Month: September 2021

I generally try not to let Twitter beefs contaminate the blog, but at least a few people last night denounced me as an idiot on the subject of airborne aerosol transmission of COVID-19. This seems to have turned into something of a truther jihad, so it's worth running down the history of this.

The whole thing started with a suggestion that perhaps Chinese virologists knew from the start that COVID-19 was transmitted via airborne aerosols while Western scientists insisted it was spread by droplets that were too big to stay suspended in the air for long. One explanation for this disagreement is that the Chinese had learned a lesson from the SARS epidemic of 2003.

But this isn't really so. Here are the basics:

• For reasons outlined a few months ago in Megan Molteni's excellent account in Wired, the epidemiological community has for years believed that aerosol transmission is only likely for particles less than 5 microns in diameter, and since most viruses are enveloped in packages larger than that they can spread only a few feet before falling to the ground. This mistaken belief was caused by a misreading of one particular study followed by inconclusive research over the next few decades. By the 1980s the tide was turning slightly on this, but scientists mostly still believed that aerosol transmission was unlikely with large particles.
• When SARS hit in 2003, epidemiologists in China believed it was mostly spread by droplets, not airborne aerosols. There was one large outbreak in an apartment complex that seemed to implicate airborne transmission, but eventually, after practically tearing the apartment complex apart, scientists concluded that the culprit was droplets spreading through a faulty sewage system. However, other studies disagreed, suggesting that airborne transmission of SARS may have played a modest role in certain kinds of settings.
• Fast forward to late 2019. The initial reporting from Chinese scientists once again suggested droplet transmission, though with some possibility of aerosol transmission. Generally speaking, there was no difference between Chinese and Western scientists on this, and Chinese treatment of COVID-19 outbreaks didn't focus on airborne transmission.
• Over the next few months, evidence for aerosol transmission began to grow. Partly this was due to new studies. Partly it was due to new evidence showing that COVID particles were not always larger than 5 microns. And partly it was due to the intervention of the physics community, which wasn't blinded by the old 5 micron standard and knew that aerosol transmission was perfectly plausible for particles considerably larger than 5 microns. By late summer of 2020, it was pretty well agreed that airborne aerosol transmission was one of the ways COVID-19 spread.

You can decide for yourself how badly this speaks of the epidemiological community. On the one hand, they obviously misjudged the role of aerosol transmission of COVID-19. On the other hand, this was a matter of considerable research in their area of specialty, and that's not something you abandon instantly based on one or two early studies.

Unfortunately, as I mentioned earlier, this subject has become something of a jihad among certain people, who are convinced that there was overwhelming evidence of airborne transmission within weeks of the first case, and anyone who denied it was simply stupid. I don't think there's any way of talking them out of this belief, but it really doesn't fit the facts. Epidemiologists clearly made a mistake here, and some of it was indeed based on stubbornness. But it was also based on a laudable desire not to overturn decades of belief overnight until the evidence was clear. There are not really any big villains here.

UPDATE: I mixed up microns and nanometers in the original post. Sorry about that. It's fixed now, along with some more precise language about exactly what it is we're measuring.

The New York Times reports that Latino support for California's governor was high when he was first elected in 2018:

But as Gov. Gavin Newsom tries to prevail in a recall election in a matter of days, the very Latino voters he is relying on appear to be disengaged and ambivalent about the prospect of his being ousted from office.

....For many Latino voters, the mixed feelings stem from a continued struggle with the pandemic, as they face higher infection and death rates, as well as unemployment. For others, there is a deep disconnect with the Democratic Party and Mr. Newsom himself, a multimillionaire Napa Valley winery owner whom they view as aloof and distant.

This represents a great mystery. As a Californian I have read at least half a dozen different versions of this story, and all of them offer different explanations for Newsom's drop in Latino support. Nobody really seems to know for sure what's going on, which is odd since there's no lack of reporting about Latino communities in California.

COVID-19 is a plausible, though not certain, explanation. Latinos contracted COVID-19 at rates 2-3x higher than whites or any other ethnic group, and crucially it struck young Latinos at unusually high rates. On the other hand, the death rate from COVID-19 was only slightly higher among Latinos than among other groups.

