When ideological narratives and data collide

Driving between one meeting and another the other day I heard an interview of Janet Reitman on NPR. She apparently had a new article coming out in the New York Times Magazine (U.S. Law Enforcement Failed to See the Threat of White Nationalism. Now They Don’t Know How to Stop It. by Janet Reitman).

I was intrigued because this has been one of the gross mismatches between mainstream media narrative and the news that I actually read - that there is a dramatic rise in right wing extremism and white supremacism. It is a claim plagued by definitional issues. Just who is far-right, who is a white supremacist, etc.? What constitutes a hate crime? What constitutes domestic terrorism?

We hear the claim repeatedly but where is the evidence? I am a pretty avid reader of a very broad range of newspapers and news sources. I listen to NPR whenever I am in the car. My only significant blindspot is that I do not watch TV news programs. But even there, it almost all ends up getting reported in print anyway.

And despite all that news absorption and awareness of the claims of a rise in far-right extremism, I have great difficulty putting my hands on actual evidence of such a rise. The commentators making the claim rarely make specific claims of a particular person being far-right or such-and-such institution being far-right. Other than Richard Spencer and other than the massively common ad hominem attacks of X-is-Hitler type. Lots of claims that X is racist, but again primarily falling into the category of ad hominem rhetoric rather than a serious claim.

You can readily identify a handful of instances in the past twenty or thirty years such as Timothy McVeigh, Dylann Roof, or Stephen Paddock where almost everyone might agree that their acts constituted some mixture of hate crimes, terrorism, and/or ideology. And even those, particularly Paddock, we don't have a completely clear picture of motives. But it gets very murky very quickly beyond some of those more obvious instances.

Listening to the interview with Reitman, there were some red flags in her argument, easily identifiable in the spoken word which are often easily overlooked in the printed version. Her claims switched dramatically between far-right, white-supremacist, and other derogatory terms rhetorically intended to incite revulsion. But these terms lack meaning. Is Rand Paul, being a libertarian and small government advocate, far-right? I would say clearly not and yet there are plenty who make that claim.

Likewise, is anyone who advocates for better or more tightly managed immigration inherently a white supremacist. Some, surely yes. But most, clearly no. And yet in so much reporting immigration control is held as a synonym for white supremacy.

Another red flag in the interview was the sources Reitman was using to support her claim that there is a sharp rise in far right white supremacy groups. The sources are primarily far left leaning groups such ADL and SPLC. When asked by the interviewer whether the government could provide documentation to support the claim that there is a rise in far-right extremism, Reitman responded with a long discussion that boiled down to - No, there is no empirical evidence.

To my data-oriented ears, another tell was that most of her claims were based on relative percentages rather than absolute numbers. A fifty percent increase in deaths sounds concerning until you discover that that represents an increase from 2 deaths to three deaths. Every death is a tragedy, but in a nation of 330 million a rise of 2 to 3 in any category of measurement is not inherently alarming. It is statistical noise on a year-to-year basis.

And that was the other empirical blind spot to the interview. Reitman provided no contextual reference for her argument. We are running, in very ball bark numbers, at 60,000 opioid deaths a year, 20,000 suicides and 15,000 murders. How many deaths are attributable to far-right actions, however those might be defined?

The interview wrapped up and I was left with the impression that this was an ideologically driven argument, not an empirical argument. But we all get wrapped up in our own cocoons. I parked in my mind to go to the ADL report on which she was drawing for evidence to support her argument as well as the FBI hate-crimes data. If we are seeing a real rise in far-right white supremacy actions, then we ought to see that reflected in hate-crime data.

I got around to that task today. The ADL data is ADL Report: White Supremacist Murders More Than Doubled in 2017
. Yikes! But this article is based on the underlying report and data in Murder and Extremism in the United States in 2017. In both these articles and reports as well as the NYT Magazine article, there is a curious and repeated argument that white supremacy murders far-outweigh Islamic terrorism in the US. Seems like an irrelevant issue in terms of whether white supremacy is on the rise.

