Researchers just found substantial evidence for something conservative scholars and criminal lawyers have long suspected.
Researchers just found substantial evidence for something conservative scholars and criminal lawyers have long suspected.
‘S ystemic racism,” at least in the criminal-justice system, has been comprehensively measured. This, following my recent review of Cory Clark and Lee Jussim’s excellent analysis of academic censorship, to which I myself contributed the odd word, will be my second “academic findings” article in a row for National Review. Get excited!
This week’s topic is at least as important as last week’s. In a recent quantitative meta-analysis, Stetson University’s Christopher Ferguson and Sven Smith found substantial evidence for something conservative scholars and criminal lawyers have long suspected. Simply put, based on models that adjust for social class as a variable, there appears to exist little or no “systemic” racial bias in the U.S. criminal-justice system.
Overall, the pair’s meta-analytic review of some 51 studies — which, combined, included around 120 distinct effect tables — concluded that “neither class nor race biases for criminal adjudications for either violent or property crimes could be reliably detected.” For drug crimes such as crack sales, some evidence of bias was found, but “effect sizes were very small” — often on the order of a few percentage points — and may be primarily artifacts of the weaknesses of past studies.
Ferguson and Smith note that the “better quality studies” that were included in this complex analysis were “less likely to produce results supportive of [racial] disparities.” Studies with “citation bias” toward one side of the debate about whether racism exists in the modern criminal-justice system — generally, a bias in a liberal direction — were far more likely to find racism than were unbiased fair-test studies, and they “consistently produced higher effect sizes.” Only the innermost hearts of a few academics truly know why that is — although the findings of my last piece might provide a clue. Overall, the Stetson pair concludes: “Narratives of ‘systemic racism’ as relates to the criminal justice system do not appear to be a constructive framework from which to understand this nuanced issue.”
On some level, and this is not to minimize a top piece of work, these findings seem gobsmackingly obvious in 2023. As Ferguson and Smith put it, “the civil rights movement was successful in creating equality before the law.” Since then, they add, “evidence from psychological studies demonstrates major reductions in both explicit and implicit racial prejudice.”
More bluntly put, about 40 percent of Americans are now members of minorities, and 90-odd percent of whites do not test as racist when this is measured by any real metric such as opposition to interracial marriage. Hell, roughly 30 percent of cops are black and Hispanic — and so are more than 10 percent of lawyers (before we throw in Asian and Jewish Americans). Setting CRT mumbo jumbo to the side, where it belongs, there seems to be no reason to expect black or Irish-American big-city judges to treat (say) Puerto Rican teen goons differently from Italian or Bosnian ones. And, overall, they do not.
Serious analyses, of the kind that Ferguson and Smith conducted, frequently come to similar conclusions across a variety of sectors. Back in 1990, the economist June O’Neill took a look at the black–white income gap — then about 18 percent annually, and always attributed to racism. She concluded that fairly simple adjustments for variables such as age (the modal age of a black American is 27, versus 58 for a white Yank), region of residence, test scores, and years (rather than quality) of education shrink this gap to roughly 1 percent. Similarly, in policing, Harvard economist Roland Fryer’s multifactorial analysis, which included controls for the behavior of suspects and racial groups’ rate of contact with police, concluded that white suspects are actually 27 percent more likely than black suspects to be shot by cops . . . rebutting the entire Black Lives Matter narrative almost in passing.
This brings us to a key point. Life, much like good analysis, is multivariate. People who differ from one another in terms of a major factor such as race or religion are also very likely — indeed almost certain — to differ as well in terms of other characteristics, like those O’Neill studied. Racism occurs only when individuals are treated differently entirely and solely because of the single factor of their race.
It’s worth spelling this out. An analysis of mortgage-loan rates could be said to have uncovered racism only if large and statistically significant gaps in lending rates to racial groups remained after a statistical adjustment for black/Caucasian/Asian differences in age, region, income, wealth, credit score, and probably default rates. Similarly, any study of arrest rates (or crime rates) by race has failed to demonstrate the existence of bias — and therefore is pretty much useless for our purposes — if it has not made the obvious adjustments for age and social class.
By this standard, a rather amazing amount of past research — journalistic “studies” and some academic work — is pretty much useless. As Ferguson and Smith note, in a surprising number even of the better-known studies in their field of criminal justice, “control variables are comparatively lacking. . . . Many studies that examine race issues don’t control for class” (emphasis mine). “Most” studies control for “age and prior criminal record,” but not all.
Interestingly, very few papers on criminal sentencing control for what may be the most critical variable of all — whether your lawyer was a public defender or private counsel and how good he was. Criminal justice is hardly unique in this regard: Conservative éminence grise Thomas Sowell devoted an entire book to debunking scholars’ and reporters’ cross-disciplinary trend of simply attributing any disparity between groups — black vs. white maternal mortality, for example — to racism. But that particular dragon even the great man could not slay.
Why is the univariate problem that Ferguson and Smith and Sowell describe so well, and which I have been known to gripe about myself, so prevalent? There are many reasons, one suspects. Many writers, especially in the journalistic sector, are almost certainly well intentioned but unskilled with advanced statistical tools such as Stata. Others likely notice real racism in today’s online public squares, jump to the premature conclusion that this kind of tawdry stuff explains serious decisions that adults such as bankers make, and don’t test for much else.
But it is hard not to notice a potential third factor. An entire sector of the economy — within which fall most left-aligned think tanks and lobbying groups and more than a few academic departments (“postcolonial studies,” anyone?) — is dependent on the continued existence of old-school bigotry in the United States. Old warhorses such as Jesse Jackson’s Rainbow Push and Al Sharpton’s National Action Network never really went anywhere, and they have been joined at the trough by more than 100 BLM groups led by Patrisse Cullors’s Black Lives Matter Global Network Foundation. Even sexual-rights groups now host seven-figure fundraisers centered on professional-looking white papers (I responded to the Human Rights Campaign’s claims of “trans genocide” here). This lucrative grift is quite hard to defend in the near absence of real and serious bias — and it might often be tempting for smart activists or even ideologically inclined academics to go out and create some.
The rest of us, however, shouldn’t go along with the radicals (or Thuh Science™) if their ideas and methods are simply bad and unscientific. To quote Ferguson and Smith one more time, when it comes to understanding at least the major and nuanced issue they study, “narratives of ‘systemic racism’ . . . do not appear to be a constructive framework.”