The Color of Crime Coverage

The preceding passage opens the article “Blacks’ presumed guilt hits a little too close to home,” written by Leonard Pitts, Jr. The experience faced by his son, daughter, and grandson is sadly typical for black families but not for white families. Pitts’s son, who obviously did not fit the description of the robbery suspect, save for the fact that both are black, spent two weeks in jail before the charges were dropped. When it comes to criminal suspects, the sentiments often seem to be that any black man will do.⁷ It is simply easier and more automatic to associate crime with blackness or brownness than to actually figure out who might have committed harm (which you might think is the proper role of police work).⁸ This is especially true for violent crime. For some, when they hear about a drug bust or a shooting, they spontaneously conjure up an image of a black man more often than that of a white man, and once that equation between blackness and crime is in place, any black person can be seen as a criminal.

That’s why Patricia Ripley, who was indicted for drowning her son in 2020, initially claimed that two black men abducted him.⁹ And why, shortly before the 2008 US presidential election, a white campaign worker for presidential candidate John McCain concocted then recanted a story that a 6-foot-4 black man attacked her and carved a “B” (for “Barack”) in her face showing his support for presidential candidate Barack Obama.¹⁰ And why “runaway bride” Jennifer Wilbanks fabricated a story in 2005 about being kidnapped and raped by a Latine man. And why Susan Smith, after murdering her sons in 1994, covered up her crime by fabricating a story about a black man abducting the boys. And why Charles Stuart, after shooting his wife in the head and himself in the abdomen, attempted to blame the 1989 incident on an African American man.¹¹ In each of these cases, the fabricator thought they could get away with these lies because they well understood the equation, in the eyes of law enforcement and the public, between men of color and crime. Blaming black people is not new. Frederick Douglass, famous orator and abolitionist, said in an 1883 speech, “Taking advantage of the general disposition in this country to impute crime to color, white men color their faces to commit crime and wash off the hated color to escape punishment.”¹²

Crime coverage in the news media helps viewers form schemas, or harden existing schemas, between criminality and color. For example, Muslims are greatly overrepresented as terrorists in news stories.¹³ This slanted coverage likely plays a role in the rise in anti-Muslim hate crimes over the last two decades.¹⁴ In fact, most terrorism inside the United States is committed not by Muslims, but by white Christian men – crimes such as mass shootings and right-wing terror. Between 2008 and 2018, white right-wing extremists have been responsible for 73% of extremist-related domestic-terror murders, compared to Islamic extremists who were responsible for only 23%. In fact, in 2018, right-wing extremists were responsible for every extremist murder in the United States.¹⁵ In 2021, the FBI and Department of Homeland Security found that right-wing and white supremacist organizations constitute the “most lethal threat categories.”¹⁶

Latine individuals are overrepresented as perpetrators compared to their real-life behavior, and they are underrepresented as victims compared to their actual victimization in real life.¹⁷ Depictions of blacks have also shown them overrepresented as criminal perpetrators and underrepresented as victims of crime, compared to actual crime rates, although a more recent study found that African Americans were accurately portrayed in terms of their numbers as victims and perpetrators.¹⁸ One pattern that has not changed is the extent to which white people are portrayed positively and sympathetically in news coverage. In news media, whites are overrepresented as homicide victims and as police officers compared to their actual numbers¹⁹ and underrepresented as criminal perpetrators compared to real-life crime rates.²⁰ Warped depictions like these teach viewers that white people are less scary than they really are and people of color are scarier than they really are. African American suspects often appear in more hostile and frightening ways than white suspects.²¹ Blacks are much more likely than whites to be portrayed in police custody and in mug shots. African American men, like those arrested in Tulia, Texas, are also more likely to be portrayed in the media as threatening and menacing (e.g., handcuffed, physically restrained by police officers, poorly dressed) than white suspects.

The cognitive association of blackness with crime leads white people to judge the same behavior by African Americans and whites through a prism of stereotypes. A study of white American university students by Eaaron Henderson-King and Richard Nisbett found that observing or even hearing about an African American man’s negative actions caused white people to avoid contact with a different black man later.²² The respondents who saw a black man behave negatively subsequently perceived African Americans in general as more hostile. In other words, the behavior of one member of a minority group is generalized to other members of that same group – individual people of color are stand-ins for their entire group. Not so for white people. Participants in Henderson-King and Nisbett’s study did not base their judgments of white men on the negative behavior of one white man. Whites were viewed as individuals, not as representative members of an ethnic group. In fact, in their study, when the student participants observed a white man behaving negatively, they subsequently avoided a black man. Simply witnessing a negative interaction, regardless of the race of the interactants, made the white participants want to be more comfortable, feeling safer associating only with “their own” versus “others” (even when it was one of “their own” who had behaved negatively). As we saw in Chapter 2, people tend to attribute the negative actions of outgroup members to dispositional rather than situational factors, whereas similar actions by ingroup members are more often assumed to be the result of situational factors.²³ The patterns revealed in Henderson-King and Nisbett’s study have ripple effects beyond the stimuli they used in their study because those who witnessed or heard about a black man doing something negative also minimized the existing power difference between whites and African Americans in the United States, compared to those who witnessed or heard about a black man doing something neutral or positive, or a white man doing something negative. In other words, when a single African American man does something bad, whites tend to generalize that to all African American men, and then the threat they feel when they witness this negative behavior makes them less sympathetic to the struggles of African Americans in the United States.

The School to Prison Pipeline

The criminalization of people of color begins early in their lives and manifests in different types of discipline in schools. The dramatic rise in the policing of schools through the deployment of so-called student resource officers has created a through line from people’s ideas of who criminals are to how even very young children are disciplined in schools. In 1975, 1% of schools in the United States had police officers; by 2021, 48% of schools did.²⁴ Even though there are tragic cases where armed officers have profoundly failed to protect students, some parents and politicians believe that armed police officers in schools keep kids safe. This assumption, however, is incorrect. A study of 25 school shootings found that none of them were stopped by armed officers. Instead, these shootings typically ended when the shooter was tackled by unarmed staff or when the shooter simply decided to stop.²⁵ The presence of police in schools blurs the line between criminal behavior and everyday youth behavior.²⁶ When children are criminalized and schools are militarized, students are arrested for non-serious issues such as cursing, not following directions, and making burping and fart sounds.²⁷ Some teachers and administrators view mischief, “attitude,” and disruption as criminal offenses worthy of involving law enforcement. School principals are more likely to discipline African American students and to see them as “troublemakers” for the exact same infractions as white students.²⁸ Police officers, not school staff, seem to be in charge of school discipline. Police in schools have replaced counselors, nurses, social workers, and teachers. According to a CBS News study, children with disabilities are four times more likely to be arrested at school.²⁹ Even though they make up only 14% of the US population, black children make up nearly half of all arrests at elementary schools.³⁰ That’s right, arrests at elementary schools.

