Links 8/15: Linkety-Split

Guys, I think Thomas Schelling might be alive and working in a Kentucky police department: Police offer anonymous form for drug dealers to snitch on their competitors.

Journalists admitting they’re wrong is always to be celebrated, so here’s Chris Cilizza: Oh Boy Was I Wrong About Donald Trump. He says he thought Trump could never sustain high poll numbers because his favorability/unfavorability ratings were too low, but now his favorability/unfavorability ratings have gone way up. But remember that favorability might not matter much.

Speaking of Trump – Why Securing The Border Might Mean More Undocumented Immigrants (h/t Alas, A Blog). Related: A Richer Africa Will Mean More, Not Fewer, Immigrants To Europe. So, if I’m reading this right, the best way to minimize illegal immigration is to have long, totally unsecured borders with desperately poor countries. Sounds like a plan! 😛

No, conservatives don’t like the Iran deal, but before you get bogged down in the debate note that they have been against pretty much every deal with hostile foreign countries regardless of the terms.

Study The Long Run Impact of Bombing Vietnam investigates whether areas in Vietnam that suffered “the most intense episode of bombing in human history” during the war are still poorer today. They find that no, areas heavily bombed by the US are at least as rich and maybe even richer than areas that escaped attack. They try to adjust for the possibility that the US predominantly bombed richer areas, but that doesn’t seem to be what caused the effect. Their theory is that maybe the Vietnamese government invested more heavily in more thoroughly destroyed areas. More evidence that compound interest is the least powerful force in the universe?

Luke Muehlhauser, working with GiveWell, has come to a preliminary conclusion that low-carb diets probably aren’t that helpful. Given that me, Luke, and Romeo Stevens have all said we’re not too impressed by low-carb, can this be declared Official Rationalist Consensus?

A lot of people on my Facebook have asked why Black Lives Matter protesters are disrupting Bernie Sanders but not Hillary Clinton. Answer is: they tried to disrupt Hillary, but she has security. I feel like this is an Important Metaphor For Something.

There’s a lot of heartbreak and emotion in this New York Times piece, but the part that really stands out for me is that Oliver Sacks and Robert Aumann are cousins. This sort of thing seems to happen way more often than chance, and I shouldn’t really be able to blame genetics either since cousins only share 12.5% of genes.

In 2000, the medical community increased their standards for large trials, requiring preregistration and data transparency. Now a review looks at the effects of the change. They find that prior to the changes, 57% of published results were positive; afterwards, only 8% were. Keep this in mind when you’re reading findings from fields that haven’t done this yet.

The FDA rejected flibanserin, a drug to increase female libido, as ineffective and unsafe. The pharmaceutical company involved got feminists to call the FDA sexist for rejecting a drug that might help women (NYT, Slate) and the FDA agreed to reconsider. But now asexuals are mobilizing against the drug, saying that it pathologizes asexuality. I look forward to a glorious future when all drug approval decisions are made through fights between competing identity groups.

Stuart Ritchie finds that we have reached Peak Social Priming. A new psychology paper suggests that there was an increase in divorce after the Sichuan earthquake because the shaking primed people’s ideas of instability and breakdown, then goes on to show the same effect in the lab. Even the name is bizarre: Relational Consequences of Experiencing Physical Instability. Despite the total lack of earthquakes in Michigan to prime me, I still feel like this finding is on shaky ground.

The most important Twitter hashtag of our lifetimes: #AddLasersToPaleoArt.

I’d like to hear more people’s opinion on this: Jayman links me to a post of his where he argues against the third law of behavior genetics (most traits are 50-50 genetic/environmental), saying they are often more like 75% genetic, 25% environmental. He argues that the 50-50 formulation ignores measurement error, which shows up as “environmental” on twin studies. As support for his hypothesis, he shows that the Big Five Personality Traits, usually considered about 30-40% genetic on studies where personality is measured by self-report, shoot up to 85% or so genetic in studies where personality is an average of self-report and other-report. Very curious what commenters have to make of this.