And, again, a lot of those claims depend on the definitions and analytic methodologies used. Islamic terrorism incidents tend to be infrequent but consequential when they happen. Run of the mill extremism tends to be more frequent but less lethal.

But it is kind of moot as the underlying data turns out to be outrageously bad. There is no clear methodology, though they do have a methodology section. Basically, whatever the ADL thinks is a domestic terrorism act gets counted but they acknowledge that it is essentially a random selection. Their first methodology paragraph raises concerns.

The Anti-Defamation League’s Center on Extremism has compiled a list of over 1,000 known murders/killings by perpetrators associated with domestic extremist movements of all types since 1970—essentially the post-Civil Rights era. These are primarily murders committed by American extremists on U.S. soil, though a few cases involving American extremists murdering other Americans abroad (such as at Jonestown) are also included.

Per their data there are 1,000 or so deaths in roughly fifty years, or 20 deaths a year. That sounds markedly low. I am not arguing that the true number is huge but 20 per year sounds surprisingly low.

The looseness of definitions gets more explicit. They seem to be driven by stereotypes rather than hard definitions.

The main limitation of cross-movement comparisons is that extremist connections to killings are easier to determine for some movements than for others. For example, white supremacists, who frequently sport many racist and white supremacist tattoos, or who may be documented as white supremacists by gang investigators or corrections officials, are often more easily identifiable. In contrast, it may be more difficult for police or media to identify, say, anti-government extremist associations that a suspect might have.

So really, their database of domestic terrorism is primarily a database of crimes committed by people with prison gang tattoos.

They then get more explicit about what are perhaps the deepest flaws. They are not measuring deaths from domestic terrorism, they are measuring murders committed by individuals whom they have almost arbitrarily designated as white-supremacist. If this sounds like splitting a hair, consider one of their reports.

Ricky Dubose, a member of the Ghostface Gangsters white supremacist prison gang, and another inmate, Donnie Russell Rowe, reportedly killed two corrections officers while trying to escape from a prison bus. They were later recaptured.

Two violent criminals escaping from prison, kill two corrections officers. Everyone is white. This is deemed an act of white supremacist domestic terrorism by the ADL because the two escaping prisoners had prison gang tattoos.

Are the two white supremacists because they have prison gang tattoos? Quite possibly but that seems entirely irrelevant here. The murders occurred because of their efforts to escape prison, not as an act of domestic terrorism.

The ADL only has 19 instances of domestic terrorism in their database for 2017 resulting in 34 deaths and as you go through the 19 instances, it is clear that the great majority are not domestic terrorism. A son who posts on white nationalist websites gets into a domestic dispute with his father and stabs him to death. An Aryan Circle member kills another white man because the other fellow is dating his ex-girlfriend. These are all sad tragedies enacted by dangerous and troubled individuals but they are not acts of white supremacist terrorism. The ADL acknowledges that it is almost arbitrary how they are classifying things:

Usually such categorizations are straightforward, but occasionally incidents emerge that are much harder to characterize. Perhaps no better example exists than Jeremy Christian, who is accused of stabbing to death two people and severely injuring a third as they tried to defend two teenaged girls—one Muslim and the other African-American—in a confrontation reportedly initiated by Christian in Portland, Oregon in May. Based on comments made by Christian, descriptions of him by people familiar with him, as well as his social media postings, Christian emerges as a volatile, angry man who appears to have fueled his rage with ideas from a variety of sources. Christian frequently expressed hatred of people on the left, and showed up at right-wing events in the Portland area, but had been a supporter of Democratic presidential candidate Bernie Sanders—seemingly because he thought Sanders would smash the establishment. Most of Christian’s influences seem to have been right-wing in nature and include some from the white supremacist movement and others from anti-government extremists such as sovereign citizens and the militia movement. It is clear that Christian belongs somewhere on the extreme right but it’s difficult to categorize him precisely. For purposes of this report, his murders have been categorized as white supremacist in nature, but others could look at the available evidence and possibly come to a different conclusion.

