Abstract
Systemic racism is a type of racism: the pattern by which racial disadvantage is produced and reproduced through the ordinary workings of institutions, markets, and shared mental representations rather than through the conscious hostility of individuals. Cognitive psychology enters through the bias-of-crowds model, which recasts implicit bias as a property of situations rather than a stable personal trait, so that many weakly biased judgments aggregate into a durable structural level no single actor intends. This article distinguishes systemic from individual racism, traces the bridge from momentary association to structure, shows how small per-decision gaps compound across sequential gates, and reviews the evidence linking structural racism to health and the interventions it supports. A worked example and three interactive demonstrations make the compounding, aggregation, and feedback mechanisms concrete.
Keywords: systemic racism, structural racism, bias of crowds, implicit bias, cumulative disadvantage
- Systemic racism locates racial disadvantage in institutional rules, market structure, and shared representations — not primarily in individual animus — so disparate outcomes can persist with few or no overtly prejudiced actors.
- The bias-of-crowds model treats implicit bias as a feature of situations rather than a fixed personal trait: aggregate bias is stable and predicts regional disparity, while the same individual's score is noisy across occasions.
- Small per-decision gaps compound multiplicatively across sequential gatekeeping stages, so a disparity that is trivial and deniable at each step becomes large in aggregate.
- Structural racism is a well-evidenced determinant of population health, operating through residential segregation, differential institutional treatment, and chronic stress.
- Because the causes are structural, the effective levers are structural — changing defaults, criteria, and institutional practice — and brief individual bias training alone shows little durable effect on outcomes.
What Systemic Racism Is
Most everyday reasoning about racism treats it as a property of persons: a racist is someone who holds and acts on hostile beliefs about a racial group, and racism is the sum of such people's behavior. This individual-level picture is not wrong so much as incomplete, and the incompleteness matters. It cannot explain why racial disparities in wealth, housing, health, and incarceration persist, and in some domains widen, during periods when surveys record a steady decline in openly expressed prejudice and when few decision-makers in any given institution would endorse — or privately hold — a discriminatory intent.
Systemic racism is the answer to that puzzle. The sociologist Eduardo Bonilla-Silva reframed racism as a feature of a racialized social system: a society whose economic, political, and ideological structures allocate advantage along racial lines, so that discrimination is carried out by the routine operation of institutions rather than requiring a prejudiced operator at each point (Bonilla-Silva, 1997). Joe Feagin's parallel account treats systemic racism as a durable, self-reproducing property of the whole social order — embedded in the distribution of wealth, in institutional norms, and in a shared white racial frame that supplies the interpretive defaults through which situations are read (Feagin & Bennefield, 2014). On both accounts the unit of analysis shifts from the individual to the system, and the mechanism shifts from intention to structure.
This shift does not deny individual prejudice; it relocates it. Gordon Allport's foundational analysis had already distinguished personal prejudice from the institutional arrangements that sustain it, noting that discriminatory structures can outlive the attitudes that built them (Allport, 1954). What systemic accounts add is the claim that the structure, once in place, no longer needs those attitudes to continue. A mortgage-approval algorithm trained on historically segregated lending data, a school-funding formula tied to local property values, or a hiring pipeline that recruits through existing employees' networks will each reproduce racial disparity whether or not anyone involved harbors racial hostility. The disparity is a design feature of the process, not a residue of the operators' minds.
Cognitive psychology's distinctive contribution is to explain how the representational layer of this system — the shared associations, stereotypes, and defaults that Feagin's racial frame gestures at — is produced and maintained. That is the subject of the next section.
The Cognitive Bridge: Bias of Crowds
For two decades the dominant cognitive account of racial bias was the Implicit Association Test, which measures the relative speed of pairing racial categories with positive versus negative concepts (Greenwald, McGhee, & Schwartz, 1998). The IAT was widely interpreted as a window onto a stable, individual trait — a reservoir of implicit prejudice carried inside each person's head, causing their discriminatory behavior and, summed across people, producing societal disparity. Two findings strained that interpretation. An individual's IAT score is strikingly unstable across occasions, with test–retest reliability too low for a trait measure; yet the aggregate IAT score of a region or institution is highly stable over time and predicts that place's racial disparities.