That armed police are now in schools in combination with the criminalization of black and brown people produces a toxic stew that harms the lives of some children while giving others the benefit of the doubt, just as we saw in studies described earlier. Unfortunately, schools mirror the general society that criminalizes and pathologizes girls and boys of color. Starting in preschool, African American children are four times more likely to be suspended than white children.³¹ Consequently, what has been described as a school-to-prison pipeline could be reasonably termed a preschool-to-prison pipeline for youth of color. Black boys are disciplined around five times as often as white boys. White individuals incorrectly read the facial expressions of grown black men as hostile when their faces are neutral, and similar patterns of misinterpretation also harm even very young African American children. The ambiguous behavior of sixth graders is interpreted as aggressive when the children are black,³² revealing just how early black children are associated with aggression and criminality. Some studies show that black children, even as young as five, evoke thoughts of guns and violence.³³ Incidents such as fights in school that could be handled by counselors, visits to the vice principal’s office, or even academic suspension are instead channeled through the criminal legal system for children of color. And once a person of any age gets caught in the machinery of the legal system, it is difficult to get out. It is well established that African Americans are more likely to be cited for low-level infractions than whites who commit infractions at the same rates. Furthermore, blacks are more likely to be indicted, less likely to be offered bail, more likely to be convicted, and more likely to serve longer sentences than whites for the same crimes. A study of mostly white women who were studying to be teachers found that these prospective teachers could not accurately assess the emotions conveyed in the faces of black and white children. Both black girls and black boys were more often viewed as angry, when they were not, whereas this perceptual bias did not occur regarding white boys and white girls.³⁴

Black girls are pathologized and criminalized in similar ways to boys of color, yet this racial difference in girls is under-studied and under-acknowledged.³⁵ Girls of color tend to be excluded from analyses of racism or sexism in schools, amounting to what Kimberlé Crenshaw calls an intersectional failure. Crenshaw notes that, in fact, the different treatment of black and white girls is more disparate than the bias against black boys relative to white boys. Specifically, black girls are disciplined around 10 times more often than white girls.³⁶ Black girls are particularly likely to be targets for harsh discipline. A risk of suspension is actually higher for black girls when compared to white girls, and higher than it is for black boys compared to white boys. In other words, racial disparities in discipline are more dramatic when we consider the treatment of girls.

According to Devon Price, who wrote Unmasking Autism, individuals with autism who are black are in particular peril. He notes that if a black boy or man is walking down the street flapping hands, flailing arms, not making eye contact, or otherwise acting in ways sometimes associated with neurodivergence, it can result in the police being called.³⁷ Poor youth of all races are criminalized as well. Police are more likely to refer lower-class youths to juvenile courts, and they are more likely to be institutionalized than wealthier kids. The police generally deal with wealthier youngsters informally (if the police are even involved in the first place), taking them home and having a talk with their parents, rather than instituting formal procedures. Wealthier kids who do end up in court are more likely to receive probation than poorer ones. These same patterns hold for black and white youth.³⁸

Automatic Associations and the Shooter Bias

Michelle Cusseaux was 50 years old when she was shot and killed by police during a court-ordered mental-health pick up. Eight officers arrived at her house and the officer who shot her invoked self-defense, claiming that she was holding a hammer.³⁹ Willie McCoy, a 20-year-old, was fired at 55 times in 3.5 seconds when police found him sleeping in his car at a Taco Bell.⁴⁰ The police insisted they had fired 55 times out of “fear for their own safety” because McCoy allegedly reached for his gun. When body camera footage was finally released, the footage showed that McCoy had merely scratched his shoulder.⁴¹ Police killed 15-year-old Jordan Edwards when they shot at a car in which Edwards was a passenger. Police were responding to a report of underage drinking.⁴² The officer who shot Edwards originally claimed that he started shooting because the car was headed toward his partner, but body cam footage showed the car was actually driving away. Police shot Donovan Lewis, aged 20, while he slept in his bed when the officers served a warrant. Body cam footage shows a bright light on Donovan and as he starts to sit up the officer fires his weapon. Lewis was unarmed.⁴³ And 61-year-old Michael Craig was shot after calling the police to report that his wife was attacking him with a knife. When police arrived, they saw the two fighting and shot and killed Craig.⁴⁴ In all of these cases, those who were shot were black.

A review of 7,675 police killings in the United States between 2013 and 2021 finds disturbing patterns.⁴⁵ As we have sadly come to expect, black victims were overrepresented relative to the general population, and white victims were underrepresented. Maybe African Americans are shot more by police because they are more likely to pose a danger? If the danger is the mere fact of blackness, then yes, but if the danger is being armed, no. Relative to their white peers, blacks were 23% less likely to be armed during the time they were killed, and Latine victims were 45% less likely. These racial patterns hold regardless of the year of the encounter, age of victim, gender of victim, and the geographical location of encounters. White victims then appeared to pose greater threats to the safety of police officers than black or Latine victims, yet were underrepresented in police killings relative to the general US population. Another study looked at police shootings in 213 metro areas in the United States over a 21-year period – cases in which the victim of the shooting died. Officers tended to use deadly force more in response to black suspects, compared to white suspects, even after controlling for criminal activity (e.g., suspect attacked officer, attacked citizen, fled from scene, resisted arrest).⁴⁶

We do not know exactly what went through the minds of those officers who make the split decision to shoot or not to shoot a person. But social psychology experiments offer a glimpse into the patterns of the quick decisions law enforcement officers make based on race. Joshua Correll and colleagues asked African American and white participants to play a video game simulation in which they had to decide whether or not to shoot a “suspect” holding an ambiguous looking object.⁴⁷ If the suspect was holding a cell phone, the respondent must refrain from shooting, but if the suspect was holding a gun, the respondent must shoot. Participants viewed both black and white men holding the objects. Monetary awards were provided to the participants to encourage speed and accuracy. Results of the study revealed that participants fired at an armed target more quickly if he was black than if he was white, and they decided not to shoot an unarmed white target more quickly than an unarmed black target. The tendency to make more false alarms (shooting the unarmed person) than misses was more pronounced when the target was black than white. That is, they tended to shoot unarmed African Americans more often than unarmed whites. If a target was African American, respondents required less certainty that he was, in fact, holding a gun, before they decided to shoot him. Thus, race (i.e., ideas about race) interfered with the ability of the respondents to be accurate. This pattern occurred among both white and black participants – that is, African American participants were also quicker to shoot at the black unarmed suspect than the white unarmed suspect. Surely respondents could have performed perfectly on the task by simply focusing on the object in the target’s hand, completely ignoring the target’s race. But participants were overwhelmed by their cognitive link between violence and black men.

