A Tabooed Inquiry

This topic is absolutely unpalatable for the general public. Most people are too simple to think about this complex issue without diverging into moralistic screeching and a priori rejection of what amounts to heresy in polite society. Nevertheless, burying our heads in the sand doesn’t change reality. One can either calmly consider the implications of this tabooed topic, or be blindsided by them in the future.



Race and Crime:

Is the Relationship Purely Social?



Violent crime rates vary greatly between racial groups. There is an abundance of sociological explanations, but there is a tendency to baselessly assume that no biological factors could possibly be important to fully understanding the phenomenon. This paper argues, primarily through evidence from genetic polymorphisms, that there are at least some biological factors that influence crime rates, which do not result from hidden social factors. Following the biological evidence, a brief philosophical discussion is presented, arguing that proper scientific inquiry into average racial differences does not necessitate bigotry.


Key Words

Polymorphism, Race, Aggression, IQ, Individualism, Murder



A person with physical characteristic x is six times more likely to commit murder than a person without physical characteristic x. This statement alone would typically motivate a full scientific inquiry into the relationship between characteristic x and violence. However, when characteristic x turns out to be African ancestry (FBI), entire segments of science become off-limits in the inquiry. Many theorists engage in a priori rejection of biological explanations for behavior. Biology is an important and valid perspective for understanding human behavior. This paper argues that behavior which differs by race is not an exception; that neurocriminology need not avoid this topic, and that with proper interpretation, science of biological differences need not lead to bigotry.

A first step in any neurobiological study is to define the subject of study. In the case of race, even this step is controversial. The American Anthropological Association (AAA) has officially stated that “the concept of race is a social and cultural construction” that “simply cannot be tested or proven scientifically.” According to the AAA statement, “it is clear that human populations are not unambiguous, clearly demarcated, biologically distinct groups,” and therefore “the concept has no validity in the human species” (Miele 110). Similarly, Jefferson Fish, in his 2002 Race and Intelligence: Separating Science from Myth, cites various historical racial classification systems in order to argue that since race is not consistently demarcated, it is not a valid object of study (Fish 140).

The mere fact that many people do not fit into any traditional racial category does reveal some limitations of the concept of race. Where to draw the line between categories is fundamentally arbitrary. But this limitation is not unique to race. The electromagnetic color spectrum from red to violet can also be categorized in infinite arbitrarily ways. Does this imply that a scientist can not validly assay a chemical compound by the red or blue color of its flame? There are after all, many purple flames which demonstrate that the boundary between the red and blue categories is arbitrary. This is not to say that the color spectrum or chemical structure have any substantive analogy in race, except for the principle that blurry categorical boundaries do not preclude valid scientific examination.  

Though most academics dismiss the possibility of racial biological behavior factors, they do not ignore the disparity in crime, victimization, and incarceration. Most academics just explain these phenomena from a purely sociological perspective. Many sociologists catalog ways in which racial or ethnic groups have disparate outcomes in life, taking this to imply unfairness in society (Sampson and Lauritsen 265). But if one is concerned about differing outcomes such as crime, victimization, and incarceration rates between races; if he perceives injustice on the basis of disparate outcomes between racial groups, then he has already implicitly accepted the validity of racial categorization for his own analytic purposes. It would be inconsistent to say ‘category x has an unfair outcome in life’ and then to also say ‘there is no such category x.’

A reasonable response by a sociologist may be, ‘Race is not metaphysically real, but when outcomes are highly correlated with race, there must be some societal mechanism that causes the average disparity. Blurry boundaries do not prevent this.’ To this response, I would completely agree, with one modification. Only add ‘biosocial interaction’ to ‘societal mechanism,’ and now we have a shared rationale for studying racial differences. The sociologist understands that blurry boundaries do not preclude research in general. Otherwise he would never study the sociology of racial disparity. It requires special pleading to say that blurry boundaries preclude study of the biology of race.

With the concept of race justified as a subject of study, we can now delve into some specific findings which could theoretically contribute to the racial crime disparity. This paper compares two general perspectives:

  1. That racial differences in criminal behavior can be completely caused by purely sociological factors.
  2. That both social and biological factors must be considered to fully understand the disparity.

