How frequent are Hate Crimes vs. Hoaxes?

https://www.wsj.com/articles/hate-cr...nk-11561503352

"Mr. Reilly eventually compiled a database of 346 hate-crime allegations and determined that less than a third were genuine. Turning his attention to the hoaxes, he put together a data set of more than 400 confirmed cases of fake allegations that were reported to authorities between 2010 and 2017. He allows that the exact number of false reports is probably unknowable, but what can be said “with absolute confidence is that the actual number of hate crime hoaxes is indisputably large,” he writes. “We are not speaking here of just a few bad apples.”

The author’s bigger concern, and rightly so, is the growing politicization of hate crimes, especially when they are directed at underrepresented groups and regardless of whether they in fact happened. The sad reality is that there is no shortage of individuals and entities with a vested interest in exaggerating racial tensions in the U.S.—from civil-rights organizations to corporate diversity officers to professors of race and gender studies."

And there - ya happy Loki, I even used the WSJ just for you.

Those numbers are meaningless without knowing how he chose his 346 cases. Needless to say, this is not from a peer-reviewed book.

https://en.m.wikipedia.org/wiki/Wilfred_Reilly

If I'm reading this correctly, he compiled a dataset of hoaxes and is using that to claim a large number of hate crimes are hoaxes...He's also claiming that because there were 400 hoaxes in the past 5 years and only 1,100 hate crimes are reported each year, 1/3 of reported hate crimes are hoaxes. That's some pretty awful math and statistics.

Faux research is a blind spot for all ideologues, left and right. Too desperate to latch on to empirical support for one of their claims, even when the claim has never been properly tested.

Quote:

---

Originally Posted by Loki

Those numbers are meaningless without knowing how he chose his 346 cases. Needless to say, this is not from a peer-reviewed book.

https://en.m.wikipedia.org/wiki/Wilfred_Reilly

If I'm reading this correctly, he compiled a dataset of hoaxes and is using that to claim a large number of hate crimes are hoaxes...He's also claiming that because there were 400 hoaxes in the past 5 years and only 1,100 hate crimes are reported each year, 1/3 of reported hate crimes are hoaxes. That's some pretty awful math and statistics.

Faux research is a blind spot for all ideologues, left and right. Too desperate to latch on to empirical support for one of their claims, even when the claim has never been properly tested.

---

The issue of course being where does the burden of proof lie? In the person who makes the claim of the hate crime or in the skeptic who thinks the person is making it up. Generally speaking usually the person first making a claim something happened should need to prove it. *Especially* when there are political, fame and monetary considerations.

Quote:

---

Originally Posted by Lewkowski

The issue of course being where does the burden of proof lie? In the person who makes the claim of the hate crime or in the skeptic who thinks the person is making it up. Generally speaking usually the person first making a claim something happened should need to prove it. *Especially* when there are political, fame and monetary considerations.

---

The claim in this case is that 1/3 of reported hate crimes are hoaxes. That claim is not substantiated.

Quote:

---

Originally Posted by Loki

The claim in this case is that 1/3 of reported hate crimes are hoaxes. That claim is not substantiated.

---

I'm clearly talking about *any* claim that a hate crime has occurred. Who has the burden of proof?

Quote:

---

Originally Posted by Lewkowski

I'm clearly talking about *any* claim that a hate crime has occurred. Who has the burden of proof?

---

We have courts for that. You're conflating the public opinion reaction with the legal reaction.

Quote:

---

Originally Posted by Loki

Those numbers are meaningless without knowing how he chose his 346 cases. Needless to say, this is not from a peer-reviewed book.

https://en.m.wikipedia.org/wiki/Wilfred_Reilly

If I'm reading this correctly, he compiled a dataset of hoaxes and is using that to claim a large number of hate crimes are hoaxes...He's also claiming that because there were 400 hoaxes in the past 5 years and only 1,100 hate crimes are reported each year, 1/3 of reported hate crimes are hoaxes. That's some pretty awful math and statistics.

Faux research is a blind spot for all ideologues, left and right. Too desperate to latch on to empirical support for one of their claims, even when the claim has never been properly tested.

---

It's ideologically driven, but what would be a better methodology in this instance?

Any design has to deal with most hate crimes not being reported and with the ones that make the news being the most sensationalist ones (and thus most likely to be hoaxes). There's also the matter of determining whether a report was fake. I doubt most are as clear cut as the Chicago actor.