Since Correll’s study, dozens of other studies have supported these findings. For example, a meta-analysis that compiled 42 studies on shooter bias found that compared with white targets, participants were quicker to shoot armed black targets, slower to not shoot unarmed black targets, and were more likely to have a liberal shooting threshold for black targets. They were not, however, more likely to have a higher false alarm rate for black (versus white) targets.⁴⁸ White Australians have also been shown to more quickly shoot armed targets who were black rather than white, and they were more likely to shoot unarmed black targets compared to unarmed white targets.⁴⁹ White Germans were found to more quickly shoot armed Arab-Muslim targets than white targets (but there was no difference when targets were unarmed).⁵⁰ The findings of simulation studies demonstrate the automatic link for many people between black men and violence. That is, for many people their schema for violence is more likely to contain a black person than a white person; their schema for black people is more likely to contain violence, compared to their schema for white people.

It seems that for many people, the categorical schema of “black person” or “black man” is quite powerful, but schemas are blunt instruments and often contain simplistic ideas and images of people. Mary Beth Oliver and Dana Fonash asked white US college students to read a news story about either a violent or non-violent crime that involved either a white or black suspect.⁵¹ They were later shown photographs of white and black men and asked whether or not they had seen the men before in the stories they had read. Readers were more likely to misidentify the (wrong) black man as the suspect than the (wrong) white man but only when the crime was a violent one. For white respondents, violent black suspects again tended to all look alike. Another study found that white North Americans were worse at judging authentic smiles in black faces than white faces. Black people were accurate for both. The researchers found that white respondents were bad at this because they spent more time looking at white faces than black faces. So, they jump to conclusions about black faces and are more careful and deliberate with white faces. White faces matter more.⁵² These findings suggest that there could be problems in identifying suspects in police photos or lineups. There are significant implications here for relying upon the accuracy of the eyewitness testimony of white witnesses in cases involving black suspects.

Stand Your Ground Laws and White Vigilantism

The preceding work addresses blacks as targets of police violence. Benign bigotry often takes the form of bias and discrimination in laws that are facially neutral, that is, laws that seemingly apply to everyone but are disparate in their application – one group benefits from the law, another group becomes targets of it. Stand Your Ground laws are one of those laws. Stand Your Ground laws are a relatively new type of law (Florida’s in 2005 is the standard) that massively expands the legal notion of self-defense. Common law self-defense requires a person to retreat if possible before using deadly force on an attacker. If no safe retreat exists, force, including deadly force, may then be used to prevent imminent death or great bodily harm. Stand Your Ground laws eliminate the duty to retreat and presumes a person using force had a reasonable fear of imminent death or great bodily harm. Nichole Hamsher offers a damning analysis of Stand Your Ground laws.⁵³ Hamsher argues that allowing anyone to assume that harm is imminent, and that they should take the first shot, blurs the line between self-protection and vigilantism. Removing the duty to retreat changes the whole idea of self-defense because all a person has to do is claim they perceived a threat and then claim their violence is justified. The consequences of this broad use of perceived threat is that there is little incentive to diffuse arguments or to retreat from the situation.

Bias in Stand Your Ground laws manifest in at least two ways. First, if an armed person believes that people of color, black people specifically, by their nature are dangerous and are threatening, they are likely to use deadly force against them for no other reason than the victim’s race, and both law enforcement and juries are inclined to allow such force. A second way in which Stand Your Ground laws are biased is who gets to invoke them in the first place. White people successfully use the Stand Your Ground defense when they kill someone, for the most part, but people of color don’t enjoy such a defense. When white people kill someone, the criminal legal system seems to take their word for it if they claim self-defense. The same benefit of doubt is not extended to people of color. In a study of more than 200 Florida cases of killings with Stand Your Ground as a defense, in 79% of the cases, safe retreat was possible, and in 68%, the person killed was unarmed.⁵⁴ Stand Your Ground laws remove imminence which has been the foundation of self-defense. Removing imminence allows space for other components to slip in, filling in the factual gaps to make a situation look like self-defense and not something more sinister – like racial bias. The risk is that someone will be afraid of a black person and will kill them in “self-defense,” not because the victim was actually threatening, but because their mere existence and presence as a black person was perceived as threatening.

The consequences are sadly predictable. When analyzed by race, homicides in Florida with a white perpetrator and a black victim are ruled justified in 16% of the cases, but only 1% are ruled justified when a black person shoots a white victim. In cases where there is one shooter and victim – both strangers, both men – the number of homicides justified by Stand Your Ground is 44.71% when the shooter is white and the victim black. In contrast, when the shooter is black and the victim white, the total falls to just 11%.⁵⁵ Allowing persons to defend themselves in any place they have a legal right to be, as Stand Your Ground laws expressly do, assumes that only one person in the scenario possesses the right to be where they are. And that person is white. Consider belief superiority – a common flaw in thinking, the idea that one’s own beliefs are superior to others’⁵⁶ – in combination with white supremacy, and you have a license to kill black people if you are white. Black men are likely to be victims of Stand Your Ground laws, but black women are particularly likely to not be able to use them as justification due to enduring stereotypes of black women as aggressive, fearless, angry, and violating traditional notions of white femininity.⁵⁷ And thus, a black woman is much less likely to be able to utilize the Stand Your Ground defense as self-defense. In addition, because Stand Your Ground laws were conceived of and in practice are meant for men to not have to retreat in an altercation between strangers, domestic violence survivors have not successfully utilized the laws for self-defense against abusers.⁵⁸ Marissa Alexander, a black woman in Florida was sentenced to 20 years in prison for firing a warning shot into a wall with her legal firearm after her estranged husband assaulted her. She was not afforded immunity under Florida’s Stand Your Ground defense.⁵⁹

Bias in Stops, Investigations, and Arrests

Again, laws tend to be facially neutral – there aren’t laws written specifically for certain racial groups so there is an air of neutrality and objectivity in the law. Stop-and-frisk is one practice that disproportionately targets people of color relative to whites. A stop-and-frisk occurs when a police officer stops a person for questioning and pats down exterior clothing to see if the person is carrying a weapon or drugs. A stop-and-frisk should be based on reasonable suspicion but often that suspicion is existing as a person of color rather than an articulable suspicion. In addition to the racially disparate impact, these stops are also generally a waste of time because they usually lack legitimate probable cause – unless the probable cause is blackness. Take New York City. Over a 10-year period, when these kinds of stops were at their peak, the overwhelming majority of those stops went nowhere – 90% of the stops did not end in a citation or arrest.⁶⁰ In addition to the practice being a waste of time and resources, and possibly unconstitutional, it puts those stopped in physical jeopardy. In 2021, only 8% of the stops were of white people, 87% were either Latine or black.⁶¹

