Links 7/15: Link-Carbon Battery
Contrary to claims of dumbing-down curriculum, it looks like schools are assigning tougher reading material at earlier ages than in the past. To riff off Woody Allen, “I read War and Peace in tenth grade. It involved Russia.”
Jai on the Copenhagen interpretation of ethics, ie “people are punished for touching a problem without solving it.”
Nowadays we say the brain is like a computer. Fifty years ago, people said the brain was like a telephone switchboard. Three hundred years ago? Human beings are basically made of fireworks.
In case you were insufficiently convinced about the crime-IQ link, the latest study from Finland demonstrates it conclusively across > 20,000 people – or you could skip the text and just look at the graph (since the same effect is found on self-report of crime, it’s not just “stupid people get caught”). h/t Stuart Ritchie.
Lovecraft fans: did you know the dhole is a real animal?
The latest promising potential obesity treatment is celastrol, a chemical isolated from the awesomely-named Thunder God Vine. I have already started confusing this with Thorazine for the obvious reason. [EDIT: This may have excessive side effects]
Mississippi is in special trouble for having the Confederate banner on its state flag, but you should know that every state flag is horrible.
Silk Road and similar bitcoin dark markets keep either getting their organizers arrested or turning out to be scams. Now some coders have a vision of creating a distributed dark market as decentralized and open as bitcoin itself. Not for illegal things, mind you. Just for totally legal things that you happen to want to sell via an untraceable currency on a heavily encrypted site. Something something call up something can’t put down.
Space probes that fly by the Earth for a gravitational assist seem to gain speed in a way inconsistent with known laws of physics. Possible explanations include dark matter halo around Earth.
Big new study claims to have found that black-white outcome differences are almost entirely a result of how children of different races are raised. They compare black people with white mothers to black people with black mothers, and find that black people with white mothers do just as well as white people; then they conclude that this means it’s the way you get raised by your mother which determines which racial pattern you fall into. Interesting result, but seems to have no awareness of the fact that people who identify as black but have white mothers are most likely half-white – this reintroduces discrimination as an explanation (since half-white people presumably will have whiter appearance and people might not recognize them as black or discriminate against them as much) as well as genetics. One might argue that in genetic or discrimination explanations they should at least be halfway between all-whites and all-blacks, but I don’t really see the study doing the analysis necessary to make that argument. Although the blindingly obvious next step is to look at children of black mothers and white fathers, either I am missing this in the study or it isn’t done. Overall interesting methodology but very disappointing; I only skimmed it but I’m interested in seeing further analysis.
Closely related: The IQ Gap Is No Longer A Black And White Issue. What should we conclude from African immigrants to the US (and their children) being among the highest-IQ and most successful people in the country, while ‘native’ African-Americans do much worse? Does this argue against discrimination-based explanations for outcome gaps, since immigrants’ blackness doesn’t seem to hurt them at all? Does it argue against genetic explanations, since both groups have broadly similar genetic makeup? (but the article seriously misunderstands regression to the mean, which probably dooms us to a slog through quantitative arguments about exactly how strong selection effects can be – see also here ). Does it argue for something about “poverty traps”? Should we take it as corroborating evidence for the above study that it has to do with cultural transmission and child-rearing?
And related: statistician looks over three of the most celebrated papers purporting to find evidence of racial discrimination and finds that they seem to have heavily fudged numbers, almost as if there were some sort of “ulterior” “motive” “incentivizing” people to publish shoddy research in the field. I don’t really understand the test used and would be grateful to someone explaining it to me using short words.
Not related: the cookbook
For-profit colleges are charging tens of thousands of dollars for courses that tend to leave most of their graduates unemployed or working below-poverty-line jobs. Cosmetologists and medical assistants hardest hit. Now the government is cracking down.
Stuart Armstrong’s top ten myths about AI risk. I would have also included “people only worry about AI risk because they think superintelligent AI is very near”.
Drone strikes in areas associated with decreased terrorism in those areas.
Scott Sumner on the startling level of inequality between the northern and southern parts of various European countries, especially Italy.
Master craftsman Shing Myongsik is apparently very serious about selling this game board
A very comprehensive meta-analysis by the Joseph Rowntree Foundation gathers studies showing that giving poor families money improves children’s outcomes considerably – for example, closing the financial gap between rich and poor kids would close half of the school achievement gap. Some obvious caveats – it doesn’t seem to have been published in a peer-review journal, and the Foundation seems to be a think tank associated with the cause of giving poor families money – but the research and methodology seem solid as far as it goes. My main caveat, and it’s a big one, is that pretty much all their studies measured very near-term outcomes – eg giving families money this year improves their childrens’ grades in school this year. No proof that there is any effect on outcomes that actually matter, like whether the children break out of poverty when they grow up. On the other hand, one would have a hard time arguing that getting better grades throughout your childhood and so getting into a better college or something doesn’t give you a leg up when you’re older. I can’t remember exactly which, but seems like a good place to point out that a lot of these study links are h/t Nathaniel Bechhofer and Ben Southwood.
