After reading the first few lines I was going to recommend RCA's old blog post. Then you linked to it.
When I was researching/debunking the lead crime hypothesis, I found an article on crime rates by IQ, in Sweden. Large cohort size, not much racial confounding:
After reading the first few lines I was going to recommend RCA's old blog post. Then you linked to it.
When I was researching/debunking the lead crime hypothesis, I found an article on crime rates by IQ, in Sweden. Large cohort size, not much racial confounding:
They found that dropping IQ by 15 points doubled the violent crime rate.
So, we do have a B/W gap of ~15 points in the US (for whatever reasons). But the crime ratios aren't 2X, they are more like 6-10X. It seems like even the people who believe in genetic differences need a better theory.
I like your theory. Though I suppose you also need to establish that the marriage gaps cause higher crime, not that both are secondary to another factor. And you might want to look back in time to see if crime ratios increased as marriage broke down. Marriage rates were much higher 60 years ago, but I think the crime ratios were similar.
I grew more doubtful as i researched it, but the thing that really clinched it for me was the 5th graph in section 3.
The increase in crime in the 60s was a period effect, not a cohort effect. Every age group became more violent at the same time, they weren't lead poisoned one after another.
Analyzing by cohory also helps to understand the peak of crime in the late 80s/early 90s. You can rule out the abortion/crime hypothesis the same way.
I should say it's hard to call either theory completely debunked -- lead or abortion could both have some subtle (10%?) effect on crime and it would be hard to prove or disprove that. But neither one is the kind of unifying theory of crime rates that it promises to be.
After reading the first few lines I was going to recommend RCA's old blog post. Then you linked to it.
When I was researching/debunking the lead crime hypothesis, I found an article on crime rates by IQ, in Sweden. Large cohort size, not much racial confounding:
https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0041783
They found that dropping IQ by 15 points doubled the violent crime rate.
So, we do have a B/W gap of ~15 points in the US (for whatever reasons). But the crime ratios aren't 2X, they are more like 6-10X. It seems like even the people who believe in genetic differences need a better theory.
I like your theory. Though I suppose you also need to establish that the marriage gaps cause higher crime, not that both are secondary to another factor. And you might want to look back in time to see if crime ratios increased as marriage broke down. Marriage rates were much higher 60 years ago, but I think the crime ratios were similar.
You're right that in the end this entire thing is an exercise in correlation, not causation. But I don't have a better theory on causation.
If you've got a link to lead-crime debunking I'd like to read it. I've always thought lead-crime was an interesting idea.
https://medium.com/@tgof137/debunking-the-lead-crime-hypothesis-949e6fc2b0dc
I grew more doubtful as i researched it, but the thing that really clinched it for me was the 5th graph in section 3.
The increase in crime in the 60s was a period effect, not a cohort effect. Every age group became more violent at the same time, they weren't lead poisoned one after another.
Analyzing by cohory also helps to understand the peak of crime in the late 80s/early 90s. You can rule out the abortion/crime hypothesis the same way.
I should say it's hard to call either theory completely debunked -- lead or abortion could both have some subtle (10%?) effect on crime and it would be hard to prove or disprove that. But neither one is the kind of unifying theory of crime rates that it promises to be.