getting $12,000 at the end of the year as a bonus, or $1,000 at the end of the month, is that bonuses are dependent on financial performance. In the real world, you might get more money from the monthly option, which chops up the risk of the company not making enough money into smaller bits.

A Simple Design Flaw Makes It Astoundingly Easy To Hack Siri And Alexa

I imagine it was easier not to take frequency into account when designing these apps. This seems easy to fix, in principle.

Voynich manuscript: the solution

This turned to a be a thick problem. You needed a lot of the right knowledge in the right head to solve it.

 My shelf [and a half] of Jerry Pournelle books

My shelf [and a half] of Jerry Pournelle books

Passings.....

Jerry Pournelle, one of my all-time favorite authors, died yesterday. I followed Jerry's website and writings for 16 or 17 years. Jerry was an early adopter of the Patreon method of earning a living, as he was an early adopter of so many things. I supported him for the last eight years or so. Jerry outlasted a stroke and brain cancer, and while those slowed him down a lot, he was actively writing and blogging until the end. 

Jerry led a long and interesting life. I would have loved to read his memoirs, which he never got around to writing. Hopefully someone else will fill the gap.

 

Is the Economy Legible

I am an amateur in most of the fields that I enjoy commenting on in this blog. In my own field, I get as annoyed as anyone else when an ill-informed outsider buts in. Nevertheless, like Chesterton, I think amateurs ought to be allowed their say, especially since so many experts are demonstrably incompetent.

 Full-time workers wage growth

Full-time workers wage growth

In this case, it was the juxaposition of two articles that came across my Twitter feed recently. Arnold Kling's Is the Economy Illegible?, and recent Federal Reserve wage growth data.

First, the wage growth data. This could be seen as a cautionary tale about over-reliance on averages. The average data doesn't look so hot, but that is masking interesting patterns in who is retiring, and who is coming back into the labor force. Even these latter numbers are still averages, but better, more informative ones. It is always worth asking yourself whether an average is the right tool to answer your question, when the metric of interest applies to an individual. 

The next one is Kling's article on the broad economic statistics we all use to judge the health and success of our economy. Kling uses the term of art "legible", which seems popular today, but I would probably ask the same question in a different way: does GDP measure something real? Does it measure the same things through time? Do you have any way to verify this?

These are standard questions of data quality that apply to any effort to track and trend data over time. I'm most familiar with this in the context of quality control data, but I think the principles still apply for something as grand as macroeconomics. I especially like Kling's example that the computing power in an iPhone 7 would have been worth $12M USD in 1991. Does this really mean technological progress has made us all that much richer since 1991? I'm dubious this is true, which means that the stats are off in a major way.

Linkfest 2017-07-14

The Last Barfighter: An Arcade Game That Pours You A Beer If You Win.

Genius.

Utilities fighting against rooftop solar are only hastening their own doom

This is an interesting article. I have two primary questions: do all utilities make money the same way? This article claims they only make a profit by building infrastructure. I can imagine other models are possible, and it is a big country. Second, here in the southwest, you can't store energy when the sun is shining and then sell it later when the price is high, because those is the same time here.

Battery storage: The next disruptive technology in the power sector

A citation from the above article.

Most Scientific Research Data From the 1990s Is Lost Forever

I've done a bit of digital archaeology myself, and I know that keeping data current is an on-going job. Access to current data is easier digitally, but archiving is worse. The most durable storage medium we have is fired clay.

Ancient Humans Liked Getting Tipsy, Too

As well they should.

The Strange Afterlife of Pontius Pilate

I don't find the idea that Pilate was not culpable for the death of Christ terribly obscure, but this is nonetheless an interesting article.

Apple’s Third Co-Founder Has Never Used an iPhone and Has No Regrets

Wayne sounds as weird as the other guys, just in his own unique way.

DARE to Look at the Evidence!

This isn't the first time I have heard the DARE program doesn't produce any good results.

Radical Book Club: the Decentralized Left

David Hines continues his look at the radical Left through the books they write about themselves. The ways in which protests are made to look spontaneous is particularly interesting to me.

Go Ahead, Put Salt on Your Food

The health risks of salt have been greatly exaggerated, mostly because existing summaries fail to make distinctions between people with much greater risks and the general population.

PIN analysis

An oldie but a goodie.

Linkfest 2017-06-16

Is Islam the rock on which the liberal order broke?