Payne, Vuletich, and Lundberg resolved the paradox by inverting the causal story. Their bias-of-crowds model proposes that implicit bias is not a trait residing in individuals but a property of situations — the amount of bias a context makes cognitively accessible at a given moment (Payne, Vuletich, & Lundberg, 2017). On this view an IAT score is a reading of the environment taken through a person, not a reading of the person. It is noisy at the individual level precisely because it tracks a fluctuating situational input; it is stable at the aggregate level because the situation — the local concentration of segregation, inequality, and historical disadvantage — is stable. The individual measurement and the aggregate measurement are not the same quantity observed at two scales; they are different quantities, and only the aggregate is the structural signal.
This reframing is what makes implicit bias a bridge between the cognitive and the systemic rather than a competitor to the systemic account. The model's strongest empirical support comes from showing that regional aggregate implicit bias tracks structural historical variables: across U.S. counties, the level of implicit bias correlates with the proportion of the population enslaved in 1860, a structural fact over a century and a half old (Payne, Vuletich, & Brown-Iannuzzi, 2019). Bias, on this evidence, is downstream of structure — a cognitive echo of the local environment's history — not an upstream cause that structure merely amplifies. Consistent with that, aggregate implicit bias changes slowly and in response to contextual change, while the individual-level signal is dominated by occasion-to-occasion noise (Vuletich & Payne, 2019).
The cultural-psychological tradition reaches the same conclusion from a different direction. Salter, Adams, and Perez argue that racism is literally built into the structure of everyday worlds — the calendars, curricula, product offerings, and physical spaces a person moves through — so that these artifacts transmit and enforce racial meanings without any individual needing to hold a prejudiced belief (Salter, Adams, & Perez, 2018). Where the bias-of-crowds model locates the structural signal in the aggregate of measured associations, the cultural-psychological account locates it in the engineered environment itself; both deny that the relevant unit is the isolated biased individual. The RegionalBiasExplorer demonstration below shows how noisy individual scores aggregate into a stable regional level as the sample grows.
Regional Bias Explorer
In the bias-of-crowds model an individual test score reads a fluctuating situational input, so it is noisy; the aggregate across many people cancels that noise and tracks the region's stable structural level. Raise the crowd size and watch the mean lock onto the structural line even as the individual dots stay scattered.
How Systemic Racism Reproduces Itself
If the structure no longer needs prejudiced operators, what keeps it running? The answer is a set of self-reproducing mechanisms, each cognitively unremarkable in isolation. Table 1 summarizes the three and the structural lever each implies.
| Mechanism | Cognitive basis | Structural effect | Structural lever |
|---|---|---|---|
| Cumulative disadvantage | Small per-gate gaps, each individually deniable | Gaps compound multiplicatively across sequential gates into a large aggregate disparity | Reduce the number of discretionary gates; audit per-stage pass rates |
| Disambiguation | Ambiguous judgments resolved with whatever associations the situation supplies | Discrimination concentrates in borderline cases, where it is most deniable | Standardize criteria; remove race-correlated cues from the decision |
| Color-blind frame | A ready non-racial explanation for every disparity | The structure's outputs are reinterpreted as fair, removing pressure to change the gates | Make the structure visible as a structure; report outcomes by group |
The first is cumulative disadvantage across sequential gatekeeping. Most consequential outcomes — getting hired, approved for a loan, admitted to a program, released on bail — are reached through a chain of decisions, each a separate gate. Barbara Reskin's analysis of the race-discrimination system emphasizes that these gates are not independent events but a linked system, so that a small disadvantage at each stage accumulates into a large disparity at the end, with no single gate showing a difference large enough to be legally actionable or even individually noticeable (Reskin, 2012). The Worked Example below re-derives the arithmetic of this compounding.
The second is disambiguation under uncertainty. When a decision is clear-cut, the relevant criterion settles it and bias has little room to operate. When a decision is genuinely ambiguous — a borderline résumé, an equivocal gesture, a judgment call — the mind reaches for whatever associations the situation supplies to resolve the ambiguity, and in a racialized environment those associations are themselves structured by race. Discrimination therefore concentrates in exactly the ambiguous cases, where it is most deniable, which is also what makes it durable: each biased resolution looks, locally, like a reasonable call on a hard case.