Since the first perspective denies all biological influences, but the second perspective does not deny any type of social, environmental, or societal influence, there is an asymmetry in how extraordinary each claim is. To disprove the first perspective, all that is needed is evidence of at least one small biological factor. All of the existing evidence of important social factors does not in any way contradict the second perspective. So the only relevant type of evidence for this paper is biological in nature (including neuroscience, genetics, and their effects on psychology and behavior).


Biological IQ Theory and Evidence

One possibility is that different racial groups have biologically influenced differing average intelligence, and that lower intelligence causes crime. This is perhaps the most taboo and controversial idea in social and behavioral sciences. For now, consider its veracity or falsity without reference to its moral palatability. Ethical interpretation will be discussed later.

According to Kevin Beaver (2011), an impressive body of research has revealed that individual-level IQ scores are negatively associated with criminal and delinquent involvement. This line of research has been extended to show that state-level IQ scores are associated with state-level crime rates. In his study, Beaver examines the potential association between county-level IQ and county-level crime rates. Analysis of data drawn from the National Longitudinal Study of Adolescent Health revealed statistically significant and negative associations between county-level IQ and the property crime rate, the burglary rate, the larceny rate, the motor vehicle theft rate, the violent crime rate, the robbery rate, and the aggravated assault rate. Additional analyses revealed that these associations were not confounded by a measure of concentrated disadvantage which is meant to capture the effects of race, poverty, and other social disadvantages of the county (Beaver 2011).

Since IQ scores have been shown to be associated with crime at many scales of analysis, it stands to reason that a difference in IQ between races would cause a difference in crime between races. According to the American Psychological Association (Neisser et al 1996), there is a long-standing 15 point or 1 standard deviation difference between the intelligence test scores of Black Americans and White Americans. The difference is largest on those tests, verbal or nonverbal, that best represented the general intelligence factor (g). Controlled studies of the way the tests were formulated and administered had shown that this did not contribute substantially to the difference. Attempts to devise tests that would minimize disadvantages of this kind had been unsuccessful. The scores predicted future achievement equally well for blacks and whites (Neisser et al 93).

Of course, environment affects IQ, so the difference can still be seen as a social effect. But the biosocial perspective does not deny any legitimate social influence. The question is whether or not biology also affects IQ. There are two levels on which intelligence could be said to have biological causes. First, and less controversially, the physical processes in the brain which underlie thinking can be shown to vary between people in correlation with varying intelligence. And secondly, intelligence can be shown to be genetically heritable.

Until relatively recently, scientists have had difficulty pointing out specific neural correlates of intelligence. In 2015, Finn et al published an fMRI study in Nature in which they showed that connectivity profiles predict levels of fluid intelligence (a factor of general intelligence). fMRI studies typically collapse data from many subjects, but brain organization varies between individuals. Finn et al used data from the Human Connectome Project to demonstrate that functional connectivity profiles act as a ‘fingerprint’ that can accurately identify subjects from a large group. Identification was successful across scan sessions and between task and rest conditions, indicating that an individual’s connectivity profile is intrinsic, and can be used to distinguish that individual from others, regardless of how the brain is engaged during imaging. Characteristic connectivity patterns were distributed throughout the brain, but the frontoparietal network was the most distinctive between individuals. And those same networks that were most discriminating of individuals were also most predictive of intelligence. (Finn et al 1668).

Finn’s fMRI study provides specific evidence of what most physical scientists already assumed; that thinking arises from neural substrates, and that physical differences in these substrates result in qualitatively and quantitatively different characteristics of thought. These physical differences are to some degree formed by the social environment, but studies show that to some degree, IQ is also genetically heritable (Plomin et al 212). Based on twin studies, sibling studies, and adoption studies, about half of the variation in IQ comes from variation in inherited genes (Bouchard et al 227).

However, these studies do not identify the specific bits of genetic code that cause variation in IQ. Without that information, along with information about the prevalence of these genetic variations in different racial groups, one can always theorize ways in which environments might still be the sole cause of racial disparity. Since environment contributes 50% to within-group variation, it seems hypothetically possible that it might account for 100% of between-group variation. This possibility (however unlikely) would only be ruled out if specific gene variations were shown to cause higher or lower intelligence, and if those specific genetic variations were shown to exist in different proportions between racial groups. This can actually be done right now.


