第1章
The Hidden Power of Bias: How Our Unconscious Mind Shapes Society
When I walked into that auditorium filled with stone-faced Oakland police officers, I never expected my five-year-old son's innocent comment would break through their defenses. Earlier that week, on a flight, my son had pointed to the only Black passenger saying, "That guy looks like Daddy," though the man looked nothing like my husband. Before I could correct him, he added, "I hope that man doesn't rob the plane." When I asked why he said that, his solemn reply broke my heart: "I don't know why I was thinking that."
This story transformed the room of officers. Their expressions softened as we connected as parents confronting how racial stratification infiltrates even innocent minds. Jennifer Eberhardt's groundbreaking work on implicit bias has earned her a MacArthur "genius" grant and recognition as one of Foreign Policy's 100 Leading Global Thinkers. Her research has revolutionized how major institutions from police departments to tech companies address discrimination. Through compelling science and personal narrative, she reveals how bias operates in our brains, our neighborhoods, and our cultural understanding, while offering practical insights for recognizing and addressing these hidden prejudices.
第2章
When Faces Blur: The Neuroscience of Recognition
My first twelve years in Cleveland's all-black Lee-Harvard neighborhood left me unprepared for my family's move to nearly all-white Beachwood. Despite my new classmates' warm welcome, I struggled with a problem I couldn't articulate: I couldn't tell them apart. Their faces blended together despite my attempts to focus on distinctive features. This inability made me awkward and withdrawn, afraid of embarrassing myself by calling someone by the wrong name.
Years later, I discovered this wasn't just my problem. Scientists call it the "other-race effect" - people are better at recognizing faces of their own race than others. This universal phenomenon appears by three months of age and strengthens through adolescence. It's not bigotry but biology and exposure; our brains become efficient at processing familiar faces. Children adopted by parents of another race don't show this effect, demonstrating how experience shapes perception.
What took me decades to learn about race and face recognition proved to be street wisdom for young criminals in Oakland. While analyzing racial disparities in policing, I learned about black teenagers targeting middle-aged Asian women for purse-snatching in Chinatown. Despite arrests, cases collapsed because victims couldn't identify suspects in lineups. The robbers explicitly targeted Asian women because "Asian people can't ID. They just can't tell brothers apart." Police data confirmed this - nearly 80% of Asian victims couldn't identify their black assailants, while black women could identify black suspects at much higher rates after just a glance.
In our groundbreaking neuroimaging study, we found the fusiform face area (FFA) - a brain region critical for distinguishing faces - responded more vigorously to same-race faces for both black and white participants. The stronger the FFA response, the more likely participants could later recognize that face. This provided the first neural evidence for the same-race advantage in face recognition, showing how race powerfully influences even our most basic brain functions.
第3章
Categories That Blind: How We Sort People
The Asian women targeted in Oakland became victims not just of theft but of racial categorization itself. To the robbers, they represented a vulnerable category; to the women, the thieves became a threatening category of "young black men." Each frightening encounter reinforced the toxic stereotype that black men are dangerous.
Categorization isn't an abhorrent feature of some minds but a universal brain function that helps us organize overwhelming stimuli. It brings coherence to chaos and enables quicker judgments. But it also impedes our understanding of people deemed unlike us, dampening our sensitivity to their individuality.
Our racial categories determine what we actually see in everyday settings. When my African American friend Marsha stood near her fair-skinned sister who passed as white, no one recognized their obvious physical resemblances. Once we decide on a category, our perception adjusts to match the label. Research confirms this: Latino participants better remembered racially ambiguous faces when shown with Latino hairstyles than the same faces with African American hairstyles.
This categorization affects neural processing - white participants exhibit less brain activity in face-processing regions when viewing black faces than white faces. Their brains process black faces categorically, as though all black faces were the same stimulus, despite being shown different individuals. This "repetition suppression" effect suggests we reserve cognitive resources primarily for those we consider "like us."
Even preschoolers quickly absorb adults' social evaluations. In a University of Washington study, after watching a brief video where an adult treated one person warmly and another coldly, 75% of preschoolers preferred the person who received positive treatment. They concluded that people treated badly must be bad people, holding the targets of bias responsible rather than the bias-holders.
When my six-year-old son asked if people see Black people differently from white people, he described witnessing shoppers creating an "invisible force field" around a Black man in our predominantly white neighborhood grocery store. After deep thought, he concluded with startling insight: "I think it's fear." This conversation revealed how adept children are at detecting social signals and patterns, watching adults to understand social hierarchies and interpersonal evaluations.
第4章
The Deadly Consequences of Bias in Policing
For fifteen years, I've trained police officers to recognize implicit bias in their work, part of a movement thrust into prominence by controversial police shootings and subsequent protests. In September 2016, while preparing for a statewide training program in California, news broke about Terence Crutcher, an unarmed Black man shot by police in Tulsa.