The third is the color-blind interpretive frame. Bonilla-Silva argues that contemporary racism is sustained not by overt supremacist ideology but by a color-blind one that recasts structural disparity as the natural outcome of individual merit, cultural difference, or market forces, and that treats any attention to race as itself the problem (Bonilla-Silva, 2015). This frame is cognitively potent because it supplies a ready, non-racial explanation for every disparity, so that the structure's outputs are continually reinterpreted as evidence that the structure is fair. The frame is the ideological maintenance system; it keeps the structure from being seen as a structure.
These mechanisms interlock. Cumulative disadvantage generates the disparity; disambiguation ensures the bias enters where it cannot be caught; the color-blind frame reinterprets the result as fair, which removes the pressure to change the gates. The loop is closed, and it closes without anyone needing to intend the outcome. The StructuralFeedbackLoop demonstration below traces how a disparity feeds back on itself across periods and settles at a self-sustaining fixed point.
Structural Feedback Loop
Each period's disparity carries into the next through the structure — segregated capital, network-based hiring, under-resourced schools — so a strong enough loop sustains disparity with no new discrimination at all. Set fresh input to zero and raise the feedback: the disparity refuses to decay. That self-reproduction is what makes the racism systemic rather than a sum of present-day acts.
How Systemic Racism Harms Health
The most extensively documented downstream consequence of systemic racism is its effect on population health, and this literature is where the structural account has been tested most rigorously. David R. Williams' research program establishes racism as a fundamental determinant of health that operates through multiple structural pathways — residential segregation that concentrates environmental hazard and limits access to care, differential institutional treatment, and the chronic physiological stress of navigating a racialized environment (Williams, Lawrence, & Davis, 2019). A systematic review and meta-analysis of the health effects of racism across hundreds of studies finds a consistent association between reported racism and worse mental and physical health, robust across populations and outcome measures (Paradies et al., 2015).
Gee and Ford drew the explicit distinction between individual and structural racism as health exposures, arguing that the field's focus on interpersonal discrimination had underestimated the health burden because it missed the structural layer — the segregation, policy, and institutional arrangements that shape exposure before any interpersonal encounter occurs (Gee & Ford, 2011). Feagin and Bennefield traced systemic racism through the specific institutions of U.S. health care, showing how the system's history and structure produce differential treatment independent of any individual clinician's intent (Feagin & Bennefield, 2014). Bailey and colleagues synthesized the evidence into a framework for intervention, arguing that because the exposure is structural the remedies must be too, and reviewing the policy levers for which there is evidence (Bailey et al., 2017).
The health literature matters to the cognitive account for a specific reason: it demonstrates that systemic racism has measurable effects that do not route through any individual's measured prejudice. A person's health can be shaped by the segregation of their neighborhood and the under-resourcing of their hospital regardless of the attitudes of the people they personally encounter. This is the structural claim made concrete — disadvantage delivered by the configuration of the system, not by the hostility of its operators.
Addressing Systemic Racism
The structural diagnosis implies a structural prescription, and the intervention evidence bears this out. The most-studied individual-level remedy — brief implicit-bias training — reliably shifts scores on the IAT in the short term but shows little evidence of durable change in behavior or, still less, in organizational outcomes. Onyeador, Hudson, and Lewis review this evidence and argue for moving beyond bias training toward structural levers: changing the criteria, defaults, and accountability structures through which decisions are made, which act on the gates themselves rather than on the operators' minds (Onyeador, Hudson, & Lewis, 2021). This is exactly what the bias-of-crowds model predicts: if bias is a property of situations, then changing situations, not retraining individuals, is the effective lever (Payne, Vuletich, & Lundberg, 2017).
Structural levers take several forms. Changing the decision architecture — removing race-correlated cues, standardizing criteria, structuring ambiguous judgments so that the gaps through which bias enters are narrowed — attacks the disambiguation mechanism directly. Changing the gates — reducing the number of sequential discretionary stages, or auditing each stage for differential pass rates — attacks cumulative disadvantage. Changing the frame — making the structure visible as a structure, so that disparities are read as outputs of a process rather than as natural facts — attacks the color-blind interpretation that protects the system from scrutiny. None of these requires changing anyone's heart; each changes the structure that produces the outcome.