Traffic stops also reveal particular targeting of people of color. For example, in California, sheriffs stop African Americans twice as often as they do white drivers.⁶² Police departments claim that the stops they do are not based on race of the driver. However, even when various factors are held constant, police generally pull over or don’t pull over people because of their race or neighborhood – which is often a proxy for race and class. For example, one report analyzed stops in San Diego County, California. Even though the US border town of San Ysidro (94% Latine) has a similar number of residents and the same number of reported crimes as La Jolla (75% white) – one of the most well-to-do communities in southern California – police pull over nearly three times as many drivers in San Ysidro as La Jolla. In San Diego, where black people are about 6% of the population, they make up 20% of the stops. In affluent neighborhoods with even higher reports of serious or violent offenses, police make less than half the stops per crime than they do in working-class neighborhoods. Over and over, San Diego police are found to stop black drivers much more than white drivers and, to be clear, these stops are initiated by officers, not by civilians calling in to report an actual incident.⁶³

An analysis of the Toronto Police Service finds similar patterns. Black people are grossly overrepresented in criminal charges. Though they represent only 9% of Toronto’s population, black people represent 32% of the charges by Toronto police. In other words, black people are nearly four times more likely to appear in the charge dataset than their representation in the general population would predict. By contrast, white people and people from other racial minority groups are underrepresented in charges. The black charge rate is four times greater than the white rate and seven times greater than the rate for people from other racial minority groups. Interestingly, the charges against white suspects were slightly more likely to lead to a conviction (23% vs. 18%) compared to charges against blacks.⁶⁴ Whereas the black conviction rate is three times higher than the white conviction rate, the black non-conviction rate is four times higher than the white non-conviction rate. These data suggest that black people are especially vulnerable to unnecessary, low-quality (flimsy) charges that ultimately lead to non-conviction.

Immigration violations may also seem like a neutral law where anyone from any country could be subjected to them, but in fact, are a ripe area for racial bias. Sahana Mukherjee and colleagues created an experiment to examine whether immigration violations are applied evenly across racial and ethnic groups.⁶⁵ Mukherjee asked white US citizens to read a fictional transcript of a traffic stop for speeding. The car carried occupants who were either white (Canadian or Irish) or Mexican in origin. The transcript described the officer asking the occupants for identification of their citizenship status and when they did not produce the documents, they were detained by the officer. The transcript of the stop varied only according to the ethnicity of the car occupants. The study respondents were asked various questions about the stop. Did participants respond to immigration violations no matter who the violators are, or do they believe some people deserve a punitive response while others do not? Participants expressed greater suspicion of undocumented status and endorsed more punishment when occupants were Mexican, compared to Canadian or Irish. Participants thus show support for punitive enforcement of traffic violations for brown people, but support little scrutiny for the same violations if they are white. Mukherjee’s research suggests if officers selectively target people based on identity cues, they are not only illegally discriminating, but also allowing potential undocumented targets who do assimilate to Anglocentric norms to avoid scrutiny.⁶⁶

Investigations and Interrogations

Contrary to a cherished judicial ideal – the assumption that a suspect is considered innocent until proven guilty – bias operates in forceful ways from the early stages in the criminal legal process. Once a suspect has been identified, police tend to stop searching for other suspects and stop pursuing other leads. This narrowed focus is due, in part, to the lack of resources departments have to maintain investigations of several suspects simultaneously. Perhaps as important as the constraints of material resources, however, is the constraints on one’s cognitive resources. Once investigators and prosecutors have targeted a suspect and have justified the massive amount of time required for a thorough investigation, there is a tendency to close off alternatives in order to further justify their pursuit of the main suspect. The reasoning runs something like this: “How can I try so hard to identify and build a case against a suspect if there might be other suspects out there?” And thus, investigators zero in on their suspect in order to justify the effort they are expending in the case.⁶⁷ Part of the reason the public believes that those who are suspected of a crime are actually guilty is due to the belief that there are many safety nets along the legal process that protect the innocent from wrongful suspicion and conviction. Many believe that if a suspect is truly innocent, the truth will come out at some time during the process. (Or like Edwin Meese in the opening epigraph, they believe that someone wouldn’t be a suspect in the first place if they weren’t guilty.) In a compilation of the literature that unites psychology and law, Saul Kassin finds, in fact, innocence actually puts innocents at risk for being found guilty.⁶⁸ In other words, suspects who are actually innocent of a crime are at more risk than those who are actually guilty, when it comes to assumptions of guilt. According to Kassin, there are five critical points in the judicial process that are especially risky for innocent suspects.

Bias and the tendency to “see guilt” affect all stages of the criminal legal process, even after the investigation, arrest, and interrogation. A dramatic example of the effects of this type of bias was revealed in the 2002 scandal that drew national attention and forced the Houston, Texas, Police Department Crime Lab to close its doors because of widespread fraud and incompetence.⁶⁹ Crime analysts were found to have acted as a “rubber stamp” for the police department, tailoring lab reports to fit the theories of the prosecution, and ignoring exculpatory evidence that would have excluded defendants.⁷⁰ It is unclear whether or not these serious mistakes were the result of conscious efforts or unconscious thinking, although the mistakes always favored the prosecution.⁷¹

The first risk to innocent suspects is law enforcement officers’ overconfidence in their ability to track the correct leads, and the fact that narrowing the pool of suspects actually tends to lead those involved in pursuing criminal investigations to presume guilt prematurely. People in general engage in confirmation bias, once they form an impression, they unwittingly seek, interpret, and create observational data that verify that impression. In contrast, evidence that contradicts the impression is minimized and avoided. This tendency reveals an inversion of the conventional wisdom: “I wouldn’t have believed it if I hadn’t seen it.” In the criminal legal process, after the early stages of investigation and identification of a suspect, the wisdom is perhaps captured in this way: “I wouldn’t have seen it if I hadn’t believed it.” Across many studies, trained professionals, compared to untrained controls, exhibit a proclivity to judge targets as being deceptive. An assumption for many police investigators seems to be: “You can tell if a suspect is lying by whether he is moving his lips.”⁷²

Once there is a suspect in a crime, they become the suspect, and this motivates investigators to look at the circumstances and facts of the case only in terms of their particular suspect – as if there are no other possible suspects. Once investigators put together a story in which the circumstances fit the suspect, it’s difficult to perceive alternatives. This creates a (sometimes false) sense of confidence for the investigators and puts them on a track that seems to inevitably point to the already identified suspect. By the time a suspect is formally questioned, the goal of the interaction is to produce a confession.