Razib Khan's Brown Pundits post on whether liberalism broadly construed can stand against the intransigence of Islam.

Think wine connoisseurship is nonsense? Blind-tasting data suggest otherwise

It is easy to make fun of the expensive culture of wine tasting, but I'm always willing to follow the data.

John Betjeman on Greece (but really on England)

A beautiful poem about the Church.

We Could Have Had Cellphones Four Decades Earlier

I like to poke fun at libertarianism, but this seems like an example where a libertarian critique is warranted. A variety of vested interests created by law defended their prerogatives to the detriment of the public interest. In this case, rather than anti-competitive practices by AT&T, we seem to have a case of neglect.

The Best Movies of the 21st Century

Ross Douthat ponders what movies deserve attention since 2001. Also see Steve Sailer's response.

How Pasteur’s Artistic Insight Changed Chemistry

Arguing that Louis Pasteur's abortive art career helped him discover chirality.

Status in the Iliad

Gabriel Rossman does a network analysis of battles in the Iliad.

That Time an Algorithm Whisperer Took Me to the Heart of Darkness

Joel Stein is a funny, funny guy.

 

Linkfest 2017-06-09

I love the movie Cars, and I am very excited about Cars 3, which looks less like a cashing-in on the initial success of the first movie sequel [you gotta pay for studios somehow] and more like a real Pixar-worthy sequel with better animation technology.

This post is nearly ten years old now, and I think still pertinent. 

I didn't learn anything new here, but I've got a better memory and a deeper interest in history than the average American. The thing that gets me is no historical figure can stand this kind of scrutiny. For example, here is Lincoln debating with Stephen Douglas. This is the argument of the Progressive Left, that no one before the present is redeemable in any way, but it surprises me when less radical people advocate for ideas that destroy their own position.

 Speech on the Kansas-Nebraska act at Peoria, Illinois

Speech on the Kansas-Nebraska act at Peoria, Illinois

It has been a while since I posted something about statistical software and graphs, so here you go!

An article attempting to link the Younger Dryas with Göbekli Tepe. I lack subject matter knowledge, but this is an interesting idea. Greg Cochran thinks the Younger Dryas impact hypothesis is bunk.

David Warren argues for the return of unsafe spaces at universities [with tobacco!]. 

I think I started reading David Warren about the same time I started reading John. J. Reilly, shortly after 9/11. He wrote a column for the Ottowa Citizen at the time, and his beat was terrorism.

Warren was better suited than most. He had been to Afghanistan in his salad days. He ended up a big supporter of George W. Bush and of the War in Iraq as a clash of civilizations. Warren spent quite a bit of time in the wilderness repenting of his sins following 9/11, and I still occasionally read his essays.

Answer: somewhat.

I've never felt like the Hobbit was a kid's book, but I have friends who feel otherwise.

College Learning Assessment Plus

The Wall Street Journal has an article up on the CLA+ test, with an accompanying data set from 68 public colleges obtained through FOIA requests. I put all the data in an EXCEL style spreadsheet as well.  

I always like to plot my data, so here is a scatterplot matrix of the whole thing:

 CLA+ Scatterplot Matrix

CLA+ Scatterplot Matrix

There are several columns that are just different ways of saying the same thing, like Freshman Score and Freshmen with Below Basic Skills. So I dropped everything that was just the same data in a different format and we get this:

 Subset of CLA+ Scatterplot Matrix

Subset of CLA+ Scatterplot Matrix

Just about the only scatterplot that stands out to me is that higher freshman scores are pretty correlated with higher senior scores. I never would have guessed.

The next most interesting is the relationship between freshman score and the difference between freshmen and senior scores. The correlation is negative, perhaps implying there is a score ceiling in the test, or that average college graduates tend to end up in about the same place by the end of school.

Both freshman scores and senior scores are correlated with graduation rates, but since we are supposed to be using this data to see whether a given college does anything useful, I plotted both freshman and senior scores against graduation rates, but I color-coded the points by the improvement between freshmen and senior scores.

 CLA+ relationship between senior score and graduation rates, color coded by score improvement

CLA+ relationship between senior score and graduation rates, color coded by score improvement

 CLA+ relationship between freshman score and graduation rates, color coded by score improvement

CLA+ relationship between freshman score and graduation rates, color coded by score improvement

The color-coding looks random on the senior scores graph, but against freshman scores the highest improvements are concentrated at the lower-left. This might be interesting, since point difference versus graduation rates in general looks pretty random in the first scatterplot matrix.