The cultural-psychological tradition adds that because racism is built into everyday artifacts, those artifacts can be re-engineered: curricula, defaults, and physical spaces that currently transmit racial meanings can be redesigned to stop doing so (Salter, Adams, & Perez, 2018). The common thread is that the effective interventions operate at the level where the causation actually lives — the structure — rather than at the level where our folk theory mistakenly places it.
Worked Example: How a Small Gap Compounds
Figure 1
The cumulative-disadvantage mechanism is easiest to see in arithmetic. Consider a selection process with four sequential stages — say résumé screen, phone interview, on-site interview, and final offer — and two groups of 1,000 candidates each, a majority group and a minority group. Suppose that at each stage the majority group passes at a rate of 0.60 and the minority group at a rate of 0.54. The minority pass rate is 90% of the majority rate at every single stage — a ten-percent relative gap that is small, plausibly attributable to any number of non-racial factors, and far too small at any one stage to prove discrimination.
Trace the survivors. Each group's count is multiplied by its per-stage pass rate four times over:
- Majority survivors: 1,000 × 0.60⁴ = 1,000 × 0.1296 = 129.6 - Minority survivors: 1,000 × 0.54⁴ = 1,000 × 0.08503056 = 85.03
The aggregate ratio of minority to majority survivors is:
85.03 / 129.6 = 0.6561
which is exactly 0.90⁴ — the per-stage ratio raised to the number of stages. The ten-percent per-stage gap has compounded into a 34% aggregate shortfall: the minority group ends with barely two-thirds the representation of the majority group at the final stage, even though no single gate disadvantaged it by more than ten percent.
This is the structural point made quantitative. A disparity that is individually deniable at every stage — because ten percent is within the range of innocent explanation — is collectively decisive. An auditor examining any one gate finds nothing actionable; an auditor examining the pipeline end-to-end finds a large disparity with no stage to pin it on. The CumulativeDisparityPipeline demonstration below varies the per-stage gap and the number of stages, showing the aggregate ratio move as the power law predicts.
Cumulative Disparity Pipeline
A gap too small to notice or prove at any one gate becomes a large disparity across a chain of gates. This is cumulative disadvantage: the aggregate ratio equals the per-gate ratio raised to the number of gates, so adding gates multiplies the shortfall even when each gate is held fixed.
Discussion
The cognitive psychology of systemic racism is, in the end, an argument about units and levels. The folk theory and much of the law locate racism in the individual — in a prejudiced belief, an intention, an actor who can be identified and blamed. The structural account locates it in the system, and the bias-of-crowds model shows why the individual-level and system-level descriptions are not two views of one thing but two different quantities, only one of which carries the structural signal. The instability of individual IAT scores, long treated as an embarrassment for implicit-bias research, turns out to be exactly what the situational account predicts, and the stability of aggregate scores is the real phenomenon.
This has a sharp methodological consequence. If implicit bias is a property of situations, then studies that measure individuals, intervene on individuals, and evaluate individual change are measuring the noisy variable and missing the stable one. The aggregate is not a statistical convenience for summarizing individuals; it is the level at which the construct actually exists. The same logic applies to intervention: the repeated finding that individual bias training does not durably change outcomes is not a failure of the trainings but a confirmation that they target the wrong level.
The account also clarifies what is and is not being claimed. It is not claimed that individual prejudice does not exist or does not matter — it does, and it is where bias enters the ambiguous cases. What is claimed is that the individual layer is neither necessary nor sufficient for the disparity: not necessary, because the structure reproduces disparity without it; not sufficient, because reducing it without changing the structure leaves the gates intact. The cognitive and the structural are complementary descriptions of a single system, and the bias-of-crowds model is the hinge that joins them.