A Bernie supporter is clearly a member of the white supremacist extreme-right? OK. Keep in mind that the two men Christian killed as well as the man he wounded were all white. That does not disprove that Christian might have been a white supremacist but it certainly raises questions. Where possible I went to independent reporting in the news or Wikipedia to get more details on the circumstances of the crime. In the case of Christian, he had a long record of behavioral issues, criminal convictions and a disturbing inclination to make provocative statements in public. From the other reporting, it seems unclear at all that white-supremacy was a motivation.

Then you have the peculiar instance of Arthurs which illustrates the challenge of concluding that affiliation with some Nazi group is defining of a crime as domestic white supremacist terrorism.

White supremacist Devon Arthurs allegedly shot to death two of his roommates for making fun of his recent conversion to Islam. All three, and a fourth roommate, were members of Atomwaffen, a neo-Nazi group.

A Nazi Muslim kills two fellow group members (also white) for making fun of his conversion. White supremacist?

It goes on and on in this vein. Arbitrarily selected cases, almost randomly classified based on stereotypes and prejudices rather than hard evidence. The ADL is padding out its numbers by including crimes committed by anyone they choose to deem as white-supremacist rather than documenting actual crimes committed as a consequence of their white supremacist beliefs.

Did Reitman not look at the underlying data? Did she just go with the headlines? It must be the case that she was a sloppy reporter. Otherwise, she must be simply lying rather than reporting.

The NYT thumbnail tilts the arrow towards lying:

Janet Reitman is a contributing writer for the magazine who is working on a book about the rise of the far right in post-9/11 America. She is also a contributing editor for Rolling Stone.

I can't imagine that she is writing a book without having looked at the data. The fact that she is affiliated with Rolling Stone and its track record for advocacy journalism even in the face of obvious falsehoods in its reporting is perhaps suggestive.

When you omit the cases that involved documented mental illness, deaths from domestic disputes, deaths arising in the commission of robbery or other crime, you are left with only 6 possible instances of domestic terrorism in 2017. Three in the name of or associated with Islam with 13 deaths, one by a black man explicitly intending to kill white people with two victims, one white man explicitly killing a black man in order to stop interracial dating, and one white man with two Hispanic victims (I could find no reporting on motive). For the Critical Theory identitarians, that is 3 Islamic domestic terrorist attacks with 13 victims, 1 Black racist with two victims and 2 white supremacists with three victims.

When you start dealing with actual numbers, Reitman's claim that white supremacism is rising and that it accounts for most domestic terrorism attacks goes out the window.

But we have already seen that the ADL numbers are a grab bag of prejudices, stereotypes and assumptions with no rigor to them. Bad data prevents strong conclusions. All we can conclude is that their own data does not support their own argument.

How about the FBI UCR Hate Crime data? Now we know that there are all sorts of issues with this data as well - It is a compilation of crime statistics from police departments all over the country and there are going to be data consistency issues. We also know that there are expenses associated with making a hate crime conviction (If you are a District Attorney, why take the risk and increase the cost by adding hate crime if you are confident of obtaining a conviction on the underlying crime already?) Increased cost and risk by adding hate crime will inherently reduce the number of claimed hate crimes reported. Etc. The saving grace of the FBI UCR data is that, regardless of the collecting department issues, there are a set of consistent definitions, there is a set of consistent processes, and there is a longitudinal record going back more than twenty years. A further complication (but a positive one) is that we are currently experiencing a secular long term reduction in all crime, including hate crimes, of pretty dramatic proportions over the past twenty-five years.

But perhaps the FBI UCR data data supports Reitman's contention that there is a rising wave of white nationalism violence. There is no category of white supremacist hate crime but it would not be unreasonable to assume that Murder and Rape based on race might reflect white supremacist violence. These are the crimes which tend to be most rigorously investigated and prosecuted, substantially improving our visibility into motives and circumstances.