I don't see anything groundbreaking here, which is probably why colleges don't talk about this much. If there was something to crow about, they would.

Linkfest 2017-06-03

Speaking from experience: a note on mysticism and religion

Niall Gooch's attempt to expand on a distinction between a mystical experience and a religious one.

Three Polish Poems

BD Sixsmith agrees with Niall in poetic form.

The History of U.S. Government Spending, Revenue, and Debt (1790-2015)

Morphing graphs on US government spending over time. 

Stratolaunch rolls out giant aircraft

Paul Allen, co-founder of Microsoft, continues testing on his launch vehicle.

How to build a fallout shelter

The Swiss have always taken preparedness to a new level.

This data set took six years to create. Worth every moment

Every US department now rolls up budgeting data into one source.

New evidence that lead exposure increases crime

A series of analyses to test Kevin Drum's theory that lead exposure explains much crime in the US.

Why we should love null results

Knowing what doesn't work is important too.

Linkfest 2017-04-07

A really fun map, and as the article reluctantly notes, not especially credulous about the tall tales travelers shared.

I've linked stories on Uber's evil genius before. This seems to be their attempt to make Hari Seldon real. The New York Times article includes spiffy simulations.

An intriguing look at what kept the Romans, so very advanced in some ways, from an industrial revolution.

Father Matthew Schneider defends oil pipelines from a Catholic point of view.

I'm not surprised, but I've also known some truly gifted scam artists.

A fun interactive map that lets you see where people live.

A labor union associated think tank argues that free trade lowered wages in Mexico.

Noah Smith defends Case and Deaton.

James Miller interviews Greg Cochran, part 1. A wide-ranging conversation, covering Less Wrong, microbiology, federal funding, and free speech. Well worth a listen. [I helped fund this.]

Greg Cochran points out the flaws and misunderstandings in Cordelia Fine's new book. [I also helped fund this.]

Hollywood accounting is pretty shady. I wish Shearer the best of luck.

Damon Linker points out that about as many people die from alcohol overdose as heroin overdose in the US, and the overall rate of alcohol problems mirrors the rate of things like depression: about 1 in 3 over a lifetime.

Kevin Drum has a chart of the rates of death by drug overdose from selected periods and causes that dovetails nicely with Linker.

Ireland's population still hasn't recovered from the combination of famine and mass emigration in the mid-nineteenth century.

  • Crime rate in the US

This chart comes from US Census Bureau data, and the paper containing it seems to be this: J. A. Miron, 1999. "Violence and the U.S. prohibitions of drugs and alcohol," American Law and Economics Association, vol 1(1), pages 78-114. I am intensely curious about the very low homicide rate at the beginning of the twentieth century. Particularly since it differs from other charts of the same thing, for example this one from Steven Pinker's Better Angels.

The solution seems to be that a 1995 article in Demography contains a model of homicide data that has been widely used to estimate missing data. Why do we need a model you ask? Because the data from the period didn't include everywhere in the United States. Not being an expert, the model seems reasonable, but don't forget it is a model.

Eckberg, D. L.  1995. “Estimates of Early 20th-Century U.S. Homicide Rates,” Demography.

I've stumbled on the same dataset twice now. The Tuskegee Institute started keeping track of lynchings in 1892. The data only goes back to 1882, which is the year the Chicago Tribune started keeping numbers. The NAACP also started collecting numbers in 1912. You can see in the chart the point when lynching stopped being just a kind of frontier justice, and started being a way to terrorize black Americans. If data existed for the entire 19th century, I think this trend would be even more clear. Data in EXCEL here.

A post on the reddit for Slatestarcodex questioning the qualifications of Emil Kirkegaard, whom I'm linked several times. I think most of the points made by the anonymous poster are reasonable, and also wrong in Kirkegaard's case.

Emil Kirkegaard's response.

Ross Douthat argues that we should just go all the way to true Imperialism. A position that I endorse. S. M. Stirling gives a notable fictional example of how an actual empire, a universal state, can be genuinely multicultural. Also, the Hapsburgs.

Right-wing terrorism in America

I saw this graph, and others much like it, frequently on social media. I think is is a good idea to look into things that seem contrary to your impressions, and this one seemed to challenge mine.