Current Directions
Active research is pressing on several fronts. The bias-of-crowds model has sharpened the measurement debate: if the aggregate is the construct, then the field needs aggregate-level measures and designs — comparisons across regions, institutions, and time — rather than the individual-difference paradigm that dominated the IAT's first two decades, and work continues on how best to estimate the situational level and relate it to structural covariates (Vuletich & Payne, 2019). The finding that county-level implicit bias tracks the 1860 enslavement rate has opened a line of historical-structural work asking how such signals are transmitted across generations and what contemporary structures maintain them (Payne, Vuletich, & Brown-Iannuzzi, 2019).
In health, the agenda set by Williams and by Bailey and colleagues is increasingly focused on measuring structural racism directly — building indices of segregation, policy, and institutional exposure — rather than inferring it from reported interpersonal discrimination, and on evaluating structural interventions against health outcomes (Williams, Lawrence, & Davis, 2019; Bailey et al., 2017). On the intervention side, the shift Onyeador and colleagues describe — from training individuals to redesigning decision architecture and holding institutions accountable — is being tested in hiring, lending, and criminal-justice settings, where the structural levers can be manipulated and their effects on disparity measured directly (Onyeador, Hudson, & Lewis, 2021). The common thread is methodological: a move from the individual to the structural as the unit of both measurement and action.
Common Misconceptions
- Systemic racism means everyone in the system is racist.
- It means almost the opposite. The structural account holds that disparity is produced by the configuration of institutions, markets, and defaults, so it persists even when few or no participants hold or act on prejudiced beliefs; a system can be systemically racist while most of its operators are well-intentioned (Bonilla-Silva, 1997).
- If each decision is only slightly unequal, the overall result must be only slightly unequal.
- Sequential gatekeeping compounds multiplicatively, not additively. As the worked example shows, a ten-percent gap repeated across four stages produces a thirty-four-percent aggregate shortfall, so small, deniable gaps at each gate yield a large disparity at the end (Reskin, 2012).
- A low or unstable implicit-bias score means implicit bias is not real.
- Individual IAT scores are noisy, but the bias-of-crowds model explains why: the score reads a fluctuating situational input, not a fixed trait. The stable, structurally meaningful signal is the aggregate, which reliably predicts regional disparity, so the instability is a feature of the measure's level, not evidence against the construct (Payne, Vuletich, & Lundberg, 2017).
- Bias training fixes institutional racism.
- The evidence does not support this. Brief training shifts IAT scores briefly but shows little durable effect on behavior or organizational outcomes; because the causation is structural, the effective levers are structural, changing criteria, defaults, and accountability rather than retraining individuals (Onyeador, Hudson, & Lewis, 2021).
Glossary
- Aggregate bias.
- The average level of implicit bias measured across the people in a region or institution. Unlike an individual score, it is stable over time and predicts that place's racial disparities; in the bias-of-crowds model it is the quantity that carries the structural signal.
- Bias of crowds.
- The model of implicit bias as a property of situations rather than of individuals: an individual test score reads the bias a context makes accessible at a moment, so it is noisy, while the aggregate across many people is stable because the situation is.
- Color-blind racism.
- An ideological frame that explains racial disparity through individual merit, culture, or market forces rather than structure, and treats attention to race as itself the problem; it reinterprets the structure's outputs as evidence that the structure is fair.
- Cumulative disadvantage.
- The compounding of small, per-stage disparities across a chain of sequential gatekeeping decisions into a large aggregate disparity, with no single stage showing a difference large enough to be noticed or actioned.
- Disambiguation.
- The process of resolving a genuinely ambiguous decision by reaching for available associations; in a racialized environment those associations are structured by race, so discrimination concentrates in ambiguous cases where it is most deniable.
- Ecosocial theory.
- Nancy Krieger's framework describing how social conditions, including structural racism, become biologically embodied over the life course, shaping the distribution of disease across populations.
- Implicit Association Test.
- A reaction-time measure of the relative strength of automatic associations between categories, such as racial groups and positive or negative concepts; widely used to index implicit bias, with noisy individual scores but stable regional aggregates.
- Institutional racism.
- Racial disadvantage produced by the policies and routine practices of a particular organization; a narrower, organization-level relative of structural and systemic racism.
- Racialized social system.
- Bonilla-Silva's term for a society whose economic, political, and ideological structures allocate advantage along racial lines, so that discrimination is carried out by the routine operation of institutions rather than by individual intent.
