Our model of racial bias in arrests predicts, and our empirical results suggest, that the policing of blacks has been disproportionately aggressive for crimes associated with high levels of police discretion to make on-scene arrests. . . . The resulting increases in the proportion of Blacks in the criminal justice system can affect the perception of officers in subsequent cases, perpetuating the imbalances even in the absence of racial differences in underlying criminality.15
A similar pattern of racial bias is seen in the makeup of the prison population. According to a 2007 report issued by the Justice Policy Institute, despite similar patterns of drug use, African Americans are far more likely to be incarcerated for drug offenses than whites. As of 2003, twice as many African Americans as whites were incarcerated for drug offenses in state prisons in the United States. African Americans made up 13 percent of the total U.S. population but accounted for 53 percent of sentenced drug offenders in state prisons in 2003. In 2002 African Americans were admitted to prison for drug offenses at 10 times the rate of whites in 198 large-population counties in the United States.16
According to data from the Innocence Project, an organization dedicated to the use of DNA to exonerate falsely convicted persons, racism is also a significant factor in wrongful convictions. African Americans make up 29 percent of people in prison for rape and 64 percent of those who were found to have been wrongfully convicted of rape. This suggests that there are disproportionately more false convictions for rape of blacks than of whites. Also, most sexual assaults nationwide are committed by men of the same race as the female victim. Only 12 percent of sexual assaults are cross-racial. This fact, in conjunction with data that show that two-thirds of African American men exonerated by DNA evidence were wrongfully convicted of raping white women,17 reveals the subtext of racial bias in false conviction. The vestigial fears and prejudice held by whites of black-on-white rapes, whose roots go back to slavery, become expressed in our current society by trumped-up accusations, mistaken identification, and false convictions.
Racial Composition of CODIS
Because of the disproportionately high rates of arrest and incarceration of people of color,18 it can be inferred that the U.S. Combined DNA Index System (CODIS), which, according to the FBI, is not coded by race, is also disproportionately composed of racial minorities. If CODIS, by and large, contains the profiles of past and present incarcerated felons, and minorities are disproportionately represented in prison, then it follows that they will be disproportionately represented in CODIS. Moreover, if Tomic and Hakes’s results are corroborated, the mere appearance of innocent blacks in DNA data banks, even if they are never charged with a crime, will impose a penumbra of bias by law-enforcement officials toward those individuals. If DNA databases are superimposed on traditional racial profiling and other forms of bias toward African Americans, the combination could lead to greater degrees of racial disparity in criminal justice. Troy Duster observes:
Indeed, racial disparities penetrate the whole system and are suffused throughout it, all the way up to and through racial disparities in seeking the death penalty for the same crime. If the DNA database is primarily composed of those who have been touched by the criminal justice system, and that system has provided practices that routinely select more from one group, there will be an obvious skew or bias toward this group.19
In chapter 2 we showed that there has been a rapid expansion of DNA data banks in certain states where the criteria have evolved from violent offenders and felons to arrestees who have not been charged with or convicted of crimes. Let us assume that this policy continues beyond the dozen or so states that have lowered the threshold of inclusion in their forensic DNA data banks, and that “being arrested” becomes the standard within the criminal justice system for requiring a biological sample and a DNA profile. How will that affect the racial composition of CODIS?
To contextualize this question, we first examine the racial composition of those who are serving sentences in federal and state prisons. According to a report published by the Pew Charitable Trusts in 2008, 1 in every 15 black males aged 18 or over, compared with 1 in 106 white males of the same age group, is in prison or jail. The highest rate of incarceration is among young black men. One in 9 black men aged 20 to 34 is behind bars.20 Although African Americans make up 12.8 percent of the U.S. population, they constitute between 41 and 49 percent of the prison population.21 Sixty-two percent of all prisoners incarcerated in the United States are either African American or Latino, but those groups constitute only one-quarter of the nation’s entire population.
There is no published information on the racial and/or ethnic composition of individuals who have DNA profiles contained in CODIS. The FBI does publish annual crime reports broken down into racial categories, but states may upload only names and DNA profiles to CODIS without personal information such as a racial or ethnic identifier. Each state retains the personal information associated with the profiles placed in the national DNA forensic database, but as far as we can tell, that information is not aggregated.
It is reasonable to assume that the racial composition of CODIS mirrors the racial composition of the prison population. The racial composition of CODIS can also be estimated indirectly. Henry Greely and colleagues use a crude measure of felony convictions to estimate the number of African Americans in CODIS, since the vast majority of states currently retain and upload to CODIS DNA from all felons:
African-Americans constitute about thirteen percent of the U.S. population, or about thirty-eight million people. In an average year, over forty percent of people convicted of felonies in the United States are African-American. . . . We assume, based on the felony conviction statistics, that African-Americans make up at least forty percent of the CODIS Offender Index.22
If we assume that the racial composition of CODIS mirrors the current prison population, we arrive at a similar estimate, with African Americans constituting 41 to 49 percent of CODIS.