As a engineer, I also always like to look at the source data, so I sought out the article from the New America Foundation, Terrorism in America, that was cited. The graph below appears in the article, and sure enough, it looks like the one above, with a big spike for Jihadist victims in 2016, since the dataset has been updated since the previous chart was generated. You can download a set of .csv files from the New America Foundation that contains all of the data they have collected. I appreciate the effort that went into compiling this, and I also appreciate the New America foundation and the authors choosing to share their data freely.

Looking at that data set [downloaded 2016-12-19], I was a little surprised at what I found. Looking at American domestic attacks only [Mumbai is a big spike in 2008], out of 210 plots, 190 had been categorized as Jihadist, as compared to 19 Right Wing and 1 Left Wing.[and one with no data, the 2010 King Salmon plot] That is an order of magnitude of difference between all three categories!

I decided to plot the data in a number of different ways, which is often a good way of looking at trends in a dataset. I excluded anything outside of the US [foreign_attack=TRUE in the dataset]. I also just summed victims by year, rather than using a cumulative sum like the New America Foundation chart. Here are my charts:

One of the biggest differences is how often a given plot is prevented. Jihadist plots are prevented much more frequently than right wing plots. In the chart of prevented/not prevented, deadly attacks are not included. You can see that in the next chart down, which compares plots per year to the number of people injured or killed. I've said before that we are lucky that our enemies are so incompetent. Here is proof it is true.

The big differences are in rate of occurrence, and in rate of prevention. I don't know enough to really understand why the FBI and other American police forces and security agencies are so much better at preventing Jihadist plots than right-wing ones, but I think that given the 10x difference in rate of plots, they are doing the right thing.

By the way, I think my initial impressions ended up justified by the data.

LinkFest 2016-11-06

This LinkFest has been delayed three weeks. I had better publish it before the election and everything gets falsified!


Divided by meaning

A great piece on how Americans are divided by their attachments to hearth and home, or the lack thereof. Fascinating to me, since by education and career, I ought to be a member of what the author calls "the front row kids" who run the country, but I have chosen to live and work in the same small town I grew up in, much like Steve Jobs.

Is there a dietary treatment for multiple sclerosis?

It wouldn't be that hard to design a double-blind RCT on this if you wanted to. You could put everyone on the vegan+fish diet and then supplement animal fats in pill form. If the IRB balked, you might then suspect they secretly believed it might work.

The Crony Economy

Everyone agrees they don't like it, but no one has produced a lasting reform.

How Half Of America Lost Its F**king Mind 

Cracked continues to impress me. Companion piece: Divided by meaning

Is there a dietary treatment for multiple sclerosis?

A really good look at the incentives that push medicine towards pharmaceuticals and away from other kinds of therapies. Related reading: Lions, Tigers, and Bears. Is the placebo powerless?

Revealed: Nearly Half The Adults In Britain And Europe Hold Extremist Views

There is no end to the humor in this, but I find the commitment to democracy kind of sweet and endearing in people who are otherwise horrified when they find out what average people really think. Can you imagine the headline if one were able to conduct the same survey world-wide?

Why Tokyo is the land of rising home construction but not rising prices

Because they almost always tear down old houses and build new ones instead of just moving into them. There are a variety of interesting cultural and practical reasons for this, but one that doesn't appear in the article is the way the Yakuza use construction as their legitimate front. A lot of blue collar work in NYC works much the same way.

How Democrats killed their populist soul

Part of my on-going series of how the Economic Right and the Cultural Left are currently dominant in the West. Until I read this, I hadn't appreciated how the economic theories of Right and Left alike had turned against trust-busting and monopoly prevention.

Estates of Mind

A bit more about anti-trust laws as applied to intellectual property.

Mergers raise prices not efficiency

Since I have worked in manufacturing for my entire career, I don't find this surprising at all. The idea that mergers allow for standardization looks a lot easier on paper than in reality. Supply chains and manufacturing lines can't change with a memo.

The Rise of the alt-Right

Definitely one of the best things I have read about the alt-right. What is going on in the US has a lot of ties to what is going on in Europe.

On the reality of race and the abhorrence of racism

Bo Winegard, Ben Winegard, and Brian Boutwell point out that studying race doesn't make you deplorable.

The ruthlessly effective rebranding of Europe's new far-Right

I said what I read above from Scott McConnell was the best thing about the alt-right, but you really need this one as background.