Evidence from Genetic Polymorphisms

Nina Volf et al (2015) have found an association between IQ and two distinct versions of a gene that codes for the serotonin transporter. Short and long alleles of a polymorphism known as 5HTTLPR in the SLC6A4 gene cause respectively low and high expression of the serotonin transporter coding gene. This polymorphism was studied in 208 Caucasian subjects. They were students, scientists, technicians and administrative staff of Novosibirsk State Universities and State Research Institutes of Russian Academy of Sciences. Higher IQ scores were observed in the subjects with two copies of the short 5HTTLPR allele (S/S genotype) compared with those who had one short and one long (S/L genotype) (p<0.002), or two long alleles (L/L genotype) (p<0.003). Based on this result, Volf et al hypothesized that there would be an increased frequency of the short (S) allele in the highly educated university group compared with a control group of Caucasian subjects from Dovolnoe village, Russia (n=83). The frequencies of the S allele and the S/S genotype in the academic sample were significantly higher than in the control group (p=0.011). The results indicate a significant association between IQ scores and the 5HTTLPR polymorphism.

In order to test for differences across races in the 5HTTLPR polymorphism, I used data from the National Center for Biotechnology Information’s Single Nucleotide Polymorphism Database, which includes data on single nucleotide polymorphism allele frequency in different populations. A single nucleotide polymorphism (SNP) is a spot in the genetic code at which it is common for different people to have a different letter. Usually, an SNP has two possible letters (alleles). For example, at a given SNP, some will have the A allele, and some will have the G allele. 5HTTLPR is not an SNP, but it includes SNPs, including rs2129785, and rs11867581. Based on a study of 2823 unrelated individuals by Anna Vinkhuyzen et al (2001), having both the A allele of rs2129785 and the A allele of rs11867581 results in a 91% probability of having the short allele of 5HTTLPR (Vinkhuyzen et al 1075).

So to impute the frequency of the 5HTTLPR S/S genotype in each race, I took the A allele frequency of rs2129785 in a given race, multiplied by the A allele frequency of rs11867581 in that race, multiplied by 91%, giving the probability of having one 5HTTLPR S allele, which I squared, giving the probability of having two S alleles, which is the S/S genotype. I performed this computation for three groups; ‘White,’ ‘Asian,’ and ‘Black.’ In the ‘White’ category, I included populations of Utah residents with Northern and Western European ancestry and Toscans in Italy. In the ‘Asian’ category, I included populations of Chinese in Metropolitan Denver, Colorado, Han Chinese in Beijing, and Japanese in Tokyo. In the ‘Black’ category, I included populations of African ancestry in Southwest USA, Yoruba in Nigeria, Luhya in Kenya, and Maasai in Kenya.

The calculations show that for the 5HTTLPR S/S genotype, Asians have a frequency of around 50%, whites have a frequency of around 20%, and blacks have a frequency of around 0%. This means that the gene that Volf associated with high IQ in Russian subjects is most common in the race with the highest average IQ and lowest violent crime; it is moderately common in the race with middling average IQ and middling violent crime; and it is least common in the race with lowest average IQ and highest violent crime. The gene does not change over the course of a lifetime, so it can not be a result of differing environment.

That the 5HTTLPR polymorphism would affect criminality through the serotonin transporter protein is also theoretically consistent with other findings, not dependent on IQ. The S/S genotype down-regulates expression of the serotonin transporter gene, which should lead to reduced transporter numbers, and thus reduced reuptake of serotonin. Thus the S/S genotype likely has an agonist effect on serotonin receptor neurons. This is the same action intended in the administration of selective serotonin reuptake inhibitors (SSRIs) in the treatment of depression (Preskorn et al 241). Agonist serotonin action through the serotonin receptor has also been shown to reduce aggression (Caspi 114). So the 5HTTLPR polymorphism may affect criminality through more channels than IQ alone.

The 5HTTLPR polymorphism is a biological factor that affects IQ, and is not caused by social factors or environment. This causal factor is present in different frequencies between races, and thus has a differential effect between races. Therefore this is an example of a way in which racial disparity can not be fully explained by social factors, no matter how many are added to the catalog.