What are the numbers for anti-black, anti-white or anti-any other minority? In 2016, the most recent data, racist crimes against whites were nearly 60% of all murders and rapes based on race (five of the murders out of seven and five of the rapes out of ten). African-Americans are the victims of about 23% of violent race crimes. And keep in mind that non-whites are responsible for a significant portion of the racial hate crimes among other racial categories.

So based on violent race crimes, it appears that white supremacists are committing hardly any violent hate crimes.

When you get to slightly less serious crimes such as aggravated assault, simple assault, and intimidation, the picture begins to align slightly more to the mainstream media narrative. While whites are the primary victim group among most violent crimes, African-Americans are the primary victim group (52%) with lesser crimes.

What about trends? How do these numbers compare to, say, 2012? Same headline story. A total of five race hatred murders and rapes in 2012, 40% against whites. Interestingly, there are no race hatred murders or rapes against African Americans in 2012. There were three rapes against Asian Americans.

The other thing we have to acknowledge is that all these numbers are so blessedly low (a total of all types of hate crime across all categories of hate of 7,321 in 2016 in a nation of 330 million), that a consequence is that there is a lot of noise in the trend lines. Very high one year, very low for several others.

Are there other sources of information which might give us reliable empirical data? Perhaps, but none spring readily to mind.

Reitman indirectly acknowledges that the Federal government is not focusing on rising white supremacism because the federal government is not seeing rising white supremacism. As opposed to the left leaning mainstream media who see it everywhere but likewise cannot actually document it.

White supremacists and other far-right extremists have killed far more people since Sept. 11, 2001, than any other category of domestic extremist.

This claim by Reitman is an ideological belief unsupported by any empirical evidence that I can see. Not only have white supremacists not killed far more people than any other category of domestic extremist, but from the best data available, including those cited by Reitman, they are killing hardly any people at all.

Now none of this is to deny that terrible crimes are committed for all sorts of reasons. All I am arguing is 1) that these are infinitesimal proportion of all crimes, and 2) that there is no ready evidence to support the belief in rising violent white supremacism.

And I guess I am making a third argument - That we need to tackle crime as it exists rather than trying to tackle crime based on ideological beliefs. The overwhelming number of murders are committed in domestic disputes and street crime. Gang murders account for approximately 10-15% of all murders (1,500 - 2,000 murders a year). Against these sorts of numbers, why would the FBI focus on a category (race-based hate crimes which only account for 10 or 20 violent crimes in a year.

From this perspective, Reitman's headline title becomes something of a tragic joke "U.S. Law Enforcement Failed to See the Threat of White Nationalism." Of course they failed to see that threat because, based on the empirical evidence, there is no material threat. And of course "They Don’t Know How to Stop It" because there is hardly anything to stop. It is random and without a pattern. Diverting resources from the well known categories of street crime and gang violence which take thousands of lives in order to focus on an ideological fantasy crime is borderline evil. And incomprehensible given the hard left fascination with critical race theory and the fact that some 75% of the victims of street crime and gang violence are poor African-Americans and Hispanics.

There is one interesting thread in here, and one which I keep harping on. A large percentage of people the ADL claims to have committed acts of domestic terrorism are found to be mentally incompetent to stand trial or have a long history of mental illness. They are crazy people whose crimes can be made to look like domestic terrorism but whose root condition is mental illness. As is the case with a huge percentage of the 20,000 thousand suicides in a year. And as is the case to some degree (substance abuse) with the 60,000 opioid deaths in a year.

Instead of diverting law enforcement to chase chimerical white supremacist crimes which are negligible in absolute terms, we should, I think, be focusing on how to better align mental health diagnostics, mental health treatment and police enforcement. Time and again tragedies occur where many or most of the danger signs were well known long in advance and yet ignored or lost in the system.

From this view, tackling real tragedies based on real conditions, Reitman's ideological convictions look evil.

UPDATE: The Mythical Rise in White Supremacist Violence by David Catron makes some of the same points as I do above with soma additional insights thrown in.