Weak Analysis, No Useful Answers

Review of Charles Murray’s Facing Reality

In Facing Reality, Charles Murray argues that America’s most urgent problems stem from a widespread failure to confront two fundamental facts about race: (1) racial minorities like Blacks and Latinos score lower on cognitive tests (on average) than Whites, and (2) these same minorities commit more violent crimes (on average) than Whites. Ignoring these facts, says Murray, has led to “false accusations of systemic racism” (p. x). “special treatment” for minorities (p. 4), portrayal of Whites as “racist and hateful” people (p. x), and growing social unrest (pp. 119-20). The only hope, says Murray, is “facing reality” about race differences.

There are problems with this book, and they begin with Murray’s organizing conception. His main claim—that racial differences in test scores and crime are largely “ignored”—is a hard one to sustain. On the day I wrote this, I Googled the terms “intelligence” and “racial differences” (in quotes) together and got 349,000 hits. For comparison, combining the word “intelligence” with the phrase “social class differences” produced 251,000 hits. Substituting similar phrases for race like “ethnic differences” and similar comparison phrases like “socioeconomic differences” produced similar results. Likewise for comparisons regarding racial and other differences in crime (e.g., 594,000 vs. 288,000). Since Murray aims much of his criticism at academics, I repeated this experiment with the PsychINFO academic database, which, again, produced the same pattern (far more hits for racial than other kinds of differences, for both cognition and for crime). Wikipedia has articles on “Race and Intelligence” and “Race and Crime in the United States,” but no corresponding articles for other demographic groups; and an Amazon search produced numerous books with titles linking race with intelligence and with crime, and far fewer linking other demographic groups (and in the case of crime, the titles with other demographic groups often included race as well). Thus, racial differences in these areas can hardly be said to be “ignored” (p. ix).

Most of this book consists of statistics to support Murray’s claims about these differences, and the bulk of the statistics concern the results of IQ tests. There are also statistics on crime, but these are fewer and somewhat spotty. In what follows, I will point out some problems with Murray’s statistical analyses, and then then evaluate some of his conclusions. The sources I cite can be found on the “References” page of this website.

A reviewer of Murray’s earlier work once observed that, like some naive psychometricians, Murray exhibits an “intoxication” with the IQ test as a device with almost mystical powers for reducing intellectual life to a single number (Gardner 2001, para. 13). In reality, as a measure of academic and cognitive ability, the IQ concept is limited in some significant and problematic respects (see, for example, Ceci, 1996; Richardson, 2017). Murray’s statistical analyses in Facing Reality are weakened by his glossing over these limitations and problems. Global IQ (the single number given by most IQ tests) is a statistical construction based on correlations among a variety of tests, It is not always clear what this construction signifies about the specific abilities of the individuals who take the test, or in which contexts. One cannot assume, for example, that IQ is the most accurate measure of the skills required for any particular task or job. Murray repeatedly treats IQ as the best predictor of job performance, but a number of studies have found that IQ is often a worse predictor of such performance than tests of specific job-related abilities, and sometimes has no predictive value at all (Thomas et al., 1996;Grobelny, 2018). Similarly, another reviewer of Murray’s earlier work discovered that Murray had excluded from his data on job earnings the strongest single predictor (a numerical operations test) because it correlated only weakly with global IQ scores (Heckman, 1995, para. 32). Nevertheless, Murray focuses on global IQ scores of White and non-White workers, emphasizing the superiority of the former over the latter across several job categories. But he fails to note that the gaps among professionals, his main focus, are more a function of above average skills of Whites than below average skills of minorities (this is obscured partly by Murray’s use of z-scores). This pattern is consistent with analyses of “White privilege” and with the evidence on causation (see below). Murray gives no recognition to this or to the ways in which global IQ scores conflate social and cultural factors with innate ability (Nisbett, et al., 2012; Suzuki & Aronson, 2005; see also Kaufman et al., 1997). And he says little or nothing about factors that may bias some of the measurements he cites as manifestations of IQ (e.g. licensing exam scores, performance evaluations, and records of complaints and suspensions). Some of these biasing factors include, but are not limited to, exam preparation opportunities and resources, preexisting social and cultural capital, discriminatory judgments, and the fact that minority employees often work in settings like inner-city schools and hospitals that face inordinate challenges coupled with insufficient resources.