Evidence of Anger and Aggression Caused by an SNP

The avoidance of racial differences by academics makes it difficult to find high quality research on the topic. But the conspicuous underdevelopment of the field makes it easy to do more original back of the envelope research. Rather than using IQ as an intermediary between genetics and violent crime, I will now consider a genetic factor for anger and aggression.

Ina Giegling et al (2006) investigated the association of four serotonin receptor gene SNPs with anger, aggression, and suicide related behavior in 566 German subjects. Anger and aggression related traits were assessed by using the State Trait Anger Expression Inventory and the Questionnaire for Measuring Factors of Aggression. Subjects with the CC genotype for the SNP rs6311 reported more anger (p= 0.004) and aggression related behavior (p= 0.011) (Giegling et al 78).

The known mechanisms of genes imply that the causality in a non-random association between a gene and a behavior must be in the direction of gene to behavior. Unlike brain function, which can be shaped throughout life by environments, a person’s genotype remains constant throughout life. So the behavior in question can not cause the genotype. Additionally, SNP alleles are passed from parents to children in a straightforward replication mechanism with high fidelity. Therefore, there is no hidden variable that could could cause both the genotype and the aggressive behavior. This means that to the statistical level of the current data, we know that the CC genotype for the SNP rs6311 causes elevated anger and aggression.

Is this biological risk factor for aggression equally distributed across racial groups? Using data from the National Center for Biotechnology Information’s Single Nucleotide Polymorphism database, I tested for differences in relative rs6311 genotype frequency between the three race categories previously described (Black, Asian, and White).

The ‘Black’ category has the highest proportion of rs6311 CC genotypes, followed by the ‘White’ category, followed by the ‘Asian’ category. t-tests showed that the differences in CC genotype frequency between ‘White’ and ‘Black,’ as well as the difference between ‘Asian’ and ‘Black’ are statistically significant, with p-values < 0.001. The difference in CC genotype frequency between ‘Asian’ and ‘White’ is not significant.


Figure 1: Three racial categories have different proportions of genotypes in an SNP that has been associated with anger and aggressive behavior. The differences between ‘White’ and ‘Black,’ as well as the difference between ‘Asian’ and ‘Black’ are statistically significant, with p-values < 0.001. The difference between ‘Asian’ and ‘White’ is not significant.
Though the association between the rs6311 SNP and aggressive behavior requires further examination, especially in how it interacts with other factors, current evidence points to a genetic influence on aggression, which differs between races, and which by definition can not be the result of the social environment. There are many social factors working in concert with this genetic factor, possibly mitigating or exacerbating its ultimate effect, but the genetic factor exists independently of any social factor.

The findings on rs6311 and 5HTTLPR likely only account for a very small fraction of variation in crime. There are certainly more important social factors, and likely many unexplored biological factors. However, these finding prove a few principles:

  1. Behavior is shaped by the physical brain.
  2. The brain is shaped by genetics and environment.
  3. Differing behavior can be caused by differing genes.
  4. Racially differing behavior is not a special exception to this general schema.
  5. If one wishes to fully explain differing behavior, he must be open to causal pathways from both differing environments and differing genetics.

There is currently an abundance of unutilized genetic data residing in databases, which could be used by researchers to answer important questions. But there seems to be a conspicuous avoidance of this subject matter in academia, possibly for good reason.


Ethical Interpretation

Scientific endeavors that intentionally search for differences between races may evoke thoughts of historical examples of scientific justification for mistreatment. Before the civil war, Josiah Clark Nott, an American physician and surgeon, used skull classification to advocate in favor of slavery (Dewbury, 2007 144). Pseudoscience was also used to justify apartheid and eugenics. To some, this seems to imply that science of racial differences inevitably leads to mistreatment. But perhaps it would be more accurate to say that mistreatment follows from pseudoscience, and that pseudoscience fills the space vacated by conscientious and legitimate scientists. The truth seeking process alone does not intrinsically entail racism, bigotry, or mistreatment. Those depend on the philosophical interpretations of the findings.