- Residential segregation.
- The spatial separation of racial groups into different neighborhoods, a core structural mechanism that concentrates environmental hazard, limits access to care and opportunity, and transmits disadvantage across generations.
- Sequential gatekeeping.
- A decision process reached through a chain of separate stages, each a gate; because a small per-stage disadvantage multiplies across gates, it is the structure through which cumulative disadvantage operates.
- Structural racism.
- Racial disadvantage produced by the configuration of institutions, laws, and markets, including segregation, policy, and differential institutional treatment, operating as an exposure independent of any individual's prejudice; often used interchangeably with systemic racism, with structural emphasizing the institutional configuration.
- Systemic racism.
- Racial disadvantage produced and reproduced through the ordinary workings of a whole social system, including institutions, markets, and shared representations, rather than through individual hostility; a racism located at the level of the system.
- White racial frame.
- Feagin's term for the shared interpretive defaults, including stereotypes, images, and narratives, through which situations are read in a racialized society, supplying the representational layer that systemic racism maintains and draws on.
Key Researchers
Gordon W. Allport
(1897–1967; Harvard University). Author of The Nature of Prejudice (1954), the foundational analysis distinguishing individual prejudice from the institutional arrangements that sustain it, on which later structural accounts build. Wikipedia - Wikidata
Eduardo Bonilla-Silva
(Duke University). Originator of the theory of racialized social systems and of the analysis of color-blind racism, which locate racism in social structure and routine institutional practice rather than in individual prejudice alone. Wikipedia - Wikidata
Joe R. Feagin
(Texas A&M University). Developer of systemic racism theory, the account of racism as a durable, self-reproducing feature of the whole social order embedded in institutions, wealth, and a white racial frame. Wikipedia - Wikidata
Nancy Krieger
(Harvard University). Social epidemiologist whose ecosocial theory describes how structural racism becomes biologically embodied, and who has assembled the evidence and intervention framework linking structural racism to health inequities. ORCID - Wikipedia - Wikidata
Jennifer A. Richeson
(Yale University). Social psychologist whose work on misperceptions of racial economic equality and on the cognitive demands of navigating racialized systems connects individual cognition to structural inequality. Wikipedia - Wikidata
David R. Williams
(Harvard University). Leading scholar of structural racism as a determinant of population health; developer of widely used measures of everyday discrimination and of the pathways linking racism to racial disparities in health. Wikipedia - Wikidata
Frequently Asked Questions
What is the difference between systemic racism and individual racism?
Individual racism is prejudice held and acted on by a person; systemic racism is racial disadvantage produced by the ordinary operation of institutions, markets, and shared representations. The key difference is that systemic racism can persist even when few participants hold prejudiced beliefs, because the disparity is a feature of the structure rather than a sum of individual attitudes.
Is systemic racism the same as structural racism and institutional racism?
They are closely related and often used interchangeably. All three locate racism above the individual, in the configuration of institutions and systems. The structural label tends to emphasize the overall configuration of laws and markets, the institutional label a particular organization's practices, and the systemic label the self-reproducing character of the whole, but the distinctions are matters of emphasis rather than sharp boundaries.
How does cognitive psychology contribute to the study of systemic racism?
Its central contribution is the bias-of-crowds model, which recasts implicit bias as a property of situations rather than of individuals. This explains why individual bias scores are unstable while aggregate scores are stable and predict regional disparity, and it provides the bridge between momentary cognitive associations and durable social structure.
Why do racial disparities persist even as expressed prejudice declines?
Because the disparity is reproduced by structure, not by attitudes. Sequential gatekeeping compounds small gaps, ambiguous decisions absorb whatever associations the environment supplies, and a color-blind interpretive frame reinterprets the results as fair, a loop that runs independently of how prejudiced individuals are.
What is the bias-of-crowds model?
It is the proposal that implicit bias is the amount of bias a situation makes cognitively accessible, read through a person, rather than a fixed trait inside the person. An individual IAT score is therefore a noisy reading of a fluctuating situational input, while the aggregate across many people is stable because the local structure is stable.
Does systemic racism affect health?