The video showed Crutcher walking slowly with raised hands before being shot. This was just the latest in a string of high-profile deaths including Philando Castile and twelve-year-old Tamir Rice. Nearly a thousand Americans were killed by police in 2016 alone.
These incidents have left mothers in tears at my community talks, desperately asking how to keep their sons safe. The old advice-be respectful, keep hands visible, don't run-no longer seems sufficient protection. Though over 99% of police encounters involve no force, Black people are disproportionately stopped and more likely to experience force.
Watching Terence's video and hearing his twin sister Tiffany's story crystallized years of implicit bias research for me. Each critical moment-from Officer Shelby's initial approach to Terence lying untended-corresponded to scientific studies on bias. While we can't know exactly what went through Officer Shelby's or Terence's minds, science allows us to examine larger patterns by controlling variables and measuring responses.
My research with colleagues on attentional bias directly relates to questions about what drew Officer Shelby to approach Crutcher. In studies with police officers, we used subliminal priming techniques to measure how crime-related words affected visual attention. Officers exposed to crime-related words (like "apprehend" or "shoot") unconsciously paid more attention to Black faces than White faces, while those in the control condition showed the opposite pattern. We found the same results with Stanford undergraduates, demonstrating how stereotypic associations drive attention regardless of profession.
Police characterized Terence as a giant, estimating his weight at 300 pounds when he actually stood 5'9" and weighed 255 pounds. Research shows that people consistently perceive Black men as taller, heavier, and stronger than White men of identical size. Even when shown racially ambiguous body images, participants rated the same bodies as larger when told they belonged to Black men.
Research using "shoot-don't shoot" simulations reveals how racial bias affects split-second decisions. Participants were quicker to "shoot" Black individuals holding guns than whites with guns, and more likely to mistakenly shoot unarmed Black people. Police officers showed similar bias in reaction time but were equally accurate in their shooting decisions regardless of race-a difference attributed to specialized training.
第5章
When Police Become Prisoners of Perception
While data collection efforts on racial disparities in policing give communities a voice, they often lead to opposing interpretations. Police see high crime statistics among blacks (83% of violent crime in Oakland) as validating their focus, while communities view stop disparities as evidence of racial profiling. The constant radio dispatches of "MALE BLACK" hundreds of times daily creates an automatic association between blackness and criminality for officers.
A young white officer who grew up in Germany shared how this environment transformed him - even off duty, he began scrutinizing every black man he saw, with innocent gestures triggering an avalanche of precautions and mental scenarios of danger, concerning his non-police friends.
Procedural justice training helps officers reconnect with their original mission to protect and serve. Rather than focusing on tactics or maximizing stops, it emphasizes building healthy relationships with the public through four key tenets: voice (allowing people to tell their story), fairness (applying law impartially), respect (treating people with dignity), and trustworthiness. Research shows people care as much about how they're treated during police encounters as the outcomes themselves.
This approach helps counter the cynicism officers develop over time-the jaded perspective that comes from witnessing violence, feeling unappreciated, and operating in constant hypervigilance. This cynicism narrows officers' vision, causing them to focus on the 3% engaged in violent crime while the other 97% of community members "go dim." Like the famous "invisible gorilla" experiment where viewers focused on counting basketball passes miss a person in a gorilla suit walking through the scene, officers' selective attention can miss the majority of law-abiding citizens in high-crime neighborhoods.
第6章
The Justice System's Bias Pipeline
On the eve of my Harvard PhD graduation in 1993, after years of feeling out of place as a first-generation college student, I experienced firsthand the reality of biased policing. Despite being selected to carry the Graduate School flag at commencement, that day I was simply exhausted after catering an event with my friend April. As we drove to her Boston housing project apartment, police stopped us without explanation. When questioned about my mother's car registration (six weeks expired), I couldn't provide her Social Security number. The officer ordered a tow truck and demanded we exit the vehicle. When I refused, five police cruisers arrived as residents gathered to watch. An officer yanked me from the car and body-slammed me onto its roof, leaving me gasping for air before handcuffing me.
Equipment-related traffic stops give officers wide discretion to investigate drivers based on hunches, which can activate racial biases. Research analyzing 18.5 million traffic stops found Black drivers are twice as likely as white drivers to be stopped for equipment violations rather than moving violations. In some jurisdictions, these stops function as an unofficial tax on Black and low-income people. A Department of Justice investigation of Ferguson revealed officers were instructed to generate municipal revenue through tickets, disproportionately targeting Black residents. Though Black drivers were twice as likely to be searched, they were 26% less likely than whites to possess contraband.