What would a well-developed science that does not avoid racial questions mean for people’s lives? The findings would be in the form of differences in group averages. So the impact of the findings depends on how people digest information about groups and individuals. Some people tend to treat group averages as if they apply to every member of the group. I will call this the collectivist interpretation. Others prefer to consider each person as an individual who may or may not align with any average trend in his group. I will call this the individualistic interpretation.

Collectivism can have either malevolent or benevolent intentions. An archetypal white supremacist is a malevolent collectivist. He’s all too eager to look for ways in which his group can be construed to be better than another group. He then makes an additional logical jump to conclude that the perceived average differences imply that he himself is necessarily superior to any given individual in another group.

In contrast, the benevolent collectivist seeks to catalog disparities in average life outcomes between groups, constructing a taxonomy of privilege and oppression. In an empathetic effort to aid the ‘oppressed,’ the benevolent collectivist makes an additional logical jump to conclude that the group averages say something about every particular individual in the ‘privileged’ or ‘oppressed’ group.

Individualists take a different approach to group differences. For an easily digestible example, consider the genetics of height and race. There are many genetic factors that cause white men to be taller than asian men. However, knowing that the average white man is taller than the average asian man does not tell us anything about the height of any particular individual asian or white man. The only reasonable way to judge any person’s height or buy him the correct pants is by his own height alone, not by the racial average. In the same way, knowledge of an average difference between races in genetic risk of violence would not inform us of how any particular person should be judged or treated. Each individual should still be treated on his own merits, no matter how much genetic evidence of average differences accumulates. To do otherwise would be a mathematical error. An average can be computed from a set of data points, but no particular data point can be computed from an average. An individual is a single data point in the population, and none of the average measures of any of the groups to which he may be assigned necessarily say anything about him.

Another example, the gender disparity in crime, demonstrates that average differences can be accepted and interpreted in a mostly individualistic way. Men are nine times more likely than women to commit murder. Society is generally aware of this, and scientists are less resistant to biological explanations as reviewed by Archer (2004). This has not lead to treating every man as if he has anything to do with the elevated male murder rate, nor treating every woman as a harmless angel. Mostly, men and women are judged to be murderous or not based on their own individual murder count, as should blacks, whites, and any other group under an individualistic interpretation.

Given a mathematically correct individualistic interpretation, discovering new scientific information about group averages should not lead to bigotry. Of course, the problem is that in the real world, people often tend to not understand the distinction between the individual and the group. There is a concern that since malevolent collectivists tend to make unjustified judgements based on collective group affiliation, any research on group differences will only add fuel to the fire of racism. This is certainly a possibility. One can find plenty of examples on the internet of white supremacists posting cherry picked and misunderstood science to justify the way they feel about people of other races (Brent 1998). It is possible that more science would just be more for them to misunderstand.

However, the benevolent collectivists should also be glad to learn more about differential genetic propensities for crime. Those who wish to achieve equity in incarceration and victimization rates across racial groups need to understand the true nature of their task in order to succeed. Stringently believing in the impossibility of physical differences leading to behavioral differences will guarantee an incomplete plan of action to reach their goal. Preemptively dismissing biology as a basis of behavior is profoundly anti-scientific, and will not lead to a realistic solution. If the benevolent collectivists truly desire racial equity, they should seek to understand all types of causes of inequity.

As shown by evidence from genetic polymorphisms, those causes that drive the relationship between race and crime are not purely social, though perhaps they may be primarily social. Just like every other variation in human behavior, racial variation in violent crime is a complex system with causal inputs from genetics and from the social environment, which interplay in the structure and processes of the physical brain. To continue enforcing the taboo against scientific inquiry into the matter is to choose to not understand the world. And any policy, with benevolent intent or not, which is based on misunderstanding is unlikely to produce the desired effect.



Archer, John. “Sex Differences in Aggression in Real-World Settings: A Meta-Analytic Review.” Review of General Psychology, Vol 8(4), Dec 2004, 291-322.

Bouchard, Thomas J.; Lykken, David T.; McGue, Matthew; Segal, Nancy L.; Tellegen, Auke (1990). “Sources of Human Psychological Differences: The Minnesota Study of Twins Reared Apart”. Science. 250 (4978): 223–8.

