Thread by @cremieuxrecueil

This post got 50,000 likes and it never even pointed out the actual issue with the calculations, it just took issue with framing and it expressed that Kareem is too inept to find sources.

But what's new?

Kareem debunking thread below

Kareem says this is a "textbook example of how to lie with statistics."

It really isn't, but let's see what he bases this on.

The first thing he says -- his "main criticism" -- is that the data isn't provided. But for Kareem, this is completely meaningless.

We know this is meaningless, because even when all the data is presented, Kareem still doesn't do anything with it, understand it, open it, manipulate it, or anything.

He says "where's the data?" and when he gets it, he just blocks you.

Example:

https://x.com/cremieuxrecueil/status/1728550278381617338

Kareem doesn't even deal in simulating data to check if his EXTREMELY BAD statistical intuitions are correct.

Remember this? If he had simulated it, he would've seen that the exact opposite of his claim was true!

https://x.com/cremieuxrecueil/status/1651674075025010699

But in any case, here's the data for the present post:

This should've taken Kareem about five seconds to find, so there goes his "main criticism".national-conservative.com/interracial-ho…

Kareem's next criticism is that the post leaves "out relevant information like how many murders happen between people of the same race."

But this is *irrelevant* information when the topic is interracial homicides.

It's only relevant if you're changing the topic!

He says that the post "makes it seem as if race is the main causal factor in murders", but that's only true if you have a reading comprehension problem.

He then says "if you look at interracial murders in the context of all murders, they are almost always a tiny fraction."

This is just not true!

Something like a fifth of murders cannot be called a "tiny fraction" given the population sizes and segregation at play.

Now, in the real world, homicide is increasingly interracial, and the Black-on-White number is large:

https://x.com/cremieuxrecueil/status/1839911540767473754

But Kareem's claim also isn't *relevant.*

Just because you want to talk about all murders instead of interracial murders doesn't mean that people are dishonest for talking about something different than what you wish they were focusing on.

This is a common ploy he does.

His claim that @GadSaad's post assumes Blacks and Whites are equally likely to murder except for race is asinine.

No one said that!

There can be tons of differentiating factors at play, but the statistics remain, and the risk to Whites from Blacks remains higher than vice-versa

Kareem then makes the baseless claim that calling an interracial homicide an interracial homicide is an "essentially racist framing". This makes no sense, but, who cares?

His claim that focusing on interracial murders is to condone intraracial murders is also not supportable.

Two things should be obvious.

Kareem deals in non-sequiturs

Kareem doesn't have any real criticisms, he just doesn't like what @GadSaad is saying, and he wants to claim that looking at stats he doesn't approve of can't be neutral when what he's replying to shows it can be.

Now Kareem isn't smart, he's a fraud, he's been caught lying, implying untruths, and more.

But what's funniest here is that there is a real statistical error and he missed it.

The image @GadSaad posted says Blacks are 23x as likely to kill Whites as vice-versa, but it's wrong!

You get to the 23x (really 23.1x) number by multiplying the ratios of Whites killed by Blacks to Blacks kill by Whites by the population percentage ratio, when that's not required.

For population pairs, you just use the ratio of killed to killed.

That number is 458/84 = 5.45x.

There's also a data issue that Kareem didn't notice:

These numbers are very unlikely to be true!

Open up EZASHR, which spans 1980 to 2020, and the yearly White-on-Black numbers are usually 400-500, and the Black-on-White numbers are usually 1,300-1,500.

Alarm bells.

This dataset starts in 2024, so we can't use overlapping years with EZASHR to see if it's correct in other years, but if we could, my bet is that they wouldn't line up.

Homicide statistics don't tend to jump around enough to get to the source's numbers.

As a final point of order, the Black-on-White number at the source is 472, not 458, and the White-on-Black number is 86, not 84. This is minor and probably just means Gad's image was a bit out of date. Not a big deal at all.

My advice to Kareem is to focus less on making threads that illustrate he suffers from cognitive difficulties and to instead focus on earning his PhD that he's spent nearly 9 years on so far.