Two experiments on interrogations and confessions conducted by Saul Kassin illustrate how investigators’ assumptions can hamper their ability to detect deception accurately. In one study, Kassin and colleagues tested the accuracy of law enforcement officers and students in judging confessions – both true and false.⁷³ Kassin showed the mostly white American students and police officers either video with sound or only audio recordings of incarcerated men giving either a true confession (of the crime they actually committed) or a scripted false confession. One might expect that police officers, with their training and experience, would be able to detect subtle clues or the common tells in a suspect that reveal deception – lack of eye contact, stammering and stuttering speech, nervous fidgeting, perhaps. The results were revealing. Neither students nor law enforcement performed better than would be predicted by chance. Rolling dice to decide who was telling the truth and who was lying would have yielded about the same results. But there were additional results that were concerning. First, law enforcement officers actually performed worse than students – students were accurate 59% of the time and law enforcement officers 48%. How can this be? Police officers, compared to students, presume guilt more often. In this study, the officers guessed that 65% of the confessions were true compared with students who guessed that 55% were true (in fact, half of the confessions were true). What’s more, police officers were more confident in their assessments than were college students, and confidence was negatively correlated with accuracy. That means the more confident the observer, the less accurate the judgment. Finally, raters (regardless of whether they were students or law enforcement officers) made more accurate judgments when they listened to audio recordings of the confessions rather than when they watched video with sound. It seems that the visual cues people use to determine guilt or innocence actually contribute to their inaccuracy – judges are better off just listening.

The implications for fairness and accuracy in the criminal legal system are widespread. Police officers are, in effect, trained to presume guilt. Their confidence in lie detection hinders their accurate evaluation of evidence. The combination of presuming guilt and their mislaid confidence in their own abilities to detect lies, leads some officers to make false-alarm responses – that is, concluding a truth-teller is lying.⁷⁴ When officers have been asked whether or not they are worried that they might be interrogating innocent people and presuming they are guilty, a common response from some investigators is: “No, because I do not interrogate innocent people.”⁷⁵ That people were more accurate when they listened to confessions without video has implications for training because many law enforcement investigators believe the way they detect characteristics about a suspect is visual – body language, visibly detectable nervousness, and so on. Despite training, professional standards, and popular notions, there is no group of people (e.g., police, FBI, psychotherapists) that reliably differentiates lies from truths at levels that are better than by chance. People from fields that provide lie detection training usually do not perform better than do civilian college students.⁷⁶

Naively believing that truth and justice will prevail, innocent suspects often waive their rights to be questioned and to have an attorney present. This is the second risk described by Kassin. Innocent suspects overestimate their ability to prove their innocence. They think that, because they have nothing to hide, they can waive their Miranda rights.⁷⁷ The Miranda warning was established in 1966 to protect suspects against coerced confessions and self-incrimination. Ever since the landmark US Supreme Court decision Miranda v. Arizona, police must warn suspects of certain rights before starting an interrogation. These rights include the right to remain silent when in police custody. Suspects do not need to talk or answer any questions law enforcement may ask. Suspects also have the right to an attorney, either hired or court-appointed if they cannot afford one.

Relevant to the discussion here, innocent individuals and those who have no prior felony record are more likely to waive their Miranda rights than those with a history of criminal legal experience. The waiving of these crucial rights means that suspects allow police to interrogate them with no protection from an attorney. Those with criminal legal experience seem to better know their rights and protect themselves in ways that innocent people and those with less experience people do not. In one of Kassin’s studies, some participants (mostly white American college students) committed a mock theft and some did not, but all were suspects.⁷⁸ Those who were innocent were much more likely to sign a waiver giving up important protections than those who were guilty (81% vs. 36%). Those who were innocent were also more likely to waive a regular police line up and instead agreed to a one-on-one “show-up” with the victim – a procedure that put them in more danger of being identified by the victim. Among guilty participants, only 47% waived the line up; but among the innocents, 100% waived the line up. Why do innocent people waive these important rights, while those who are guilty tend not to? According to Kassin, there are at least two reasons. First, innocent people tend to believe in an illusion of transparency, which is the tendency to overestimate the extent to which people’s true thoughts, emotions, and motivations can be seen by others. Their thinking seems to be, “If I am honest and show that I have nothing to hide, they will see that I am not guilty.” In terms of waiving a line up, the innocent suspect figures, “I didn’t commit the crime so a witness cannot identify me.”

A second reason why innocent suspects put themselves at risk by waiving important rights is because they tend to believe in a just world.⁷⁹ Belief in a just world is a construct that explains a host of phenomena from why people tend to blame rape survivors for being raped to why the guilty-of-something myth persists. The belief in a just world is the belief that good things happen to good people and bad things happen to bad people. It’s the notion that if we are good, decent, hard-working people, good things will come our way. If something bad happens to us, we must have done something to deserve it. Believing in a just world is functional in that it makes people feel better. It allows people to believe that the world is a fair and predictable place. For the purposes of our discussion here, innocent people’s belief in a just world contributes to their sometimes faulty belief that the truth will come out in the end. From the perspective of those judging others, the belief contributes to the myth that if someone is accused of a crime, they must be guilty of something because it would be too difficult to cope with the fact that sometimes an innocent person is railroaded into being found guilty. People want to believe that the legal system is just.

The third risk to innocent suspects is that, in spite of or because of their plausible and vigorous denials, they trigger highly confrontational interrogations by interrogators. Kassin’s study comparing students’ and law enforcement officers’ ability to detect true and false confessions only looked at the judgments of confessions, not the interactive dynamics of interrogations. Do police question innocent and guilty suspects differently? Are there techniques that interrogators use that put innocent people at risk of prosecution and even conviction? Another set of studies conducted by Kassin addresses these questions.⁸⁰

Again, the interrogation is a guilt-presumptive process. While innocent suspects might hope that interrogators will remain open-minded enough to hear them out, once an interrogation begins, the focus for investigators becomes narrowed to the point that investigators have a set of beliefs that they are committed to and now they just need to obtain a confession. What happens when Risk 2, which often involves waiving the right to counsel, and the presumption of guilt of Risk 3 combine? Kassin refers to a three-step chain of events: “a perceiver forms a belief about a target person; the perceiver behaves toward the target in a manner that conforms to that belief; and the target responds in turn, often behaving in ways that support the perceiver’s belief.”⁸¹ Thus the interrogation itself amounts to a process that confirms guilt in the interrogator’s mind. And again, because innocent suspects tend to waive their rights to an attorney, these interrogations tend to occur without the presence of counsel.