Do Latinos actively blame Newsom for this, or are they just generally unhappy about the pandemic and have lost faith that it matters much who's running the state? That's unclear, and it highlights a weird disconnect in American politics right now. Newsom is generally given good marks for his handling of COVID-19, but he's unpopular. But in red states where governors have pointedly refused to take the pandemic seriously and death rates have skyrocketed, governors remain popular.

My take on this is twofold. First, of course, is the culture war aspect: lots of red-state voters care more about their governors sticking it loudly to the man than they do about case rates and deaths. Second, in many states, people care more about keeping stores open and jobs available than they do about fighting the pandemic with annoying regulations. You can tag this as good policy being punished because it causes hardship, or you can say that it simply represents different preference curves in different states. Either way, it doesn't seem to have done Newsom any good.

Of course this is just one theory for why Latinos seem to have soured on Newsom, and the results aren't in yet anyway. Maybe when the ballots are actually counted, Latinos will break for Newsom at their usual rate.

In the meantime, Californians should be aware that the recall is exactly one week away. If you have a mail-in ballot—which you should—vote NO and send it in now. Don't put it off and then forget about it until the last minute, when you risk a chance of it not counting. All you have to do is mark one box (or two if you care about the bottom part of the ballot) and seal it up. It takes about one minute.

POSTSCRIPT: But what about the bottom part of the ballot, which lists all the candidates who are competing for the governorship if the recall succeeds? Who should you vote for?

I honestly don't think it makes much difference, but for the record I marked my ballot for San Diego mayor Kevin Faulconer. I didn't give this a great deal of thought beyond the fact that Faulconer is non-insane, and among California Republicans that's a rare commodity these days.

I don't have any pictures of people laboring, so instead here's a picture of the new Gerald Desmond bridge in Long Beach harbor. It's at least a product of labor.

I've shown you the bridge before, but this picture is the result of an exhaustive search for the best possible view I could get. My conclusion is that there isn't one. This particular angle was better than any other I could find, but the weather was bad on the day of my search so I went back the next week to actually take the picture. By then, however, a whole pile of electrical lines had been put up. What's more, this view is from a vacant lot that turned out to be private property, so I got chased off.

Eventually they'll dynamite the old bridge and it will then be possible to get a good view from the other side. However, that's still a couple of years away.

Do genes play a role in cognitive traits like shyness, memory, language skills, and so forth? Of course they do. The only real question is how big a role they play. Take language, for example. In the case of which language you speak, the role of genes is 0%. It's all environmental. But in the case of how well you speak a language, it's much higher.

This is obvious enough that it should be uncontroversial, and it would be except for one thing: if genes have an effect on cognitive traits, it means that genes have an effect on the trait we call intelligence. And if genes affect intelligence, then it's possible that racial differences in intelligence are partly the result of genetic factors.

The fear that this could be true, even though the evidence is currently against it, has driven left-wing opposition to the whole notion of genes and behavior for decades. Luckily for everyone involved, the evidence for the impact of genes is based almost solely on ecological studies, usually of twins. But no matter how suggestive such studies can be, they will never be proof positive of anything. The only thing that has a chance proving anything is biochemical: that is, finding specific gene complexes that affect personality traits. That's basically impossible, which meant everyone could keep merrily arguing forever, safe in the knowledge that no one would ever be conclusively proven wrong.

But lots of things are impossible until suddenly they aren't. Readers with very good memories may recall that a few years ago I wrote about Genome Wide Association Studies (GWAS, pronounced jee-wass). These do the impossible: they allow genetic researchers to find single-nucleotide polymorphisms (SNPs, pronounced snips) that are associated with cognitive traits. At the time, I linked to a paper that claimed to have found SNPs that explained about 5% of the variance in intelligence. But work was ongoing, and the latest studies have gotten up to 20% or so. There's no telling where this number will eventually end up, but it's almost certain that within a few years we'll get to one that's high enough to prove to all but the most recalcitrant that genes do in fact have a considerable effect on human intelligence.