The election that forgot about the future

In John's review of The Fourth Turning, one of the things that Strauss and Howe said made the Civil War worse than it could have been was the failure of the aging Transcendentalists to step aside and let someone else solve new problems. According to Strauss and Howe's model, the Baby Boomers are currently filling the same role in the United States. And you might note, one of two aging Baby Boomers is about to win the Presidency in a bitterly contested election.

Hacksaw Ridge

Mel Gibson makes another great war movie, about a guy who would not carry a gun.

LinkFest 2016-04-01

April Fool's Edition

The CDC is trying to make 86 million Americans sick

I've long thought the pre-diabetes thing was a bit foolish. While it is a good thing to be able to quantitate, if you don't understand what you are doing it can make you far too certain. Pre-diabetes is a lot like a risk-factor; something that is correlated with diabetes, but is in totality a poor predictor.

It's all Geek to Me

Neal Stephenson's review of the movie 300 is now nine years old, but I still enjoyed reading it. I liked 300 when it came out, and mostly for the same reasons Stephenson did.

My journey through Molenbeek

A nice synopsis of the way in which not particularly devout partially assimilated children of immigrants get radicalized.

These unlucky people have names that break computers

Parsing text is hazardous.

A researcher explains the sad truth: we do know how to stop gun violence. But we don't do it.

Unfortunately for this well-meaning researcher, his suggestions involve pattern recognition, which is currently disfavored.

Peak Water: United States water use drops to lowest level in 40 years

The story is similar for gasoline. Technological progress means we do more with less.

HVAC Techs — Hackers who make house calls

The kind of unglamorous but well-paid job Mike Rowe likes to talk about.

America may DUMP algebra as new study finds it is the main cause of high school drop-outs - and only 5% of jobs need it

This is a fantastic idea. We have raised the bar to graduate high school so far that we are penalizing people of normal intellectual ability.

Immigration and the Political Explosion of 2016

This is a recurring pattern in United States history.

Philosophical Reflections on Genetic Interest

Frank Salter's concept of genetic interest is a philosophical concept that is muddled up with a scientific one. Unfortunately, his philosophy isn't too sharp.

How much of the placebo 'effect' is really statistical regression?

Courtesy of the ever contrary Greg Cochran, a reason to doubt the placebo effect. Here is a recent blog post expanding on this idea, with further reading suggestions.

Bryan Caplan on the Signaling Theory of College

Bryan Caplan is a popular blogger and economist at George Mason University. Caplan was recently interviewed on EconTalk about the value of a college education. Short version: a college education doesn't have much intrinsic value. I'm simplifying a bit, but only a bit. Caplan argues that higher education is more important for sorting out the smart and hard-working from the rest than in teaching anything specific.  Of course, I might very well say that. I skipped out on grad school and went into the workforce precisely because I was convinced that more school wouldn't make me any smarter, or teach me anything useful. Of course, I also hated academia.

The signaling theory of education is nothing new to me. I've certainly pooh-poohed American higher education on multiple occasions here. However, it is easy to go too far. Caplan is too careful to say one learns nothing in college. What he is saying is that overall, and for the most part, specific skills are less important than intelligence, the capacity to work hard, and a willingness to play by the rules. These are the things college selects for. These are also correlates of success in America and similar societies.

On the other hand, I can certainly point you to plenty of disgruntled college graduates who cannot readily find work because they have the wrong degree. So there is a sense in which college functions in this signalling fashion that Caplan posits, but there is another sense in which employers, particularly in the much-vaunted STEM fields, really do expect college graduates to know things, very specific things. Personal experience suggests to me that this tendency is perhaps somewhat stronger than necessary to ensure competence, but one needs to understand the difference between ability and skill, or potentia and actualia. Since Caplan is an economist, perhaps he can look at the opportunity cost of hiring an able but unskilled graduate.

[Caplan] The human capital story says that you go to school; they actually teach you a bunch of useful jobs skills; you then finish and the labor market rewards you because you are now able to do more stuff. The signaling model says, no, no, no, no; that's not what's going on. What's going on is that people go to school; they don't actually learn a lot of useful stuff; however, the whole educational process filters out the people who wouldn't have been very good workers. So people who are lower intelligence, lower in work ethic, lower in conformity--those people tend to not do very well in school. They drop out. They get bad grades. And that's why the labor market cares. It's not that the school actually transforms you to a good worker from a bad worker. It's that the schooling, the school puts a little sticker on your head--you know, Grade A student, Grade B student, Grade C student.

...