In other parts of the book, Murray presents statistics demonstrating that violent crimes in the U.S. occur primarily among minority populations, especially in highly segregated portions of large cities. He focuses on the most violent crimes because statistics on these offenses are generally held to be more accurate than those regarding lesser, nonviolent crimes (pp. 49, 136). But like his decision to look only at IQ scores after the age of 12 (p. 26), his decision to focus mainly on the most violent crimes skews his analysis primarily toward social problems that are fully entrenched, while ignoring the long chains of events that precede them. Thus, while the majority of prisoners in state and federal facilities are violent felons, the number of nonviolent felons passing through the prison system is proportionately higher because of their greater turnover (Rutherford, 2015). This, in effect, makes prisons training grounds for younger, initially nonviolent, prisoners whose options are increasingly narrowed because of their criminal involvement. These younger prisoners are largely individuals of color who have moved through the “school-to-prison pipeline,” where minority youths are disproportionately targeted for punishment and suspension in schools (Skiba & Losen, 2015-2016), and are prosecuted for nonviolent crimes at higher rates than are their White counterparts, even when criminal record and severity of offense are controlled for (Kansal & Mauer, 2005; Bishop et al., 2020). For example, while Blacks and Whites use and sell drugs at about the same rate, Blacks are 12 times more likely to be imprisoned for a drug offense than Whites (Mauer & Cole, 2011). Murray gives no attention to dynamics like these, arguing instead for the even-handedness of the criminal justice system, sometimes ignoring obvious likely confounding factors. For example, he cites a study reporting higher arrest rates for Whites than for Blacks as evidence for lack of bias (pp. 52, 140-41), but seems never to have considered the possibility that these rates might reflect resource and case load disparities between police departments in White versus Black communities.

Murray claims to be a truth-teller, but he is consistently evasive about the reasons for the racial differences he reports. If the differences among the races are as important as Murray holds them to be, one would expect to see something from him about the causes of these differences. But Murray is emphatic: the causes are “irrelevant” (p. 47). Irrelevant? Even Murray’s most sympathetic readers have been puzzled by his indirectness on this point. For example, conservative writer Glenn Loury recently pressed this issue in an interview with Murray, and Murray would say only that the reasons for the differences were “intensely controversial” (The Glenn Show, 2021, 18:50). Since Murray himself did much to promote this controversy in 1994 by suggesting that racial differences in cognition were at least partly genetic (Herrnstein & Murray, 1994, p. 311), his response here seems strikingly disingenuous. Murray was similarly evasive when questioned in another recent interview by Tucker Carlson. Murray responded that “certain kinds of outcomes exist that are not explained by racism, let alone systemic racism. They are explained by differences that exist, for whatever reasons, between different ethnic groups” (MediaMatters, 2021, paras. 15-16). But this is tautological nonsense. The differences between ethnic groups are the very facts that need explaining, not the explanation themselves. And in refusing to supply an explanation, Murray has no grounds for his claim that racism and systemic factors can be ruled out.

Why would Murray be so evasive about a position for which he is already well-known? Evidently, because he wants to have it both ways: Murray’s inferences about genetic causation elevated him to notoriety in 1994. Over the next few years, his data were re-analyzed by a number of researchers, who repeatedly found that his analyses were wanting, that the most important social and environmental causes of racial differences had not been considered, and that the more thoroughly these social and environmental factors were included, the further Murray’s genetic explanation shrank toward implausibility (e.g., Dickens et al.,1994; Phillips et al., 1998; Fischer et al., 1996; Devlin et al., 1997). Similarly, the U.S. crime statistics that Murray reports are part of a larger worldwide pattern in which violent crimes regularly increase in contexts of income inequality and poverty; when these factors are held constant, race disappears as a correlate and a possible explanation (Neapolitan, 2008; Pridemore, 2011). Murray does not confront the implications of these findings, nor has he ever refuted—or even effectively disputed—the criticisms of his previous work. And why should he? The public remembers his genetic claims and is largely unaware of the subsequent research debunking them. So he’s better off avoiding this topic.

If the causes of racial differences are social, as indicated by the bulk of the research, then interventions at the social level would seem to be a reasonable response. But this is a reality that Murray himself does not want to face, because it clashes with his libertarian sensibilities (pp. 1-3, 8). In a particularly revealing section of the book (pp. 93-94), Murray acknowledges that many of the racial differences he has identified are potentially addressable, but that multiple issues would need to be confronted to do so (e.g., neighborhood dynamics, de facto segregation, family structures, employment opportunities, and political interference). He focuses instead, however, on the “dismal record” of the limited governmental and other programs that have been attempted instead. Two things need to be said about this.