Why mention this? Because in its current issue the New Yorker has a profile of Kathryn Paige Harden, a professor of psychology at the University of Texas at Austin who has written a new book, The Genetic Lottery: Why DNA Matters for Social Equality. Harden has been doing GWAS work of her own and her conclusion is unsurprising: both genes and environment play significant and intertwined roles in most cognitive traits. But there's a depressing coda:

In my conversations with her colleagues, Harden’s overarching idea was almost universally described as both beautiful and hopelessly quixotic....James Tabery, a philosopher at the University of Utah, believes that underscoring genetic difference is just as likely to increase inequality as to reduce it. “It’s truly noble for Paige to make the case for why we might think of biological differences as similar to socially constructed differences, but you’re bumping into a great deal of historical, economic, political, and philosophical momentum—and it’s dangerous, no matter how noble her intentions are, because once the ideas are out there they’re going to get digested the way they’re going to get digested,” he said. “The playing board has been set for some time.”

"Hopelessly quixotic" is a fancy way of saying that no matter what the science says, Harden will never convince people on the left. As Harden puts it, the life of a behavior geneticist resembles “Groundhog Day.” Always the same arguments no matter what.

In fairness, the reason for lefty intolerance of cognitive genetics is obvious and righteous: It's been violently misused for a very long time as a way of proving that certain kinds of people are inferior to others. As Tabery says above, the playing board has been set, and it's almost certain that any new results, no matter how carefully explained, will be used as an excuse by some people to dismiss the possibility of ever improving the lives of the poor, the black, and the oppressed.

But as understandable as this is, it has a big problem: it looks as if we're getting close to a genuine understanding of how genes affect cognitive traits—and the answer is not going be "they don't." At that point the left had better have an argument to make, because they're certain to lose if they just bury their heads in the sand.

The funny thing is that I've never entirely understood lefty opposition to the notion that genes have a significant impact on cognitive abilities. My view has always been close to Harden's: if genes do have an impact, then it makes the case for social safety nets incomparably stronger. It becomes impossible to argue, for example, that poor people are merely lazy if you can point to SNPs that have a clear association with poverty. At that point, it's provably the case that being poor is mostly a matter of bad genetic luck. So what argument is left for leaving anyone in poverty?

Beyond that, as Harden points out, if you know the genetic foundations for a particular trait then it's easier to disentangle its genetic and environmental causes. This makes it easier to accurately identify the environmental causes, which in turn makes it more likely that you can construct social interventions that actually work. In other words, knowledge of genetics is a key part of the liberal project of doing everything we can to improve lives via social programs that are truly effective.

But most people don't see it that way. And beneath it all lurks the deep fear that someone doing GWAS research is eventually going to find SNPs associated with both race and intelligence. I continue to think that's unlikely in anything more than a trivial sense, but I may be wrong. And if I am, what are we going to do?

Hopper is doing fine. She has grown more accustomed to the cone and is no longer banging into things as she walks around. We've also discovered the easiest way for her to eat while wearing the cone, and she's pretty happy with that.

Hilbert is a whole different story. He is completely nonplussed by events in his house, and apparently he doesn't recognize Hopper at all. Whenever they're close, he hisses and then takes the widest possible path around her. He seems to think that he's been invaded by some alien creature carrying around a book called To Serve Cats. When Hopper shakes her conehead Hilbert jumps a foot in the air and scurries away.

But I guess we should have expected this. A few years ago a friend of ours came to Southern California to pick up a kitten, and afterward he spent a couple of days with us. The kitten was a tiny furball about 10 weeks old and Hilbert was terrified of it. A kitten! But Hilbert spent a good part of the next two days hiding under the bed.

Cowardice, thy name is Hilbert.

This chart has become so depressing I can hardly stand to post it anymore. Our death rate from COVID-19 is nearly three times higher than the next highest country and more than double that of Europe as a whole. Ditto for our case rate. And both are rising at a much faster rate than in Europe.

Is this entirely due to our lower vaccination rate? Or our mask idiocy? Or something else? If you look at excess mortality instead of official COVID data, the US looks fairly normal compared to Europe. Is this all just an artifact of different countries accounting for deaths differently?