[Interviewer] I have a natural skepticism about it. And I think a lot of labor economists do as well. And the reason is that it's an extremely expensive signal. So, you are saying, for 4 years, I give up the chance to work; I pay this tuition, whether it's $5000 or $10,000, or $30,000, or $40,000--at a private university. And for that enormous amount of money, I prove that I am a good worker and I get a sticker on my head. Wouldn't there be an easier, cheaper way to get the sticker? If all it's doing is measuring ability, this 4-year slog that's extremely expensive? That's the best way that people have come up with to get the sticker?

...

[Caplan] Yes, it's an arms race. And the fact, if it is a fact, the private return is high is really a very bad argument for pouring more money on. Now, the other point, as we were saying, the return that you should be looking at in terms of this argument of not being able to borrow against your future earnings--what you are looking at is return for the marginal people who are just on the edge of going or not going. And as we've seen, the return for those people is actually not, is actually quite mediocre. And then finally if you adjust for ability and everything else, really I would say that once you appreciate signaling you realize that, so we have subsidized education way past the point of [?] returns. So by my calculations, actually, the social return to education is now quite negative. And it would be a much better policy to drastically scale it back, so rather than encouraging more people to go, I think it's better to discourage them from going or at least to encourage them less. So in fact--so, the biggest policy implication that's going to come out of my book is we just have way too much education. I call this the white elephant in the room. There are way too many people going to school, maybe not from their own selfish point of view, but certainly from a social point of view to go and pour more money on this really is just throwing gasoline on the fire. And we need to do less of it.

h/t DarwinCatholic

A decline in science?

Greg Cochran continues to be one of my current favorite reads. I learned something entirely new to me, the breeder's equation, from his recent post of the same name.

R = h2 S.

R is the response to selection, S is the selection differential, and h2 is the narrow-sense heritability. This is the workhorse equation for quantitative genetics. The selective differential S, is the difference between the population mean and the mean of the parental population (some subset of the total population).

In an aside, Greg also explained regression to the mean using the breeder's equation. I hadn't truly understood what regression to the mean meant until I saw Greg's explanation. Greg is a physicist who went into evolutionary biology, so it isn't surprising to me that I find his point of view readily comprehensible.

The comments on Greg's blog are also very informative, although truth be told sometimes I just browse the comment threads for Greg's smackdown on some poor soul. There were a lot of good links in the comments, and some books I wish I had time to read. Greg and Bruce Charlton have the opportunity to exchange words on the recent Victorian reaction time data, needless to say, Greg is not impressed.

Where it really gets interesting, is when Greg brings up an objection to a decline in intelligence that I also thought of: with that big of a drop, you would see a complete collapse of intellectual endeavour at the highest levels, by which Greg means math and physics. Bruce rejoined: of course it has! Greg's rebuttal was scoffing, Bruce's counter equally scoffing, and ultimately Greg fell back on the argument from authority: Bruce is ignorant of the current state of math and physics, and so cannot really assess them as well as Greg can.

I've pondered whether the rate of intellectual discovery is slowing, but the data I looked at was patent filings and other technological items rather than science per se. Recently, I also suggested there is something intellectually barren about the twentieth century compared to the rest of modernity, especially in the way the guiding lights of Western Civilization seem frozen in time.

Yet for all that, I side with Greg here. Bruce has never struck me as a very quantitative thinker, so I think Greg's accusation is largely correct. Also, despite it's reputed fallaciousness, I think the argument from authority is perfectly appropriate when assessing a question of this type. Greg really is better placed than Bruce to assess math and physics, because he understands these subjects better.

I think the one way Bruce's argument still has force is the way in which science is becoming disconnected from technology and society. The early modern scientists were all about useful inventions, in contrast with the ancient and medieval scientists who were more interested in knowledge for its own sake. Science seems to be trending back towards the knowledge for its own sake model, and as a consequence most people are less interested. When Bruce says,

"I suppose this is hard to settle – but to an outsider, a collapse in the highest levels of physics has very obviously happened, since the quality of the best modern physicists seems qualitatively *much* inferior to those of, say, 100 years ago – or, to put it another way, the supply of geniuses has almost completely dried-up. Maths looks much the same – but I don’t know so much about it."

