Broken Windows Theory - Criminology Theories

Broken windows theory proposes that visible signs of disorder and minor incivilities in a neighborhood — graffiti, litter, abandoned buildings, public intoxication, panhandling — signal that no one is in charge and that criminal behavior will be tolerated, triggering a developmental sequence in which unchecked disorder erodes informal social control, drives out law-abiding residents, attracts criminal offenders, and produces escalating levels of serious crime. Introduced by James Q. Wilson and George L. Kelling in a 1982 article in The Atlantic Monthly and established as one of the most influential and contested frameworks within Criminology Theories, broken windows theory transformed American policing by providing the theoretical justification for order-maintenance strategies — including zero-tolerance policing, quality-of-life enforcement, and aggressive misdemeanor prosecution — that dominated urban crime control from the 1990s through the 2010s. This article examines the theory’s origins, presents its causal logic and key propositions, traces its implementation in New York City and other jurisdictions, surveys the empirical evidence for and against the disorder-crime relationship, discusses the theory’s racial dimensions and civil liberties implications, and assesses its current standing in light of more than four decades of research and debate.

Outline

I. Introduction

II. Origins of the Theory

a. Zimbardo’s Experiment and the Metaphor
b. Wilson and Kelling’s Atlantic Article
c. The Causal Logic: From Disorder to Serious Crime

III. Key Concepts

a. Physical and Social Disorder
b. Fear of Crime and Community Withdrawal
c. The Role of Informal Social Control

IV. Implementation: The New York City Model

a. Bratton, CompStat, and Quality-of-Life Policing
b. The New York Crime Decline
c. Stop, Question, and Frisk

V. Empirical Evidence

a. Skogan’s Support and Its Reanalysis
b. Sampson and Raudenbush: The Collective Efficacy Challenge
c. Experimental Tests: Keizer, Lindenberg, and Steg
d. Taylor’s Baltimore Longitudinal Study
e. Braga and the Hot Spots Alternative

VI. The Racial Dimensions of Broken Windows Policing

a. Disparate Enforcement and Racial Profiling
b. Floyd v. City of New York
c. The Social Construction of Disorder

VII. Criticisms and Limitations

a. The Causal Arrow: Does Disorder Cause Crime?
b. Harcourt’s Critique: The Illusion of Order
c. The Displacement and Net-Widening Problems

VIII. Alternatives and Contemporary Assessment

a. From Order Maintenance to Problem-Oriented Policing
b. Community-Based Approaches
c. The Current Consensus

IX. Conclusion

X. References

Introduction

Can cleaning up graffiti prevent murder? Can arresting panhandlers reduce robbery? Can fixing broken windows stop a neighborhood from descending into serious crime? These are the provocative claims — or at least the provocative implications — of broken windows theory, one of the most famous, widely implemented, and fiercely debated ideas in the history of criminology and criminal justice policy.

Wilson and Kelling’s (1982) broken windows thesis offered a deceptively simple account of how neighborhoods decline: if a window in a building is broken and left unrepaired, observers conclude that no one cares about the building, and soon more windows are broken. The unrepaired window is a signal that vandalism and disorder are tolerated, and this signal invites further disorder and eventually serious crime. Applied to neighborhoods, the theory proposes that visible disorder — litter, graffiti, abandoned buildings, loitering, public drinking — communicates to potential offenders that the community has lost the capacity or willingness to maintain social order, and that criminal behavior will therefore go undetected and unpunished.

Broken windows theory within Criminology Theories had an enormous practical impact. It provided the intellectual justification for a transformation in American policing — from a reactive, incident-driven model focused on responding to serious crimes after they occurred to a proactive, order-maintenance model focused on preventing serious crimes by aggressively addressing the minor disorders that were theorized to be their precursors. The most famous implementation — New York City’s quality-of-life policing under Commissioner William Bratton in the 1990s — was credited by its advocates with producing the dramatic crime decline that transformed the city from one of the most dangerous in America to one of the safest. The theory’s critics, however, argue that the evidence linking disorder to serious crime is weak, that order-maintenance policing disproportionately harms racial minorities and the poor, and that the New York crime decline was driven by factors unrelated to broken windows policing.