According to the FBI, the National DNA Index System (NDIS) held 8,201,707 “offender” DNA profiles as of April 2010.23 On the presumption of 40 percent African American entries, they would constitute 3,280,683 entries in CODIS. This means that approximately 8.6 percent of the entire African American population is currently in the database, compared with only 2 percent of the white population.
Suppose that we asked this question: how would the composition of the database change if all states collected DNA profiles of arrestees? If blacks in American society continue to be stopped, searched, arrested, and charged at a rate in excess of nonminorities, and if collecting DNA samples from arrestees becomes the norm, then it would seem that the racial disparity of blacks in CODIS would continue to grow. D. H. Kaye and Michael Smith make the case for strong racial skewing of U.S. DNA data banks as the system currently exists:
There can be no doubt that any database of DNA profiles will be dramatically skewed by race if the sampling and typing of DNA becomes a routine consequence of criminal conviction. Without seismic changes in Americans’ behavior or in the criminal justice system, nearly 30% of black males, but less than 5% of white males will be imprisoned on a felony conviction at some point in their lives. . . . A black American is five times more likely to be in jail than is a white.24
The authors estimate that the profiles of about 90 percent of urban black males would be found in DNA data banks if all states required DNA samples from arrestees. At the same time, they contend that racial disparities would be diminished by expanding the databases to include arrestees because many more whites would be brought into the databases: “Racial skewing of the DNA databases will be reduced somewhat if the legal authority to sample and type offenders’ DNA continues to expand and come to include the multitudes convicted of lesser, but more numerous, felonies and misdemeanors.”25
We can use current FBI arrestee statistics to evaluate the claims about racial disparities in expanding CODIS. According to the 2007 FBI data on national arrests, there were 9,014,180 ind
ividuals 18 years or older arrested in the United States; 70.2 percent (6,327,954) were white and 27.7 percent (2,496,928) were black.26 If we use these figures, adding all the arrestees to CODIS would tend to bring down its racial composition from an estimated 40 percent blacks to slightly less because we are diluting what is believed to be a higher relative percentage of blacks with a lower relative percentage. Because DNA profiles might be taken only of arrestees who commit violent crimes, let us consider those figures. The FBI classifies violent crimes as murder or nonnegligent manslaughter, forcible rape, robbery, and aggravated assault. In 2007, 376,745 individuals were arrested for violent crimes as defined by these four categories, of whom 63 percent were white and 36 percent were black. If we added these arrestee profiles to CODIS, the percentage of blacks would be reduced slightly from 40 percent. In other words, adding arrestees who had committed violent crimes to CODIS would increase the racial composition of blacks only if their current composition in CODIS were less than 36 percent.
Of the four categories of violent crimes, only two (murder/manslaughter and robbery) have a higher percentage of black arrestees than whites (50 versus 48 and 53 versus 46, respectively). The racial disparity of blacks in CODIS would increase only slightly if arrestees charged with murder/manslaughter and robbery felonies (and then released) were the only violent perpetrators added to the database because of the relatively small numbers of these crimes.
However, we should not be concerned only with the relative proportion of whites and blacks in the database; we should also consider the proportion of the African American population that is in the database compared with the proportion of the white population. As stated earlier, currently 8.5 percent of African Americans are in the database. If we assume that these are predominantly males (approximately 80 percent of individuals arrested for violent crimes are male, for example), then we are approaching a situation where as many as 14 percent of African American males are in the database, even before we have started to upload arrestees to the system. Of the approximately 2.5 million arrests of blacks that occur each year,27 a large proportion of those arrests will not result in conviction. The U.S. Department of Justice reports that 47 percent of the more than 140,000 individuals arrested for a federal felony offense in 2004 were not convicted.28 Similarly, more than 30 percent of the hundreds of thousands of individuals arrested on suspicion of a felony each year in California are never convicted of a crime.29 Conviction rates for lower-level crimes are even lower; a study in California revealed that 64 percent of drug arrests of whites and 92 percent of those of blacks were not sustainable.30 Assuming a 50 percent conviction rate for all arrests, this means that of the 2.5 million blacks arrested each year, half will be added to the database who would not have been added under a system that requires conviction for inclusion in CODIS. In just the first year the percentage of the African American population represented in CODIS, from the addition only of unconvicted blacks, would rise to 11.8 percent of the African American population, or close to 19 percent of the male African American population. In comparison, still only 3 percent of whites would be in the database (or as many as 5 percent of white males, under the same assumptions). Predicting how this would play out over time is difficult because one would have to account for repeat arrests or, alternatively, have an estimate for one’s lifetime risk of being arrested. Such estimates have been reported as 80 to 90 percent for urban black men.31 If we layer on top of this the fact that blacks are inappropriately arrested at higher rates than whites, then blacks who are innocent under the law will also be disproportionately represented in the database. This may be especially true where data banks are expanded to arrests for nonviolent (“victimless”) crimes, where police have significant discretion in making arrests. As stated earlier, a 1993 California study revealed that while 64 percent of drug arrests of whites were not sustainable, a full 92 percent of the black men arrested on drug charges were subsequently released for lack of evidence or inadmissible evidence. There are nearly 2 million drug-abuse arrests annually.32 Furthermore, if the use of controlled substances as a percentage of population is equal to or greater in the white community than in the black community while the arrests are disproportionately higher for the latter, then not only are blacks treated unequally by law-enforcement officials but their DNA makes them and their family objects of continued genetic surveillance.