A laboratory study can tell us how the interrogation process might be different, depending on whether the interrogator approaches the procedure presuming guilt or innocence. Kassin performed such a study in which he could control relevant variables, such as the type of crime, and actual innocence or guilt. In this two-phase study, Kassin and colleagues instructed mostly white American college students to steal $100 (other students were placed in a similar situation but were instructed not to steal the $100).⁸² After the theft, the thiefs and the non-thiefs were interviewed. Other students served as investigators and were divided into two groups: those who were told that most suspects were innocent and those who were told that most suspects were guilty. The “investigators,” girded with the assumption of either guilt or innocence, were instructed to interrogate one of the suspects. Did the presumption of guilt or innocence affect how the investigators behaved toward the suspects? Yes. Investigators who were led to expect guilt rather than innocence asked more guilt-presumptive questions (e.g., “How did you find the key that was hidden behind the computer?”) and exerted more pressure (e.g., repeated accusations) to get a confession. In the second phase of the experiment, audio recordings of the interrogations were played to independent observers during which they judged the presumed guilt or innocence of the suspects they heard being interrogated. According to the neutral observers, the suspects judged to be the most anxious, defensive, and guilty, were the suspects who were actually innocent (as opposed to those who stole the $100) and who were paired with guilt-presumptive (as opposed to innocence-presumptive) investigators. According to observers, the suspects who told the most plausible denial stories, yet brought out the worst in the interrogators, were innocent suspects. Thus, the most pressure-filled sessions occurred when innocent suspects were paired with guilt-presuming investigators. Again, the police interrogation is a guilt-presumptive process that sets into motion a set of assumptions and biases on the part of investigators when hearing an innocent person resisting guilt and telling a plausible story of denial, instead of carefully weighing the information and adjusting their judgments of the suspect, denials seem to make interrogators work doubly hard at obtaining a confession.

Certain interrogation techniques, such as isolation and the presentation of false evidence, increase the fourth risk to innocent suspects, the risk of a false confession. But surely, truly innocent people would never confess to a crime they didn’t commit, right? People simply do not believe someone would make a false confession. Many believe that under no circumstances would they give a false confession. If they didn’t do it, they didn’t do it. But another study performed by Saul Kassin shows that getting people to make false confessions involves a lot less pressure than torture. Kassin and Katherine Kiechel asked mostly white American college students to type letters on a keyboard in what was supposedly a study on reaction time – participants had to keep up with typing a fast-paced reading of information on a keyboard.⁸³ At one point, participants were accused of causing the experimenter’s computer to crash by pressing on the keyboard the ALT key they had been instructed to avoid. Subsequently, the students were asked to sign a confession. All participants were actually innocent, and all “suspects” initially (correctly) denied that they pressed the ALT key. In some sessions, a planted accomplice of the experiment told the experimenter that they had witnessed the participant hit the forbidden key; in others they said they had not seen what happened. Thus, in some versions of the procedure there was an “eyewitness.” This false eyewitness evidence significantly increased the number of students who eventually agreed to sign a written confession, from 65% (when there was no witness) to 100% (when there was a witness). The presentation of this false evidence also increased the number of participants who eventually believed they were truly responsible for this outcome to 100%. That is, while no one initially admitted to hitting the dreaded key, after some questioning by the experimenter and the presentation of a witness, everyone signed the confession. Furthermore, when the participants were asked how the accident happened, 35% made up details consistent with their confession that they pressed the ALT key. That is, they confabulated details that were in line with their (false) confessions. For instance, they stated, “I hit it with the side of my right hand after you called out the ‘A’.” Again, all participants were actually innocent.

While the Kassin and Kiechel study was a laboratory experiment, whose participants were college students and the study did not mimic the high stakes and extreme stress involved in being falsely accused of an actual crime, it does illustrate how susceptible one can be to influence. In Kassin and Kiechel’s study, there really was no motivation to confess, but with the presentation of (false) incriminating evidence in this minimally stressful situation (fast-paced key strokes), eventually all “suspects” did confess. Replications of this study have been conducted with similar results in conditions which include financial consequences,⁸⁴ and also with juveniles, who are especially vulnerable to making false confessions.⁸⁵

Three processes commonly used in interrogations and taught in interrogation training manuals are: isolation, confrontation, and minimization. These three processes can lead to false confessions. Prolonged periods of isolation, coupled with fatigue and sleep deprivation, can heighten susceptibility to being influenced and can impair one’s ability to make complex decisions. Interrogations producing confessions that are later found to be false tend to be longer than other interrogations. As Kassin reports, typical interrogations last three to four hours, but interrogations that produce confessions that are later found to be false averaged 16 hours. Once a suspect is isolated, interrogators can confront suspects with false DNA evidence supposedly linking them to the crime, they can present phony eyewitnesses, and they can claim that the suspect failed a polygraph. All of these techniques are legal and admissible in US courts.⁸⁶ As we saw in Kassin and Kiechel’s “ALT key” study, false evidence greatly increases the risk that innocent people will confess to acts they did not commit. The third step in interrogations is a process called minimization. To ensure a confession, the interrogator uses minimizing techniques such as providing a moral justification for committing the crime such as, “I’m sure she had it coming,” face-saving excuses, explaining that the crime was accidental, or saying that the suspect was provoked or impaired by drugs. Minimization techniques can also include promises of leniency by the court. All along, the interrogator implies that as soon as the suspect confesses, they can go home.

Again, police, judges, jurors, and lay people have extreme difficulty believing anyone would confess to a crime they hadn’t committed. The fifth risk to innocent suspects then, according to Kassin, is the resistance to believing that people would actually confess to a crime they did not commit. It has already been established that law enforcement officers are no better (and are, in some cases, worse) than college students at differentiating true confessions from false.

For juries, confessions carry more weight than eyewitness testimony. People do not fully discount a confession, even when it is logically and legally appropriate to do so. The belief that suspects are guilty of something is so powerful that prosecutors sometimes continue to deny innocence even after DNA tests unequivocally clear the wrongfully convicted person. Their gut tells them the person is still guilty; or the shame and embarrassment from admitting they were wrong prevents them from accepting the truth. Kassin states three reasons why innocence is not often detectable. First, people tend to take “facts” at face value. If people confess, we believe them. People also tend to discount the extent to which situational factors (e.g., a stressful interrogation) impacts behavior. Second, as we already established, people are not very good at deception detection. Third, police-induced false confessions often contain content cues, such as vivid details, that people associate with truth-telling. Kassin summarizes his five points this way:

Kassin’s studies were conducted with white participants and white inmates. From what we have learned about race and the criminal legal system and the sweeping criminalization of black people in particular, this process for a person of color is even more fraught.

One reason people confess to crimes they have not committed is because police can and do lie to suspects about the evidence they have about the case.⁸⁸ The general public likely does not realize that it is legal for interrogators to falsely tell a suspect that their fingerprints were found at the crime scene, or that their friend and co-defendant confessed. Interrogators use these lies to make suspects feel hopeless with the only way getting away from their hopeless condition is to confess. This is what happened in the infamous “Central Park Five” case when interrogators got five black and Latine teenagers to confess to a crime none of them committed. Years later, after serving long prison sentences, they were finally exonerated when a DNA match was found for the actual perpetrator. An ancillary to falsely confessing to a crime not committed is the act of pleading guilty to a crime not committed. While it may seem unimaginable, if a suspect is faced with overwhelming evidence, or, like in the Tulia case, sees peers convicted on similar evidence, or faces draconian minimum sentences, suspects might plead guilty to a crime they did not commit in order to avoid the risk of going to trial and facing a decades-long prison term.