This means you can't get your cases from news reports. Nor can you make claims about hate crimes in general; you can only talk about reported ones. I'd say you can start by picking a city and examining every single hate crime reported there in a given year (or look at a number of years, but have a random number generator choose the cases for you). Then you come up with as an objective set of criteria as possible for determining whether a report was a hoax. Then you train people who aren't ideologically committed to a certain conclusion to consistently apply those criteria. Lastly, you have 2-3 people examine each case to make sure there's agreement on the coding. If there's disagreement, you have more people look at it or create a separate variable that mentions your level of confidence in a particular coding decision.

Quote:

---

Originally Posted by Loki

We have courts for that. You're conflating the public opinion reaction with the legal reaction.

---

No I'm not - who has the burden of proof in public opinion? Example if a shitty actor says they were attacked by two white guys win MAGA hats, what should the default stance be?

Quote:

---

Originally Posted by Lewkowski

No I'm not - who has the burden of proof in public opinion? Example if a shitty actor says they were attacked by two white guys win MAGA hats, what should the default stance be?

---

There is no such thing as burden of proof in public opinion. You can't tell people what to believe. This isn't North Korea. :bored:

Quote:

---

Originally Posted by Lewkowski

No I'm not - who has the burden of proof in public opinion? Example if a shitty actor says they were attacked by two white guys win MAGA hats, what should the default stance be?

---

burden of proof in public opinion

burden of proof in public opinion

burden of proof in public opinion

burden of proof in public opinion

:bulb:



Lewk, you're a lovely little gem, you really are. :kiss:

Quote:

---

Originally Posted by Loki

There is no such thing as burden of proof in public opinion. You can't tell people what to believe. This isn't North Korea. :bored:

---

Continue to dodge the question if you must but it really is a simple one. When someone makes a public claim that X happened, where *should* the burden of proof lie?

Quote:

---

Originally Posted by Loki

Any design has to deal with most hate crimes not being reported and with the ones that make the news being the most sensationalist ones (and thus most likely to be hoaxes). There's also the matter of determining whether a report was fake. I doubt most are as clear cut as the Chicago actor.

This means you can't get your cases from news reports. Nor can you make claims about hate crimes in general; you can only talk about reported ones. I'd say you can start by picking a city and examining every single hate crime reported there in a given year (or look at a number of years, but have a random number generator choose the cases for you). Then you come up with as an objective set of criteria as possible for determining whether a report was a hoax. Then you train people who aren't ideologically committed to a certain conclusion to consistently apply those criteria. Lastly, you have 2-3 people examine each case to make sure there's agreement on the coding. If there's disagreement, you have more people look at it or create a separate variable that mentions your level of confidence in a particular coding decision.

---

I see. You're describing a good methodology, but I think you're demanding that research into one thing also measure another (far less measurable) issue that is related, but not necessary.

One can study and quantify hoax crime reports and compare to real crime reports; I don't think one must determine the number of unreported crimes out there if studying crime reporting. It's certainly a related issue, but the ratio between hoax/non-hoax reports is just as interesting as the ratio between reported/non-reported crimes.

I'm perfectly willing to believe the number of hoax hate-crime reports is very high, and the number of unreported hate-crimes is also very high.

You cannot draw any firm quantitative conclusions about the ratio of hoaxes to real hate crimes using the methodology employed in this book. Loki's proposed approach, applied to a single manageable region, would at least help ameliorate some of the concerns about cherry-picked data. Your defense implies that you believe very bad research on a topic is better than no research on that topic, but that is a questionable stance. Research isn't a one-way knowledge-ratchet; bad research decreases knowledge, and the harm it does to our understanding of the world is often disproportionately difficult to remedy.

I don't disagree with anything you're saying. But I don't agree with the idea that research into hoax crime reporting must by default also discover the number of real unreported crimes.

It doesn't have to deal with the issue of unreported crimes. But it must make it clear that the findings only apply to reported crimes.

The book in question doesn't even do that; it only looks at purported hate crimes that were reported ON, and not even all of those.

For a professor, you'd think he'd know what selection bias is.

Quote:

---

Originally Posted by Loki

For a professor, you'd think he'd know what selection bias is.

---

I'm certain he does. "Selection bias is what other people use to get results that are different from mine"

In a better college, he'd really be screwing his career by publishing something of this quality, even if it was in a non-peer-reviewed source. But he's not.