Because blacks are represented in CODIS and in prisons at a much higher percentage than their relative composition in the general population and at a higher percentage relative to nonminorities, if we expand the use of forensic DNA technology to the existing disparity, it will produce downstream effects that will further exacerbate racial prejudice. This is illustrated by the use of familial searching, as in the case when a close, but not exact, match is made between a crime-scene profile and an individual who has a DNA profile in CODIS. When arrestees who have not been convicted of a crime are added, their families are also brought under the surveillance lens of criminal justice. This conclusion was reached by Greely and colleagues:
Assume that, either using the current CODIS markers or an expansion to roughly twice as many markers, partial matches of crime scene DNA samples to the CODIS Offender Index could generate useful leads from an offender’s first degree relatives—parents, siblings and children. . . . Using some additional sampling assumptions, the percentage of African-Americans who might be identified as suspects through this method would be roughly four to five times as high as the corresponding percentage of U.S. Caucasians. . . . More than four times as much of the African-American population as the U.S. Caucasian population would be “under surveillance” as a result of family forensic DNA and the vast majority of those people would be relatives of offenders, not offenders themselves. . . . African-Americans are disproportionately harmed by crimes committed by other African-Americans.33
In the United Kingdom, England and Wales have the largest per capita forensic DNA data bank in the world. At the end of March 2009 there were 4.8 million DNA profiles in the database, representing about 7 percent of the population in England and Wales, compared with the national DNA data bank in the United States, which contains 2.6 percent of its population. Recent data reveal that more than one-third of black men in the United Kingdom have DNA profiles in the national DNA data bank, which includes three out of four black males between the ages of 15 and 34.34 If the United States were to reach that percentage, CODIS would have 13 million DNA profiles of African American males.
Rounding Up the Usual Suspects
Human DNA is shed continuously in all the environments in which people interact. Our saliva is left on paper cups we discard in a coffee shop or empty soda cans we leave behind at a park. We leave hair fragments in bathrooms of restaurants or on shirts we bring to the cleaners. Swabbing for DNA at a crime scene has become almost as common as screening for fingerprints. Newer techniques allow investigators to retrieve trace DNA samples from touched objects. Obtaining profiles of DNA left at a crime scene may sometimes be useful to investigators, but it also may be irrelevant without other forms of evidence. With massive DNA data banks, every DNA sample found at a crime scene that matches someone in the data bank will become grounds for investigation—for no other reason than that the evidence shows that either the person or his or her discarded objects were once at the place of interest.
Let us suppose that there is a robbery and shooting at a Starbucks coffee shop. An employee at the coffee shop believes that the hooded robber had a cup of coffee before the robbery. Police obtain all the discarded cups for DNA samples on the premises. Once the profiling of the DNA is completed, investigators submit the profiles to the national and state DNA data banks, which by 10 years from today may have more profiles of innocent people, including juveniles, than of convicted felons.
If African Americans are disproportionately overrepresented in CODIS, not because of prior convictions, but because blacks are stopped and arrested more frequently than whites, then they are more likely
to become suspects in any crime that involves a DNA sweep (see figure 15.1). An old adage claims that crime-scene investigators will go where the evidence takes them. In this technological era we can say with some degree of confidence that crime-scene investigators will go where the DNA takes them. The greater the disparity in the percentage of African American DNA profiles in CODIS, the greater will be the disparity in suspicion of blacks, false arrests, and false convictions.
FIGURE 15.1. The cycle of racial disparity in DNA data banks. This flowchart depicts the increasing and compounding racialization of CODIS under a system of expanded criteria for collection and retention of DNA profiles in state and federal databases. Minorities, more likely to be arrested, are disproportionately represented on the database. Racial disparities in the criminal justice system are then further exacerbated, since, once in the database, these individuals are more likely to be arrested again in the future. Source: Authors.
In the hypothetical Starbucks shooting, perhaps there were a dozen cups in the trash with sufficient cells for DNA analysis. If only two profiles showed up in CODIS that matched the DNA profiles on the cups, those would be the people the police would investigate, interrogate, and follow around for the reason that one pursues the leads one has even though what the police have became potential evidence because the database used to match “abandoned” DNA to profiles in CODIS was weighted disproportionately toward minorities. Given the racial composition of the data bank it would be no surprise to learn that both of the profiles were from black males.
Genetic Justice Page 32