How do people view exonerees? Do they accept their innocence? Simon Howard conducted a study on people’s reaction to those who falsely confessed to a crime and who were later exonerated.⁸⁹ American participants (mostly white non-students) read a fictional newspaper article that featured a story about a man who was exonerated by DNA evidence after serving 10 years in prison for a murder he had falsely confessed to committing. The article was formatted like other articles on exonerations. It included an interview with the exoneree and his mother. To manipulate the race of the exoneree, the article included a headshot and the name (DeShawn Cooper or Daniel Cooper) in order to convey either a black or white man. Respondents were then asked various questions about the exoneree. Interestingly, participants were no more likely to believe the man actually committed the crime if he was white or black. That is, respondents for the most part accepted the man’s innocence and did so regardless of whether they believed he was a black or white man. However, there were some racial patterns detected. For instance, the black exoneree was perceived as more aggressive than a white exoneree. Participants believed that a black exoneree was less deserving of governmental assistance (i.e., job training, psychological counseling, and subsidized housing) than a white exoneree. And they believed the black exoneree was more likely to end up back in prison for committing a crime post-exoneration than the white exoneree.⁹⁰ The different perceptions of the black and white men are stunning given that their circumstances conveyed in the newspaper article was identical except for their names and photos. Howard’s study suggests that those in charge of deciding people’s fate in criminal proceedings are not able to set aside their racist schemas. These findings exemplify the forgiving nature of attitudes toward white individuals and the lingering suspicion about African Americans. We can see some of the implications in the next section.

Bias in the Courtroom: How Jurors Perceive Defendants and Interpret Evidence

This section deals with sociocognitive processes that are relevant after a criminal defendant has been arrested and charged. It addresses how juries interpret information and how their interpretations influence the decisions they make. In a review of the research on the effects of race and ethnicity on jury behavior, Jennifer S. Hunt documents several patterns.⁹¹ Defendant race and ethnicity influences verdicts and sentencing decisions made by juries. In experimental research using mock jurors, jurors tend to make more favorable judgments for ingroup (e.g., same-race) defendants and make harsher judgments against outgroup defendants. This pattern is known as the similarity-leniency effect and seems to be particularly strong for white jurors making judgments about white (vs. Latine and black) defendants. Echoing the previous chapter, juror bias can be found in the attributions they make about defendant behavior. They make more internal attributions when explaining the behavior of an outgroup member and more external attributions when the defendant is an ingroup member. These attributions suggest that the criminal behavior of an outgroup member is believed to be due to stable personality characteristics, but the criminal behavior of an ingroup member is due to situational, temporary, factors.

Bias in Convictions and Sentencing

The United States incarcerates a larger proportion of its citizens than any other country in the world,⁹⁷ and incarceration rates have been supercharged for decades starting in the 1990s. For example, in 2021, the nation incarcerated five times more people per capita than it did 40 years prior.⁹⁸ But, who is being incarcerated? According to Jeffrey Reiman and Paul Leighton, “For the same criminal behavior, the poor are more likely to be arrested; if arrested, they are more likely to be charged; if charged, more likely to be convicted; if convicted, more likely to be sentenced to prison; and if sentenced, more likely to be given longer prison terms than members of the middle and upper classes.”⁹⁹ And, according to a review of the literature by Reiman and Leighton, the face of the criminal legal system is not only poor but also black.¹⁰⁰ African Americans are incarcerated in state prisons at nearly five times the rate of white Americans. In 2019, the rate of incarceration of black women was double that of white women; for black men, it was nearly six times more than for white men.¹⁰¹ Latine people are incarcerated at a rate of 1.3 times the rate for white people.¹⁰²

Once again, controlled experiments allow us to expose double-standards and benefits-of-the-doubt that make up subtle forms of prejudice related to the criminal legal system. Experiments have found, for example, when a suspect was black and was accused of committing a violent crime, they were more likely to be misidentified and also harshly penalized than a white suspect accused of similarly egregious crimes. This happened for black respondents as well as for other minority ethnicities.¹⁰³ When another experiment compared the treatment of an undocumented immigrant and non-immigrant accused of a crime, it found that the suggested punishment was harsher for the undocumented immigrant.¹⁰⁴

One experiment looked at one proxy of race and ethnicity – accents – in a civil case involving a car accident. Mostly white college students served as mock jurors who reviewed a two-car accident case to determine guilt, technically negligence in this case.¹⁰⁵ The mock jurors heard the audio recording of a black, Latine, or white defendant. In each of the three versions of the audio recording, the speaker spoke with a stereotypical accent associated with his race, or spoke without an accent. So, in total, there were six versions of the audio report of the accident. In all versions, the plaintiff (i.e., the alleged victim) was white. Mock jurors evaluated various aspects of the case. Several findings had nothing to do with accents, but did have to do with racial/ethnic bias. For example, regardless of accent, white defendants were viewed as least negligent compared to the black and Latine men. Mock jurors awarded more to plaintiffs if the defendant was black, compared to a white or Latine defendant. Jurors awarded significantly higher amounts in damages to the white plaintiff when a black defendant, as compared to a white or Latine defendant, was found negligent. Jurors also saw the black defendant more negatively, in general, and perceived the black defendant’s actions as more avoidable than those of a white defendant. There were also several findings associated with bias toward defendant accents. For example, black defendants with stereotypical accents were most likely to be found negligent by the generally white participants. White defendants, regardless of accent, were least likely to be found negligent. Accented defendants were also believed to have a lower socioeconomic status than unaccented defendants, particularly Latine defendants. Accented defendants were seen as more prone to automobile accidents, particularly the black accented defendant.

In addition, mock jurors with more conservative social beliefs were more likely to find an accented defendant negligent regardless of race, and participants with more liberal social beliefs were less likely to find the defendant negligent across all conditions, particularly for the Latine defendant. Women jurors were more likely than men to find the black defendant negligent. On the other hand, respondents had less confidence in their verdicts when the defendant was black compared to when the defendant was white. Respondents also perceived the police officer as less truthful when testifying against an accented defendant, across all race conditions. Thus, even though respondents were more likely to find a black defendant with a stereotypical accent negligent, and then award significantly more damages to the white plaintiff, the participants expressed some reluctance in their negligence judgment and skepticism in trusting the police officer’s testimony.