What I really hear is: scientists of 2013 are less famous than scientists of 1913. Sure, there are TED talks, but nobody is lining up to hear Faraday's lectures on electricity any longer. Now it is Malcom Gladwell. Nobody really thinks Gladwell is part of the leading edge of science himself. Part of the problem here is modern physics and mathematics are way over the public's heads'. I can almost hear Bruce's rebuttal here, but I think the problem is the work has gotten much more abstract over the last two centuries, so much so that it hard for non-specialist scientists to truly understand. Faraday or Maxwell or Darwin have a far wider audience, because their work is both simpler and broader. Greg suggested to Bruce that he study the Princeton Companion to Mathematics, and then they could discuss the subject fruitfully.

I am far from Greg's level, but I don't have any doubt that both mathematics and physics have progressed in impressive ways, on their own terms. What this has not done, however, is given us flying cars. So, no one cares. Science has abandoned it's Baconian roots for an older Aristolielian set of roots, but no one seems to have noticed yet. Plus, nobody really knows what Aristotle actually said anymore, but that is a different post.

Has Intelligence Declined in the Modern Era?

I mentioned this subject a couple of times before.

Simple reaction time data from Woodley et al.Bruce Charlton is the first person I know of to discuss the implications of the change in simple reaction time data in the paper by Woodley et al.  Charlton claims the data shows a 1 standard deviation reduction in IQ based on the correlation [r = .3] between simple reaction time and IQ or the g factor. I remain unconvinced that that intelligence has declined in the last 100 years. For one, that is a terrible fit [see scatterplot]. For another, with that kind of correlation, I need better evidence that the model is correct. Charlton's best arguments focus on the way in which intellectual inquiry in general still seems to be dominated by the same famous men of the last century.

My late friend John Reilly used to sum it up thus:

How do we know that the 500 or so years of the Modern Age are at last drawing to a close? Lukacs's answer to this, and it's a good one, is to draw attention to the remarkable intellectual barrenness of the 20th century. In 1914, when the century began to manifest its characteristic features, the guiding spirits of the time were Freud and Marx and Darwin and Einstein. In 1989, when in a political sense the 20th century was already over, the guiding spirits of the time were Freud and Marx and Darwin and Einstein. There was no other century of modern times that produced so little new intellectual history. Indeed, all but the earliest part of the Middle Ages was livelier.

In a general sense, I agree with Charlton that the rate of the big gosh-wow discoveries in science seems to have slowed remarkably. I've come to the same conclusion myself. Where I differ is the cause. It is not apparent that the problem is our minds have gotten weaker. Rather, I think we [we meaning the West] have chosen to focus on other things. The problem is cultural.

What I lack here is any sort of quantitative data. My sense is that the tasks to which we apply ourselves have if anything, gotten harder. We just are getting less output partly because our best minds think about other things, and partly because the easy scientific discoveries have already been made. Now if only I could prove it.

There is a cracking good discussion of the technical details of the SRT paper over at Greg Cochran and Henry Harpending's blog, West Hunter, that makes some of the same points I want to make here. However, I am more interested in the cultural aspect than the scientific aspect at present.

I think we don't make big scientific discoveries anymore because we don't want to. This seems strange, this sort of thing is frequently in the news and a subject of discussion. However, if you look at our priorities as a civilization, we have decided to do other things. You might also say that our civilization is decadent. I use this in the technical sense proposed by Jacques Barzun in From Dawn to Decadence: 500 Years of Western Cultural Life 1500 to the Present 1st (first) edition, a decadent society is one that wills the ends, but not the means. We say we want big science, but we have a remarkably inefficient way of doing it.

So what are we interested in? I think the things we are interested in are actually very good things, but they don't produce good headlines. For example, one thing the Victorians were terrible at was industrial hygiene and product safety. A fad for green dye made with arsenic in Victorian England lead to thousands of deaths. The manufacturing process for safety matches produced a condition known as phossy jaw, which could actually make your bones glow in the dark before it killed you. The discovery of radium by Marie Curie lead to the incorporation of radium into a wide variety of products, contaminating much of Paris. It is not very hard to multiply examples like this ad infinitum. The Victorians discovered lots of new things, but they left a lot of collateral damage in their wake in the form of dead and maimed factory workers and widespread environmental pollution.

There are now systems in place to verify that products are safe before they enter the market. Designing nearly any product is vastly more complicated now than 100 years ago, not only because science and engineering have progressed and we are pushing the limits of our knowledge, but because you have to make certain every material and manufacturing process you use is safe not only for consumers, but for the workforce that creates it. The big scientific discoveries made life better for everyone on average, and the systems and techniques and regulations we have developed help make sure each individual is able to benefit from that knowledge personally. We have chosen to spend our effort on mitigating the clear downsides of progress, rather than plowing ahead with new discoveries.

