Many responses to this misunderstand it. A lot of commenters, in their confusion or ignorance, assume that it must necessarily simply be due to abject poverty and lack of access to other food.

In reality, this is an old cultural tradition that persists from West Africa.

https://twitter.com/jeremykauffman/status/1833965281838006327

Geophagy – the scientific term of the eating of dirt, earth and mud – has been common in many places in Africa, for a variety of purported purposes.

John Hunter (1973) discusses this, and also how the practice diffused to the United States by the African diaspora.

Why poverty is not a major cause of violent crime: macroeconomic evidence

While it is well known that poverty tends to be correlated with violent crime, causally informative evidence usually tells a different story. One important source of evidence is the macroeconomic kind.

If economic factors causally affected violent crime, then, when macroeconomic conditions change within a country or region, you should expect violent crime rates to follow. Economic downfalls should lead to increases in violent crime, and vice versa.

One illustrative example is that the homicide rate actually fell in America during the Great Depression–the opposite of what the causal hypothesis would expect.
But it's not just the Great Depression. In the Handbook of Crime Correlates, Lee et al. review the macroeconomic evidence and find that economic downturns have no systematic effect on violent crime.

More recently, Ludwig & Schnepel (2024) also reviewed macro evidence and also concluded that there is no systematic effect on violent crime.
The COVID economic crisis is yet another example. Hoeboer et al. (2024) reviewed the evidence and found that globally there was no increase in violent crime. If anything, it was in the opposite direction.

The United States famously did experience a homicide rate increase starting in 2020. But a careful analysis of the timing of that homicide increase shows that this was the result of a homicide spike immediately following the death of George Floyd. It should not be considered an economic effect.

Overall, there is no clear indication that macroeconomic downturns result in an increase in violent crime. Often the evidence finds the opposite. This is yet another strong line of evidence against the idea that poverty has a major causal impact on violent crime.
There is also evidence from within-individual studies:

https://x.com/Scientific_Bird/status/1831406728878961131

The moment when the West changed forever, on its trajectory towards the modern world. This is at least what I have argued. This chart is based on my analysis of a database of notable people. But, if you think such a database is too biased to consider, fear not, for much other evidence points to this being an important transition period.

Immigration and crime in Sweden

Recently, I wrote a piece in which analyzed the topic of immigration and crime in Sweden. I have posted some of the findings individually, but I thought it would be useful to compile all the major findings in one place. So, thread below. Immigrants and second-generation immigrants have higher crime rates than people with Swedish background.

As shown in the image, this is true after adjusting for age and sex compositions, and it's true for all major categories of crime.

This chart is wrong.

It might be comparing gun homicides to gun deaths, or something to that effect. Whatever the case, the data is clearly nonsense.

That's unfortunate, because there is a legitimate point to be made.

Short thread.

https://twitter.com/ElijahSchaffer/status/1822843896705028258

Cursory research shows that the firearm homicide rate in West Europe is far lower than portrayed in that figure. A firearm homicide rate of ~0.1 per 100k is common. Even in Sweden, known for having serious issues with gun killings, has a rate of ~0.5 per 100k.

I have written a short piece about Sweden's unwillingness to publish information about immigration.

The Swedish government publish very little information regarding immigration and crime or financial impacts. When compared to, say, Denmark, the contrast is glaring. Beyond that, certain data is considered "sensitive" in Sweden. This includes, e.g., nation of birth data.

Note, it's not sensitive in the usual sense of personally identifiable information, but instead in the sense that it's *politically* sensitive.

My newest post is the first in a new series where, starting with Denmark, I will investigate the topic of immigration and crime in a given country.

Are immigrants overrepresented in crime? If so, which immigrants? And why? In Denmark, immigrants are overrepresented in criminal convictions and in imprisonments.

The figure below illustrates crime rates relative to Danish origin for various offense types.

In a new piece, I discuss more myths about Norwegian recidivism and rehabilitation. Many say that prior to '90s rehabilitation reforms, Norway had a recidivism rate of 60 to 70%, and reforms greatly reduced the recidivism rate. Is this true? These sorts of claims are made on the Wikipedia page for "Incarceration in Norway", and a large number of articles repeat them.

In a new post, I show that the Nordic countries, despite common belief, have not rehabilitated their way to remarkably low recidivism rates.

When properly compared, their recidivism rates are, perhaps surprisingly, not much better than the United States.

Short thread. It is common to see comparisons between the United States and the Nordic countries. For example, US NEWS reports that Norway has a recidivism rate of only 20%, compared to 76.6% in the United States.

In a new post I argue that poverty has little causal effect on violent crime. I interrogate why poverty and violence are correlated and conclude that it is primarily due to selection and reverse causality.