Brent, James “Links to Hate Groups.” San Jose State University webpage. (1998) http://www.sjsu.edu/faculty/Brent/190/hatelinks.html

Caspi N, Modai I, Barak P, Waisbourd A, Zbarsky H, Hirschmann S, Ritsner M.; Modai; Barak; Waisbourd; Zbarsky; Hirschmann; Ritsner (Mar 2001). “Pindolol augmentation in aggressive schizophrenic patients: a double-blind crossover randomized study”. Int Clin Psychopharmacol. 16 (2): 111–115.

Dewbury, Adam, 2007. “The American School and Scientific Racism in Early American Anthropology”, in Darnell, Regna; Gleach, Frederic W., Histories of Anthropology Annual, 3, p. 142

Finn ES, Xilin Shen, Dustin Scheinost, Monica D Rosenberg, Jessica Huang, Marvin M Chun, Xenophon Papademetris & R Todd Constable “Functional connectome fingerprinting: identifying individuals using patterns of brain” Nature Neuroscience 18, 1664–1671 (2015)

FBI. UCR Expanded Data Table 6: Murder https://ucr.fbi.gov/crime-in-the-u.s/2013/crime-in-the-u.s.-2013/offenses-known-to-law-enforcement/expanded-homicide/expanded_homicide_data_table_6_murder_race_and_sex_of_vicitm_by_race_and_sex_of_offender_2013.xls

Fish, Jefferson M. Race and Intelligence: Separating Science from Myth. Lawrence Erlbaum Associates, Publishers. 2002

Giegling I, Hartmann A, Hans-Jürgen Möller, Dan Rujescu. “Anger- and aggression-related traits are associated with polymorphisms in the 5-HT-2A gene.” Journal of Affective Disorders. Volume 96, Issues 1–2, November 2006, Pages 75–81.

Miele, Frank. Intelligence, Race, and Genetics. Westview Press, 2002.

National Center for Biotechnology Information. “Reference SNP (refSNP) Cluster Report: rs2129785.” Single Nucleotide Polymorphism databasehttps://www.ncbi.nlm.nih.gov/projects/SNP/snp_ref.cgi?rs=2129785

National Center for Biotechnology Information. “Reference SNP (refSNP) Cluster Report: rs11867581.” Single Nucleotide Polymorphism databasehttps://www.ncbi.nlm.nih.gov/projects/SNP/snp_ref.cgi?rs=11867581

National Center for Biotechnology Information. “Reference SNP (refSNP) Cluster Report: rs11126630.” Single Nucleotide Polymorphism databasehttps://www.ncbi.nlm.nih.gov/projects/SNP/snp_ref.cgi?rs=6311

Neisser, Ulric; Boodoo, Gwyneth; Bouchard, Thomas J., Jr.; Boykin, A. Wade; Brody, Nathan; Ceci, Stephen J.; Halpern, Diane F.; Loehlin, John C.; et al. (1996). “Intelligence: Knowns and unknowns”. American Psychologist. 51 (2): 77–101.

Plomin, R.; Pedersen, N. L.; Lichtenstein, P.; McClearn, G. E. (1994). “Variability and stability in cognitive abilities are largely genetic later in life”. Behavior Genetics. 24 (3): 207–15

Preskorn SH, Ross R, Stanga CY (2004). “Selective Serotonin Reuptake Inhibitors”. In Sheldon H. Preskorn, Hohn P. Feighner, Christina Y. Stanga, Ruth Ross. Antidepressants: Past, Present and Future. Berlin: Springer. pp. 241–262

Sampson RJ, and Janet L. Lauritsen, “Racial and Ethnic Disparities in Crime and Criminal Justice in the United States,” Crime and Justice 21  pp. 311-374 (1997)

Vinkhuyzen, Anna A. E. et al. “Identification of Tag Haplotypes for 5HTTLPR for Different Genome-Wide SNP Platforms.” Molecular psychiatry 16.11 (2011): 1073–1075.

Volf N.V., Sinyakova, N.A., Osipova, L.P., Kulikov, A.V. and Belousova, L.V. (2015). Association between intelligence quotient and the 5HTTLPR polymorphism of human serotonin transporter coding gene. Annals of Neuroscience and Psychology, 2:6.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s