This dynamic is illustrated by the development policy the Chinese seem to be following: explicitly trading pollution and worker safety for faster progress in order to catch up with the West. A brutal policy, but in fairness to them, didn't we do the same thing, with perhaps a little less self-awareness? The current Western policy is far more boring, even though it probably has greater benefits for the common man. For example, who knows the name of the man who invented the plastic barrels that sit at exit ramps and highway abutments? That man has saved thousands, maybe tens or hundreds of thousands of lives. Preventing something from happening will always be less interesting than making something happen, no matter how useful it is. Yet, our culture happens to promote this. Is that really a bad thing?

Simple Reaction Time Data

There has been a story trending recently in the news about the Victorians being cleverer than us. I've been following this story since Bruce Charlton broke it in February of 2012. I've never been that impressed, but I've thought about looking up the data to see if its as overblown as it sounds. I still haven't done that, but William Briggs does provide us with a scatterplot from the paper:

Intelligence and reaction times Woodley et al.Nice model-fitting boys. Keep up the good work.

Immigration and High-Skill Employment in the US

I had been wanting to look into this subject, and lo, I find that someone else has already done it.

Guestworkers in the High-Skill U.S. Labor Market.

I've long been dubious that we don't have enough native-born STEM graduates in the US. My reasoning is based on my own experience rather than exhaustive data, but I think the things I have noticed are nonetheless potent arguments. First, wages in technical fields are good, but not that good. If we truly had a shortage, in the economic sense, of programmers or engineers or what have you, wages should respond appropriately. If there were a shortage of programmers, I would expect to see companies fighting to pay 18 year old high school graduates $50,000 a year. A real worker shortage looks like thisInstead, you can make that kind of wage, but you need at least a bachelor's degree, if not a master's. Programming is still a field where a true talent can rise to the top regardless of credentials, but in any discussion of larger job trends, such folks are an exception.

Mark Zuckerberg has been in the news lately pushing for more H-1B visas for high-skilled tech workers, but many tech moguls have been pushing for expansion of the H-1B program for many years, along with other different, but related programs. Zuckerberg et al. have been claiming that there aren't enough quality workers available in the US, but I find this less plausible than they simply want to keep their labor costs down by importing cheaper workers.

I can think of one explanation that would support the claim there are not enough skilled STEM workers available in the US. Plausibly, one could claim the existing industry has already skimmed the high-IQ high-motivation graduates, and the remaining workers aren't good enough to do the needed work. However, I don't think the available data supports this either. Let's look and see. Here is what the EPI report notes:

Consider that U.S. students make up one-third of the entire global population of high-performers on tests of science knowledge.

...

For STEM graduates, the supply exceeds the number hired each year by nearly two to one, depending on the field of study. Even in engineering, U.S. colleges have historically produced about 50 percent more graduates than are hired into engineering jobs each year.

ln the computer and information sciences field, which is where nearly half of high-skill guest workers are employed, we see that about two-thirds of college graduates find jobs in their field 1 year after graduation, and of those who don't find a job, nearly half cite low pay and poor working conditions as the reason. For some of these, I would find it plausible that they weren't actually qualfied for the job. There is a skill distribution amongst college graduates, and not everyone is qualfied to do every job in this field. However, it does seem that the higher number of guest-workers in IT compared to engineering may be associated with a higher number of US college graduates being unable to find IT work because the pay is too low.

College grads working in fieldReasons for not finding work in field

This matches the rate of growth we see in guestworkers compared to US citizens and permanent residents graduating with degrees in computer and information sciences.

Guestworker growth

The authors of the study go on to look at the responsiveness of the domestic supply of graduates to wages over time, looking at both IT and petroleum engineering. In both cases, they find that increasing wages lead to increasing numbers of graduates in the field, while decreasing wages lead to a decrease in domestic graduates, with a suitable time lag. So in a sense, Zuckerberg is correct that the domestic supply of graduates in the computer and information sciences field is decreasing. However, the problem is he's not paying well enough, and importing workers only makes the problem worse. Unless, the domestic supply isn't what Zuckerberg is after, in which case he's getting exactly what he wants. Last time I checked, Mark Zuckerberg is really, really smart. Which is more likely?