But this is SSC and we are totally going to try! Estimating The Return To College Selectivity Over The Career Using Administrative Earnings Data by Dale and Krueger (of minimum wage fame) finds that, after accounting for the unobserved differences in ability that allow some students to get into better colleges than others, the return to going to a better college is basically zero. An unrelated analysis in the Economist on a similar topic comes to the same conclusion with a striking graph. [EDIT: Robin Hanson tears this apart]
Also, a while back, China increased years of required schooling, leading to a perfect quasi-experiment to see whether more schooling led to better outcomes. Authors conclude that “our results are consistent with the signaling story [ie education being mostly about signaling]; further consistent with such a story, we estimate that the labor market return to another year of schooling is very small, though greater for the less-educated.”
Some people volunteer to moderate for AOL. Later, they decide to launch a class action suit against AOL for not paying them for the volunteer work they volunteered to do as volunteers. They argue that they worked really hard, so it was kind of like being an employee, and so AOL owes them lots of back pay for all the work they weren’t compensted for. The court decides this makes perfect sense and AOL eventually settles for $15 million. Am I misunderstanding the legalities here, or has volunteering just been made illegal? [see discussion in comments]
Florida is the only state with its own embassy in Washington DC, for some reason.
There’s a common talking point that the US’ private health system is demonstrably inferior to Europe’s public health systems because it spends more money to get worse results. I was recently linked to two articles that challenge that assumption. Random Critical Analysis argues that GDP per capita is the wrong measure to compare healthcare to, and when we use the right denominator the US’ health care expenditure is no more or less than expected given its size and wealth; I don’t know enough economics to be sure it’s right, but I admire the number of graphs included in the argument. And a team from University of Pennsylvania claims that the US’ sub-par life expectancies are unrelated to the healthcare system, although their methodology consists of some measurements of health care quality that for all I know could be cherry-picked or nonrepresentative.
Language Log on the International Classification of Diseases, the official medical coding manual which goes above and beyond in trying to have a code for every possible reason somebody might come to a doctor – plus extra commentary on the inherent difficulty of ontologies. They include some classic ICD codes like “bitten by orca”, “burn due to water-skis on fire”, and “struck by tortoise”. But I have had occasion to use the ICD myself and would add personal favorites V97.33 (sucked into jet engine), V90.23 (drowning due to jumping from burning ship), W58.13 (crushed by crocodile), W50.3 (accidental bite by another person), Y38.5X3 (terrorist injured by their own nuclear bomb), Y92.72 (injury occurring in chicken coop) and W22.02 (walked into lamppost).
Interfluidity gives the Greeks’ side of the Greek financial crisis; Tyler Cowen gives the creditors’ side. Somewhat related: at least according to Forbes (whose ideological committments give them little reason to lie), despite their lazy reputation Greeks work the longest hours in Europe [EDIT: but see comment].
I get accused of overusing the word “orthogonal”, so I want you to know that the Supreme Court has officially ruled that this is cool. (h/t Oscar Cunningham)
Relevant to previous discussion on this blog: Colorado finds that giving out lots of birth control does, indeed, decrease teenage pregnancy rates.
This is a complicated paper and I’m not sure I’m understanding it right, but I think it says that genetic factors and childhood trauma both play independent roles in increasing people’s risk for psychosis, and that the chilhood trauma -> psychosis link isn’t just an artifact of genetics where psychotic people have psychotic parents who cause them trauma.
I was sure that this picture of a deep blue tarantula was photoshopped, but it’s actually a pretty typical poecilotheria metallica.
Congratulations to everybody starting, graduating, or progressing up in medical residencies this July (the rest of you might want to stay away from hospitals for a while). If you’re feeling overwhelmed, you (and medical students also) might want to check out a new book written by LW/SSC community member / new radiology resident Peter Wei: Learning Medicine: An Evidence Based Guide. It talks about how to best use spaced repetition software and similar tools to memorize and understand medical knowledge as quickly as possible.
A lot of the links this week have been studies, some more credible than others, purporting to find that genetics don’t matter as much as we thought. We had better hope they’re right: according to Pew only 46% of Americans support genetic engineering to reduce disease risk, and only 15% of Americans support genetic engineering to make kids more intelligent.