When arrested in Boston, I was handcuffed to a wall but fortunately had Harvard Dean Margot Gill's number. Unlike most arrestees, we were released without bail because someone important vouched for us. Each year, over eleven million people are jailed locally, with three-quarters held for nonviolent offenses while awaiting trial because they can't afford bail. Black defendants face pretrial detention rates four times higher than whites with similar charges and pay 35% higher bail premiums.
The death penalty represents America's ultimate punishment, reserved for those deemed "too evil to live." While thirty-one states still permit capital punishment, the United States stands nearly alone among industrialized nations in executing criminals. Research consistently shows disturbing racial disparities in death sentencing - murderers of white victims are significantly more likely to receive death sentences than those who kill Black people, even when controlling for non-racial factors.
My research revealed that physical appearance itself can determine who lives and who dies. We analyzed photographs of Black men convicted of capital crimes in Philadelphia between 1979-1999, having participants rate how "stereotypically Black" each defendant appeared. For Black defendants who killed Black victims, these ratings had no impact on sentencing. However, when Black defendants were convicted of killing white victims, those with more "stereotypically Black" features were more than twice as likely to receive death sentences (57% versus 24%), even after controlling for crime severity, aggravating circumstances, socioeconomic status, and attractiveness.
第7章
The Dehumanization Machine: Science and Stereotype
My research into the history of race and neuroscience revealed the disturbing foundations of scientific racism. Though Darwin's theory of evolution revolutionized science by establishing monogenism (all humans are one species) and locating humanity's origins in Africa, the belief in Black inferiority remained entrenched. Scientists simply reframed their narrative: instead of viewing Blacks as a separate, inferior species, whites became seen as "the most complex, most intelligent, most evolved humans."
Throughout the early 20th century, psychologists replaced crude skull measurements with IQ tests to institutionalize bias. By 1910, these tests were used to quantify supposed mental deficiencies in Blacks and immigrants. At Ellis Island, Howard Knox's wooden jigsaw puzzle became a "mental measuring scale" to identify "feebleminded" immigrants, supporting the eugenics movement. The Immigration Act of 1924 used these test results to drastically reduce immigration from "undesirable" southern and eastern European countries.
Working with colleagues like Phillip Goff, I discovered the Black-ape association remains powerfully embedded in modern consciousness. Our studies revealed that subliminally exposing people to ape images directed their attention to Black faces, even more strongly than crime-related images did.
This association operates through selective attention - we notice what aligns with existing mental frameworks. When we primed participants with stereotypically Black names before showing them the famous "invisible gorilla" video, they were 25% more likely to notice the gorilla than those primed with white names - jumping from 45% to 70% detection rates.
This dehumanization extends beyond laboratories into law enforcement. LAPD officers' computer chats after the Rodney King beating referred to Black people as "monkeys" and used the code "NHI" (No Humans Involved). As recently as 2016, San Francisco police exchanged texts describing minorities as "wild animals" and "cockroaches."
Our research showed that people subliminally primed with ape-related words were more likely to justify police brutality when they believed the victim was Black. Most disturbing was how readily my scientific colleagues accepted these findings, with many suggesting the association existed simply because "Blacks actually look more like apes" - revealing how deeply ingrained this dehumanization remains even among educated professionals.
第8章
Technology's Double Edge: Amplifying and Fighting Bias
Migration mirrors journeys still made by ethnic groups worldwide - more than just family mobility, it reflects how space has been used to subjugate people. Residential restrictions diminish potential, while physical relocation can provide access to social mobility and blunt bias's impact. Yet northern cities weren't the refuge migrants imagined. My great-cousin Jean found segregation in Chicago more severe than in the South, with real estate agents stoking white fears about property values as Black families moved in, causing white flight.
In today's digital age, bias has found new expression through technology - neighborhood surveillance systems and social networks can foster tunnel vision that distorts danger perception and heightens suspicion, allowing residents to circulate photos of "suspicious" strangers without evidence of wrongdoing, while simultaneously documenting disturbing displays of bias that might have previously remained invisible.
Nextdoor, a neighborhood social network connecting millions worldwide, faced a crisis when its "crime and safety" category became filled with racially biased posts labeling Blacks and Latinos as "suspicious" for ordinary activities. Though such posts were statistically rare, founder Sarah Leary recognized even one was problematic. Her team studied bias research and discovered that fear and rapid decision-making create perfect conditions for stereotyping. While technology typically prizes frictionless speed, this very quality makes platforms vulnerable to bias expression. Nextdoor implemented a checklist requiring users to focus on specific behaviors rather than vague racial descriptions before posting about "suspicious people." This deliberate friction slowed users down but reduced racial profiling by 75% without significantly decreasing engagement.