Of course, convictions and sentence disparities occur outside the laboratory – in real life. For instance, it has long been known that in the United States the typical sentence for the possession or sale of crack cocaine was 100 times longer than the typical sentence for the possession or sale of powder cocaine.¹⁰⁶ Pharmacologically identical, crack cocaine is manufactured from powder cocaine. Crack is made from a mixture of powder cocaine, baking soda, and water that is cooked and cooled until a solid “rock” is formed that can be broken and sold in small quantities. It was widely understood that differences in the penalties associated with rock and powder cocaine exist because white and middle- and upper-class users and sellers deal with powder, while non-white and lower-class users deal with crack.¹⁰⁷ It wasn’t until 2022 when the US attorney general directed federal prosecutors to eliminate the disparity.¹⁰⁸

A review of the literature by The Sentencing Project also found that, in the United States, young black and Latine men receive harsher sentences for the same crime than similarly situated white men.¹⁰⁹ Differences exist in terms of employment status. Unemployed black men are sentenced more harshly than unemployed white men. African Americans even pay a higher trial penalty than similarly situated whites. A trial penalty occurs when defendants are given harsher sentences for going to trial and being found guilty rather than pleading guilty, thereby avoiding a trial. The trial penalty may be a specific punishment for African Americans who might be perceived by the judge as “uppity” and resisting rather than admitting guilt and accepting their punishment.

One review of sentences of federal crimes in the United States found patterns consistent with benign bigotry. When sentences for similar crimes were compared, black offenders were sentenced similarly to white offenders. However, when criminal history was taken into account, an interesting pattern emerged. For those offenders with little criminal history, blacks and Latines received longer sentences than whites, but for offenders with extensive criminal histories, whites received longer sentences.¹¹⁰ Maybe judges are on guard with those who have long histories to not be racist, but then let themselves punish people who are not big-time criminals in biased ways. In federal sentencing, one study found that Native Americans received the longest sentences relative to whites, blacks, Latines, and Asians. Like the preceding finding, the disparity was most dramatic for those offenders with lower levels of criminal history.¹¹¹

Samuel Sommers and Phoebe Ellsworth used an aversive racism framework to examine whether there is bias against black defendants when race is made salient versus when race is not made salient.¹¹² They conducted an experiment in which white mock jurors rated various aspects of a defendant’s case. Respondents were presented with a trial summary of an interracial battery case. One version had a white defendant and black victim, the other version had a black defendant and a white victim. Half of the versions were “race-salient” (a racial slur was made toward the defendant, either white or black, during an earlier altercation) and half just specified the race of both the defendant and the victim. The defendant and victim were part of the same high school basketball team and the defendant was accused of attacking the victim. One might predict that when race was made obvious (in the race-salient condition) there would be anti-black bias among white jurors. But this prediction comes from an old-fashioned racism perspective – whites feeling fairly comfortable showing overt bias against African Americans. A prediction based on the assumptions of aversive racism and benign bigotry, would suggest that if race were made salient, whites would not feel comfortable showing bias against African Americans because they would be afraid that they would appear racist. If race was merely implied, however, but was not an overtly significant aspect of the study, white respondents might feel that they could appear as though they were responding to the conditions of the crime, not to the race of the defendant or victim.

Study results were consistent with Sommer and Ellsworth’s assumptions. White jurors demonstrated a significantly higher conviction rate for the black defendant in the non-race-salient condition than in the race-salient condition. Also, white jurors were more likely to convict the black defendant than the white defendant; the case against the African American defendant was judged to be stronger than the case against the white defendant; respondents rated the white man’s defense as stronger than the black man’s; and jurors recommended a more severe sentence for the African American than the white. In contrast, when race was made salient, the conviction rates for whites and blacks were comparable.¹¹³ Sommers and Ellsworth conclude that the assumption that race-salient trials are most likely to elicit white juror prejudice is incorrect. Instead, when race is salient, white jurors tend to make an effort to appear unbiased toward black defendants. When race is only an underlying factor, white jurors are likely to demonstrate bias.

Decisions to make race salient or not in a trial matters because stakes are high for people of color. In the United States, African Americans are about 7.5 times more likely to be wrongfully convicted of murder than are white defendants, and about 80% more likely to be innocent than others convicted of murder, according to research by the National Registry of Exonerations.¹¹⁴ The already disproportionate risk of wrongful conviction, the registry found, was significantly worse if the murder victim in the case was white. Similarly, the report Race and Wrongful Convictions in the United States 2022 reviewed the cases of 3,200 innocent defendants exonerated in the United States since 1989. African Americans, the report found, were seven times more likely to be wrongfully convicted, were more likely to be the victims of police misconduct, and more likely to be imprisoned longer before being exonerated.¹¹⁵

Capital Punishment and Death Qualification

Wrongful convictions cannot be reversed if someone has been executed. Most countries in the world have abolished the death penalty and executions are declining even where it is legal. The United States is one of the few countries left in the world that regularly executes its citizens.¹¹⁶ By 2022, just over half of the US states, 27, execute people.¹¹⁷ Support for the death penalty has decreased over several decades.¹¹⁸ In the United States, executions have decreased.¹¹⁹ Still, in 2021, about 60% of Americans supported the death penalty, although 78% acknowledge the risk of executing innocent people.

Proponents of the death penalty argue that capital punishment serves as a deterrent to criminal activity. Obviously, the person executed will not commit future crimes, but does capital punishment deter others from killing? Studies consistently conclude that the death penalty does not affect the rate of violent crimes. Proponents of the death penalty also maintain that murderers are such dangerous people that allowing them to live increases the risk of injury or death to other inmates or correctional officers. However, systematically performed studies find that capital murderers (those who have been convicted and are on death row) are no more violent than life-without-parole and parole-eligible inmates.¹²⁰

The death penalty is administered only in a minority of eligible cases, and its determinants are inconsistent and arbitrary. One factor accounting for the inconsistency regarding who is actually executed, is race – both the race of the victim and the race of the perpetrator. Jurors make harsher judgments of defendants of color when a victim is white versus another person of color.¹²¹ According to the Death Penalty Information Center, more than 75% of murder victims in cases resulting in an execution are white, even though nationally only 50% of murder victims are white.¹²² When considering persons executed for interracial murders, 94% of those were made up of African Americans executed for killing a white person, and only 6% for whites who killed African Americans.¹²³ Jurors are more likely to give the death penalty in cases in which a Latine defendant has been convicted of killing a white person.¹²⁴ These patterns hold even when various factors related to the crime and the defendant are controlled for. As with other verdicts, ethnic minority defendants are more likely to receive the death penalty when juries are made up of a higher proportion of white people.¹²⁵ These gross discrepancies reveal much more about whose lives are valued and whose lives are not than which crimes should be punishable by death.

Reducing Bias during Investigations

Jury Instructions

Changes in Policy and Law