We have lots of job openings these days and lots of unemployed workers. So why aren't those jobs being filled? The Wall Street Journal offers up this explanation:

Employers today rely on increasing levels of automation to fill vacancies efficiently, deploying software to do everything from sourcing candidates and managing the application process to scheduling interviews and performing background checks. These systems do the job they are supposed to do. They also exclude more than 10 million workers from hiring discussions, according to a new Harvard Business School study released Saturday.

....Many company leaders—nearly nine out of 10 executives surveyed by Harvard—said they know the software they use to filter applicants prevents them from seeing good candidates. Firms such as Amazon.com Inc. and International Business Machines Corp. said they are studying these tools as well as other hiring methods to understand why they can’t find the workers they need. Some said the technology can be changed to serve them better.

Hmmm. The robots are attacking humanity by interfering with our labor market. Interesting. Here's another view courtesy of labor economist David Autor:

Let’s start with the causes of the current labor shortage. Research on this question is unambiguous: We don’t know what’s going on.

I think I prefer this explanation.

John McWhorter says we should quit arguing about how to teach reading and just accept that we already know perfectly well how to do it:

In a word, phonics....Phonics works better for more children. Project Follow Through, a huge investigation in the late 1960s led by education scholar Siegfried Englemann, taught 75,000 children via the phonics-based Direct Instruction method from kindergarten through third grade at 10 sites nationwide. The results were polio-vaccine-level dramatic. At all 10 sites, 4-year-olds were reading like 8-year-olds, for example.

....However, there is a persistent disconnect between the world of reading science and the world of people teaching children to read. Only 15 percent of programs training elementary-school teachers include actual instruction on how to teach children to read. There remain people who favor the whole word method, or a combination of whole word and phonics, or even no particular “method” at all.

There's a wealth of research that confirms this, but unfortunately reading instruction has become part of the culture wars, with conservatives taking the side of phonics while university education departments tend to favor other methods.

This is unfortunate. Phonics works, and to the extent that you can invent add-ons that are potentially a little bit better it's really not worth the effort. DI-based phonics instruction is so good that we'd be a lot better off simply making it universal since it works well with both poor and affluent children. In addition:

There is a racial angle to this....We have known how to teach Black children, including poor ones, how to read since the Johnson administration: the Direct Instruction method of phonics. In this case, Black children don’t need special materials; districts need incur no extra expenses in purchasing such things. I consider getting Direct Instruction to every Black child in the country a key plank of three in turning the corner on race in America (the other two are ending the War on Drugs and sharply increasing funding and cultural support to vocational education).

Liberals should get on this train. Stop resisting just because conservatives have been pushing this for decades. In this case, they're right.

Poor little Hopper is stuck in a cone for the next ten days:

Around noon yesterday we noticed a long string hanging off her tail. At the base of the string there was a small wound covered in blood. So I headed off to the vet.

Sure enough, she had injured her tail in some way, and then spent the past day or so worrying at it and pulling off a long stretch of skin and tendon. So now she has a bandage on her tail and a cone around her neck. We are crossing our fingers that this is enough. If it's not, and there's nerve death toward the end of her tail, a little piece will have to be amputated.

Hopper is not adjusting well to the cone: she keeps bumping into things like a drunkard, and then getting stuck. We're taking her in today to get the bandage rewrapped, and I'll ask about this. She seems to be getting a little better as she explores the house, but still, I've never quite seen this behavior in a cat with a cone and I wonder if it means there's something off about her eyesight.

Here is a headline and subhead from the New York Times today:

I continue to be stupefied by this. Why are people so convinced that both of these things can't be true at once? A military evacuation of 120,000 people is not a NASA moon landing. It's inherently turbulent, crowded, and desperate. The kinds of things the Times describes are practically the baseline expectation in an evacuation of this size, even if it took two months instead of two weeks.

So yes: rogue flights, crowded tents, hope, and chaos.

And also yes: An enormous number of people rescued with minimal casualties and now on their way to asylum. It really was done about as efficiently as possible.