While Nextdoor connects neighbors, Airbnb connects travelers with hosts worldwide, promising belonging in every community. Research confirmed racial disparities in bookings across all cities, property types, and host demographics - even Black hosts discriminated against Black guests. After initially responding slowly, Airbnb's leadership eventually recognized discrimination as "the greatest challenge we face as a company," and brought in civil rights attorney Laura Murphy to address it. Murphy helped implement solutions including a mandatory "community commitment" against discrimination, promoting the "instant book" option that eliminated approval bias, and providing previous guest reviews to hosts to replace stereotypes with concrete information.
第9章
Education's Unfulfilled Promise: Integration and Inequality
In 1957, six-year-old Bernice began her education in a two-room Mississippi schoolhouse with no running water, using discarded textbooks from white students. Despite school segregation being ruled unconstitutional in 1954, it took until 1967 for her to attend an integrated school through Mississippi's "choice" program. As one of only four Black girls at all-white Olive Branch High, Bernice faced hostility from students and teachers alike. In algebra class, her teacher isolated her in the front row where no white student could sit near her, teaching past her as if she didn't exist. Despite the social trauma, Bernice valued the resources unavailable at her old school - speech therapy, a well-stocked library, and higher academic standards.
School integration advocates in the 1950s and 60s believed bringing children of different races together would naturally improve race relations by replacing stereotypes with personal connections. This optimism was partly justified - Black students from integrated schools earn 25% more as adults, enjoy better health, and people of all races from diverse schools have more cross-racial friendships and civic engagement. However, psychologist Gordon Allport's 1954 work showed that contact only reduces prejudice under specific conditions: equal status interactions, authority support, and personal rather than superficial connections. When students like Bernice were stigmatized and separated, contact actually reinforced inequality.
While school integration initially appeared successful in the 70s and 80s, it has reversed course over the last twenty years as legal rulings eliminated desegregation tools and charter schools created more racially homogeneous environments. Today, intensely segregated schools have tripled, with black students increasingly attending racially isolated institutions. Even in integrated settings, black and Latino students face an achievement gap linked to both social class disparities and psychological factors. Research shows that students need to feel valued and respected to succeed academically.
Despite evidence supporting empathy, wise feedback, and affirmation as effective interventions, schools commonly embrace "colorblindness" - the misguided notion that ignoring race prevents bias. This approach fails scientifically and practically. By age ten, children avoid discussing race even when relevant, having absorbed adult discomfort with the subject. Research shows colorblindness actually impedes equality: when children were exposed to either colorblind messaging ("race is not important") or diversity-valuing messaging ("racial differences make each of us special"), those primed with colorblindness were significantly less likely to identify racial discrimination in scenarios.
第10章
Breaking the Bias Cycle: Practical Solutions for Change
Understanding bias is only the first step - the real challenge is keeping it in check. Speed and ambiguity are two powerful triggers for bias. When hiring managers spend only six seconds reviewing each resume, they often resort to hunches and familiar patterns, allowing unconscious stereotypes about names to raise barriers for minorities.
To counter this, candidates can provide concrete metrics about their accomplishments rather than vague descriptions. Instead of saying "grew the client base," specify "produced a 50 percent increase in sales over six months." Such objective standards help neutralize subjective stereotypes. Similarly, recruiters can minimize bias by using structured interviews with standardized questions for all applicants.
Bias also flourishes when decisions remain unmonitored, with no checks and balances on our spontaneous choices. This was demonstrated in baseball, where umpires showed racial bias in close calls - favoring pitchers of their own race - but this bias disappeared in stadiums with playback cameras monitoring their decisions. Even with monitoring, however, incentives matter: umpires working unpaid extra innings tend to make calls that end games faster, regardless of who benefits.
Beyond situational factors, personal connections can override implicit bias, especially relationships involving intimacy or mutual dependence. As Tanisha, a Black nurse in a Jewish convalescent home discovered, caring for vulnerable patients creates bonds that transcend racial differences. Even patients with dementia who use racial slurs come to associate blackness with competence and tenderness rather than stereotypes. These intense cross-racial relationships quickly undo associations built over time.
Institutional values and practices can also drive social change. Starbucks' response to the arrest of two Black men - closing stores for bias training - may not have eliminated bias but legitimized efforts to address subtle discrimination. Companies are now modifying job descriptions to remove gender-biased language and using technology to reduce bias in hiring. Orchestras using "curtain interviews" have significantly increased the number of women hired.
While diversity itself isn't a remedy for bias, it creates opportunities for change, though with growing pains. Diverse groups make better decisions but face more discord as privilege shifts and new voices emerge. Success requires tolerating this discomfort while changing the underlying cultures that perpetuate bias. Society is increasingly intolerant of explicit bigotry, as seen in the swift consequences for Papa John's founder and Roseanne Barr after their racist remarks. Implicit bias may be harder to recognize, but addressing it is possible - and believing in that possibility is crucial to progress.