Yes. A large literature, including a systematic review and meta-analysis across hundreds of studies, links racism to worse mental and physical health, and structural racism operates through segregation, differential institutional treatment, and chronic stress. These pathways shape health independently of any individual clinician's attitudes.
Does implicit-bias training reduce systemic racism?
The evidence says little. Brief training shifts test scores in the short term but shows scant durable effect on behavior or organizational outcomes. Because the causation is structural, researchers increasingly argue for structural levers, such as changing criteria, defaults, and accountability, rather than individual retraining.
How can a system be racist if no one in it is racist?
Through its design. A lending algorithm trained on segregated data, a school-funding formula tied to property values, or a hiring pipeline that recruits through existing networks will each reproduce racial disparity regardless of anyone's beliefs. The disparity is built into the process, so the system produces racist outcomes without requiring racist operators.
References
Allport, G. W. (1954). The nature of prejudice. Addison-Wesley.
Bailey, Z. D., Krieger, N., Agénor, M., Graves, J., Linos, N., & Bassett, M. T. (2017). Structural racism and health inequities in the USA: Evidence and interventions. The Lancet, 389(10077), 1453–1463. https://doi.org/10.1016/S0140-6736(17)30569-X
Bonilla-Silva, E. (1997). Rethinking racism: Toward a structural interpretation. American Sociological Review, 62(3), 465–480. https://doi.org/10.2307/2657316
Bonilla-Silva, E. (2015). The structure of racism in color-blind, "post-racial" America. American Behavioral Scientist, 59(11), 1358–1376. https://doi.org/10.1177/0002764215586826
Feagin, J., & Bennefield, Z. (2014). Systemic racism and U.S. health care. Social Science & Medicine, 103, 7–14. https://doi.org/10.1016/j.socscimed.2013.09.006
Gee, G. C., & Ford, C. L. (2011). Structural racism and health inequities: Old issues, new directions. Du Bois Review: Social Science Research on Race, 8(1), 115–132. https://doi.org/10.1017/S1742058X11000130
Greenwald, A. G., McGhee, D. E., & Schwartz, J. L. K. (1998). Measuring individual differences in implicit cognition: The implicit association test. Journal of Personality and Social Psychology, 74(6), 1464–1480. https://doi.org/10.1037/0022-3514.74.6.1464
Onyeador, I. N., Hudson, S. T. J., & Lewis, N. A., Jr. (2021). Moving beyond implicit bias training: Policy insights for increasing organizational diversity. Policy Insights from the Behavioral and Brain Sciences, 8(1), 19–26. https://doi.org/10.1177/2372732220983840
Paradies, Y., Ben, J., Denson, N., Elias, A., Priest, N., Pieterse, A., Gupta, A., Kelaher, M., & Gee, G. (2015). Racism as a determinant of health: A systematic review and meta-analysis. PLOS ONE, 10(9), e0138511. https://doi.org/10.1371/journal.pone.0138511
Payne, B. K., Vuletich, H. A., & Lundberg, K. B. (2017). The bias of crowds: How implicit bias bridges personal and systemic prejudice. Psychological Inquiry, 28(4), 233–248. https://doi.org/10.1080/1047840X.2017.1335568
Payne, B. K., Vuletich, H. A., & Brown-Iannuzzi, J. L. (2019). Historical roots of implicit bias in slavery. Proceedings of the National Academy of Sciences, 116(24), 11693–11698. https://doi.org/10.1073/pnas.1818816116
Reskin, B. (2012). The race discrimination system. Annual Review of Sociology, 38, 17–35. https://doi.org/10.1146/annurev-soc-071811-145508
Salter, P. S., Adams, G., & Perez, M. J. (2018). Racism in the structure of everyday worlds: A cultural-psychological perspective. Current Directions in Psychological Science, 27(3), 150–155. https://doi.org/10.1177/0963721417724239
Vuletich, H. A., & Payne, B. K. (2019). Stability and change in implicit bias. Psychological Science, 30(6), 854–862. https://doi.org/10.1177/0956797619844270
Williams, D. R., Lawrence, J. A., & Davis, B. A. (2019). Racism and health: Evidence and needed research. Annual Review of Public Health, 40, 105–125. https://doi.org/10.1146/annurev-publhealth-040218-043750