Short thread below. It is well known that a correlation between poverty and violent crime exists. Here I have illustrated it for poverty- and homicide rates across counties in the United States. Unsurprising, but importantly it does not show that poverty causes violence.

In a new post, I investigate whether black-white homicide disparities in the United States can be accounted for economic disparities.

In short, I find that large systematic homicide disparities remain, even when economically similar people are compared. To address the question, I combine economic data from the American Community Survey, and homicide victimization data from the CDC; both at the geographical level of counties. Importantly, both being separated by race, so I can properly compare the groups.

What does research say about the risk of divorce after a spouse becomes ill, and particularly about the sex differences in divorce risk?

Recently I noted that one of the papers with claims of sex differences was retracted after an error was discovered.

https://twitter.com/Scientific_Bird/status/1743746334551658931

In the comments, critics argued that the original conclusion of the retracted paper was correct, pointing to a different 2009 paper in support.

But as this was the sole paper forwarded, and the sample was small (515 patients), I wanted to check whether the results replicate.

A new meta-analysis finds that undergraduate IQ has declined to a mere 102 on avg.

As they note, a great decline is a necessary consequence of the increasing share who enroll into college.

Imagine the extreme: if everyone goes to college, their IQs must necessarily be average. Imagine that 50% of the population go to college: the average IQ can at most be that of the top 50% (112).

In practice, the average IQ will be substantially lower as there is not perfect sorting into college (some high IQ individuals don't attend college, and some lower IQ do).

When did the rise of the West happen, and why?

I document the West's ascent by quantitatively tracking the birth of notable people of science. I then put this in context with other indicators of development. Blog post here and a short thread below.

inquisitivebird.substack.com/p/the-rise-of-… I use a database of notable people, combined with historical population estimates, to track per-capita rates of notable people of science throughout history. I focus mainly on 1000-1500, the period prior to colonialism.

nature.com/articles/s4159…

In this thread I will collect predictors of violent perpetration that have been established in large datasets, preferably with sibling comparisons.

People are very welcome to suggest any good studies I haven't included. (1/n) (2/n) A sibling comparison is one that associates a sibling difference in some characteristic to a sibling difference in violent perpetration. The advantage here is that it controls for things shared by siblings, making the association a more plausible causal candidate.

The cursed OkCupid graphs. What do they show? I might as well address it once and for all, so I don't feel the need to talk about them ever again.

The common criticism is that they are misleading, because they don't include message rates. This criticism makes sense. (1/n) (2/n) Here are the figures with the message distributions.

The difficulty here, however, is that the attractiveness ratings aren't comparable between the sexes (e.g., there are far more rated highly attractive by men than vice versa), leading to difficulty of interpretation.

This is one of those claims that survives in public discourse purely because it has been repeated so many times.

As far as I can tell, there is no, and there never has been, any compelling quantitative evidence demonstrating this claim.

https://twitter.com/vgonlin/status/1489293302998487052

She cites Crenshaw and Tim Wise, none of whom provide any quantitative evidence for this claim. She also cites a few other articles, but these articles themselves mostly just refer back to Crenshaw and Tim Wise (who didn't evidence the claim).

https://twitter.com/vgonlin/status/1489408743762714625

@rasmansa I'm not really particularly invested in whether "race" exists biologically or not, but to the extent that they exist as genetic categories, it has nothing to do with the mere dichotomous existence of an allele in a population. That's a red herring. @rasmansa Instead of simply "does the gene exist in the population, yes/no?", one would have to look at the differences in the overall structure of allele *frequencies.*

Consider that two people each have a coin, but the one individual's coin is biased 55% in favor of head when flipped.

The Danish government has released their 2021 report of net financial contribution of immigrants.

With access to population registry data, Denmark has the ability to calculate more accurate estimates than most other countries.

I will translate a few of the figures.

[Thread] First, Figure 2.7 shows net financial contribution by age and region of origin (in currency of Danish kroner).

The countries are organized into a few regional categories: Western, MENAPT (Middle East, North Africa, Pakistan, Turkey), and other non-Western countries.

(2/n)

Sex differences in personality — a thread with studies of large samples (1/n)

In the commonly used "Big 5" model of personality, the largest differences are found in neuroticism and agreeableness (women higher on average).

Source (N > 120,000; Denmark):
doi.org/10.1016/j.jebo… (2/n) With a sample of over 200,000 across 53 nations, this study finds that the differences in neuroticism and agreeableness are quite consistent across a large set of nations.

Source:
link.springer.com/article/10.100…

Statistics consistently show that men are more likely to be victims of stranger violence than women are. Yet, women are purportedly more afraid of such situations.

What are we to make of this?

Thread (1/n) (2/n) I think the first thing to say is that:

Feeling unsafe ≠ Actual risk

Some individuals are reckless. Without worry, they do things that put themselves in great physical danger. On the other hand, some individuals live in extreme debilitating paranoia.