Both Kareem Carr and his critics are wrong.
it feels like Kareem Carr has been a PhD student on Twitter forever lol
I just checked his LinkedIn and yup he is in his 8th year of his PhD
Very good. Thanks for eliminating the confusion. One should always examine the procedure used to process data.
The offending rate for violent crime is higher among black people than other ethnicities. There is no doubt about this, hasn't been for decades and it's true across all Western societies. In itself it is arguably no great big deal because few people - even black people - are violent criminals. So the BIG story here is that a culture that is terrified of acknowedging a manifest truth is a culture that does not value truth over comforting myth.
Even if we go with your methodology which results in a 9.2:1 ratio, does this take into account that Hispanics and Arabs (now MENA) have historically been counted as "White"? Steve Sailer worked up these numbers and came up with 12:1 B on W vs W on B, by breaking out the estimated portion of whites who should be classified as Hispanic or MENA.
And I'm still having a hard time believing that Asians commit violent crime against whites 1.8:1. I've never even seen an Asian person litter, let alone commit any violent crime, and I live in a major US city with a high crime rate.
"Suppose they... do not target each other"
This assumption is doing a lot of work here. In particular, it's likely that in at least a subset of cases people are targeting, or biased towards, particular racial groups, and to the extent that interracial attacks are racially targeted (e.g. "hate crimes") we should correct for the perpetrator race but not the victim race (since if someone intends to find a victim of a given race, the likelihood of success is not linearly proportionate to the victim population size, whereas more potential perps will lead to more attacks).
This is relevant for two reasons. First, we just don't actually know what percent of attacks are racially targeted which means the truth is likely somewhere between your non-targeted value (9.2x) and the original version implicitly assuming 100% racial targeting (42x). So it's useful to have both figures to bound the actual ratio (between 9.2 - 42x), and not really fair to say that the per capita figures are "wrong". Another issue is that many progressives tend to assume that interracial crimes are targeted due to racism (esp white on black), so it's useful to show that on the assumption of 100% targeting, the perpetrator rates are much higher for black-on-white than white-on-black.
That said, I suspect most attacks are not strongly targeted by race, so probably your 9.2x figure is closer to reality, but it seems pretty hard to assess.
As someone who works with data on a daily basis, the most Important fundamental factor has to be the GIGO principle (garbage in, garbage out). All crime data and statistics are heavily biased if they’re even collected and variance in quality and focus is wildly different between counties and cities (here in the UK) let alone between countries. Anyone drawing conclusions on such soiled data is questionable and that would have to include the author.
Thanks for clarifying all that. I’ve been wondering about how to calculate these numbers correctly for years.
A beautifully clear explanation.
You state that Asians commit crime at a lower rate than whites, well certainly Northeast Asians commit crime at a lower rate than whites. Bur what about all those whose ancestors are from the continent of Asia, are they not included in these crime statistics? Ethnicities such as Syrian and Pakistani are Asian. Do the crime statistics separate out different kinds of Asians, or combine them? Pacific Islanders might even be included in the label "Asian".
Strange that people struggle with this. Should be intuitive for anyone with the slightest grasp of math or statistics.
Should a variable for class/SES of victim be included in the data? If blacks (and whites) disproportionately target wealthier individuals, and whites are wealthier than blacks, then that would skew things a bit.