First, even these more limited programs have been impeded by determined resistance throughout their history. For example, in the mid-1960s, after 350 years of ignoring educational deprivation and inequality, the federal government made a commitment to raise the academic performance of minority and other children through elementary and preschool “compensatory education” programs like Head Start. Within five years of this decision, Arthur Jensen published an article in the Harvard Educational Review declaring, in the first sentence, that these efforts had “failed” (Jensen, 1969, p 2) and adding that genetics must be the cause (p. 82). Jensen had concluded this based on a mass of misinterpreted data, dubious studies (several apparently nonexistent), and a plethora of fallacious, untenable, and untestable assumptions. All of this was thoroughly critiqued, over the next few years, by a number of authors, including Cronbach (1969), Lewontin (1970), Coleman (1972), Moran (1973), Layzer (1976), and, especially, in a deep and painstakingly detailed analysis, by Kamin (1974). Nonetheless, Jensen immediately became a widely cited “authority” by those who opposed compensatory education for ideological reasons, including Charles Murray, whose 1994 book cited Jensen more than any other author. Thus, Jensen’s article reanimated a longstanding symbiosis between academic authors making bogus arguments about race and conservative politicians working tirelessly, especially after the election of 1980, to limit social programs designed to reduce inequality. Likewise, affirmative action, which initially had widespread public support, was targeted with intense opposition from resistant politicians, and was systematically curtailed by increasingly conservative court decisions after the late 1970s.

Second, despite this political resistance, these programs were more successful than Murray and other critics have acknowledged. For example, many early negative evaluations of compensatory education were seriously flawed methodologically (see Bauer, 2019), failed to recognize real gains (see Deming, 2009), and were overly focused on the augmentation of global IQ and other relatively abstract measures, which tend to stall when children are re-immersed into ordinary learning environments (see Protzko, 2015). When a wider range of outcome variables has been examined (e.g. educational attainment and income), compensatory education has been found to be significantly successful (e.g., De Haan & Leuven, 2020). Murray himself notes that the Black-White gap in test scores shrank throughout the 1970s and early 1980s, and then leveled off during the more conservative era that followed (pp. 33-35)—a trend which he seems to regard as a mystery of nature. Similarly, affirmative action produced a number of positive results in the earlier years. For example, the number of Black executives more than doubled for some positions (Glass Ceiling Commission, 1995), and Black students demonstrated levels of post-graduate success that were close to those of White students (Bowen and Bok, 1998). And even Murray implicitly acknowledges that relatively few White students lost opportunities due to affirmative action (pp. 69-71; 80-83—see also Menand, 2020, para. 18).

None of this dampens Murray’s conviction that a widespread “unwillingness” to face the “intractability” of racial differences (pp. ix, 7) is the driving force behind a national crisis that is tearing at the social fabric. While he acknowledges that polarization and extremism come from both sides of the political spectrum, this acknowledgement is perfunctory (p. 8), and nearly all of Murray’s concern is focused on the progressive theories of social constructionism and systemic racism, and on the “new ideology” of identity politics (pp. 10, 4-5). In doing so, Murray aligns himself with naïve leftists in interpreting these theories as claiming that individuals are “inescapably defined” by the groups into which they are born (p. 5) and that Whites are therefore “irredeemably racist” (p. 115). But these are serious distortions. The whole point of social constructionism is that personalities are not fixed, that group identities are constantly reconstructed by social interactions, and that identities are potentially open to change (for example, see Myers, 2021). Likewise, the concept of “systemic racism” refers not to racist individuals, but to institutionally embedded laws, standards, and procedures that unjustly reinforce and perpetuate existing inequalities. On p. 109, Murray himself acknowledges the power of such forces to create racial injustices, but he seems unable or unwilling to draw the obvious conclusion that they therefore constitute a form of structural—as opposed to personal or individual—racism. And far from being a “new ideology,” the practice of identity politics has existed throughout American history (e.g. in immigrant communities and political machines). In fact, White identity politics was probably the earliest variant, espoused by authors who praised the beauty and superiority of Caucasians (e.g. Blumenbach, 1795), justified slavery and segregation (e.g., Cartwright, 1851; Garrett, 1961), and, in recent decades, promoted White working class populism—sometimes overtly linked with white supremacy, sometimes not (see Jardina, 2019).

These dynamics do raise important issues that need to be considered in any thoughtful attempt to confront America’s race problems: How much responsibility can any one generation take for the injustices perpetrated by previous generations? How much initiative can those who inherited those injustices be fairly expected to take in attempting to counter them? When are societal interventions appropriate in this process, and how can the costs of such interventions be distributed in a just and reasonable manner? These are difficult questions to which there are no simple answers. Unfortunately, Murray’s analysis—which consists of nothing more than an inventory of racial differences and a willful refusal to examine the causes of these differences—is not a resource for approaching honest answers to any of them.

8/1/21