Brainwashing children can sometimes persist long-term, as long as you’ve got the whole society working on it. A new study finds that Germans who grew up in the 1930s are much more likely to hold anti-Semitic views even today than Germans who are older or younger, suggesting that Nazi anti-Semitic indocrination could be effective and lasting. Contradictory more optimistic interpretation; in no generation were more than like 10% of Germans anti-Semitic, so the indoctrination couldn’t have worked that well.

The Catholic blogosphere is talking about how fetal microchimeralism justifies the Assumption of the Virgin Mary or something.

A new meta-analysis finds that the paleo diet is beneficial in metabolic syndrome and helps with blood pressure, lipids, waist circumference, etc. Seems to have outperformed “guideline-based control diets”, although I can’t get the full-text and so can’t be sure exactly what these were – and one of the easiest ways to get a positive nutrition study is to use a crappy control diet. But if that pans out, all the people talking about how the paleo diet has no evidence will have egg in their face (YES I JUST USED AN EGG PUN AND A PAN PUN IN A SENTENCE ABOUT THE PALEO DIET). And here’s an interview with the authors

A subreddit of words that are hard to translate. “I will zalatwie this” means “it will be done but don’t ask how.”

Study discovers dramatic cross-cultural differences in babies’ sitting abilities; African infants seem to be able to sit much earlier and much longer than Western ones. Possible reasonable explanation: we coddle our babies and keep supporting them when they could perfectly well learn to sit on their own if we let them.

A while back I made an extended joke comparing gravitational weight and moral weight. Well, surprise, surprise, somebody did a social priming study showing that they were in fact related. Now the inevitable negative replication is in.

Archaic Disease of The Week: Eel Thing

Since we’ve been discussing coming up with numbers to estimate AI risk lately, try Global Priority Project’s AI Safety Tool. It asks you for your probabilities of a couple of related things, then estimates the chance that adding an extra researcher into AI risk will prevent an existential catastrophe.

Reason article on how a chain of New York charter schools catering to poor minority students manages to vastly outperform public schools, including the ones in ritzy majority-white areas. Wikipedia appears to confirm. My usual suspicion in these cases is that it’s selection bias; the “poor minorities” thing sort of throws a spanner in that, but here is a blogger suggesting they use attrition rather than selection per se, and here is someone else arguing against that blogger. And here is a charter school opponent saying this chain is mean and violates our liberal values, which I am totally prepared to believe.

The latest in this blog’s continuing coverage of weird Amazon erotica which totally really exists: I Don’t Care if My Best Friend’s Mom is a Sasquatch, She’s Hot and I’m Taking a Shower With Her

Cognitive behavioral therapy can cut criminal offending in half – this study should be read beside Chris Blattman’s work showing similar effects in Africa. I am usually skeptical of large effects from social interventions, but after thinking about it, CBT is at least more credible than poster campaigns or something – it’s the sort of thing that in theory can genuinely have a long-term effect on people’s thought processes. If this is even slightly true then of course we should teach CBT in elementary schools. Maybe those New York charter schools will go for it.

I should probably link to this study “showing” “that” a “low-fat” “diet” is “better” than a “low-carb” “diet”, but lest anyone get too excited it really doesn’t show that at all. It shows that in a metabolic ward where everyone’s food is carefully dispensed by researchers and monitored for compliance, people lose a tiny amount more weight on low-fat than on low-carb over six days. This sweeps under the rug all of the real world issues of dieting like “sometimes diets are hard to stick to” or “sometimes diets last longer than six days” – in their defense, the researchers freely admit this and say the experiment was just to figure out how human metabolism reacts to different things and we shouldn’t worry too much about it on the broader scale. Some additional criticisms regarding ketosis, etc on the Reddit thread

Some countries have problems with annexing neighboring lands that later agitate for independence. Switzerland has a problem with neighboring lands agitating to join them even though it really doesn’t want any more territory.