Origins of the Theory

Zimbardo’s Experiment and the Metaphor

The metaphor that gave the theory its name was inspired by a 1969 experiment conducted by Stanford psychologist Philip Zimbardo. Zimbardo arranged for two automobiles without license plates to be abandoned — one in the Bronx, New York, and the other in Palo Alto, California. The car in the Bronx was attacked by vandals within ten minutes: within 24 hours, virtually everything of value had been stripped, and the car was systematically destroyed. The car in Palo Alto sat untouched for more than a week — until Zimbardo himself smashed part of it with a sledgehammer. Once the car showed visible signs of damage, passersby quickly joined in the destruction, and within hours the car was reduced to a wreck.

Zimbardo interpreted the experiment as a demonstration of the power of situational cues to influence behavior. The undamaged car in Palo Alto was left alone because it presented no signal that vandalism was acceptable. The damaged car — whether damaged by prior vandals (in the Bronx) or by the researcher (in Palo Alto) — communicated that the norms governing property respect had been suspended, releasing inhibitions and inviting further destruction. The experiment suggested that physical disorder operates as a behavioral cue that can trigger a cascade of increasingly destructive behavior.

Wilson and Kelling drew on Zimbardo’s experiment but extended its implications far beyond the vandalism of automobiles. They proposed that visible disorder in a neighborhood — a broken window left unrepaired, graffiti left uncleaned, litter left uncollected — functions in the same way as Zimbardo’s damaged car: it signals that the community’s norms of order and civility have broken down, that no one is monitoring behavior, and that further disorder and crime will be tolerated. Ansfield (2020) later argued that Wilson and Kelling selectively appropriated Zimbardo’s findings, dispensing with Zimbardo’s own emphasis on inequality and community anonymity as the underlying causes of the behavioral cascade in favor of a simpler story about the contagious effects of visible disorder.

Wilson and Kelling’s Atlantic Article

James Q. Wilson and George L. Kelling published “Broken Windows: The Police and Neighborhood Safety” in the March 1982 issue of The Atlantic Monthly. The article was not a formal academic study but an essay — a persuasive argument directed at a broad audience of policymakers, practitioners, and the educated public. Wilson, a political scientist at Harvard, and Kelling, a criminal justice researcher who had been involved in evaluating the Newark Foot Patrol Experiment, drew on the Newark findings to argue that foot patrol officers did not reduce crime rates but did reduce residents’ fear of crime and their perceptions of disorder — and that these effects were important because fear and disorder, left unchecked, would eventually produce real increases in serious crime.

The article presented the broken windows metaphor as a causal mechanism linking minor disorder to serious crime. Wilson and Kelling argued that disorder generates fear among law-abiding residents, who respond by withdrawing from public spaces — staying indoors, avoiding eye contact, declining to intervene when they observe threatening behavior. This withdrawal of law-abiding residents from the street weakens the informal social controls that normally regulate behavior in public spaces: the eyes on the street, the willingness of residents to challenge strangers and confront incivility, the sense that the community is watching and will respond to threats. As informal social control weakens, disorderly individuals claim public spaces as their own, and the neighborhood becomes increasingly attractive to serious criminal offenders who recognize that their activities are unlikely to be challenged or reported.

The article also argued that police should shift their focus from responding to serious crimes after they occur to maintaining public order by addressing the minor disorders that precede and produce serious crime. Wilson and Kelling drew on the Newark foot patrol findings to argue that officers on foot — visible, accessible, and engaged with the community — could maintain order more effectively than officers in patrol cars, who were isolated from the public and focused on responding to 911 calls. The policy prescription was clear: police should be reoriented from crime fighting to order maintenance, with priority given to addressing the signs of disorder that signal community decline.

The Causal Logic: From Disorder to Serious Crime

The causal logic of broken windows theory can be summarized as a developmental sequence with several stages. In the first stage, minor disorder appears in a neighborhood — a window is broken, graffiti appears on a wall, litter accumulates on the sidewalk, a group of young men begins congregating on a street corner. In the second stage, residents observe the disorder and conclude that the neighborhood is declining and that the authorities (both police and community institutions) are unable or unwilling to maintain order. In the third stage, fear rises and residents withdraw from public life — spending less time on the street, avoiding unfamiliar areas, declining to intervene when they observe threatening behavior, and eventually moving out of the neighborhood if they have the means to do so.

In the fourth stage, the withdrawal of law-abiding residents weakens the informal social controls that previously regulated behavior in public spaces. Without the eyes on the street, the willingness to challenge incivility, and the sense of community ownership that characterized the ordered neighborhood, disorderly individuals — panhandlers, drunks, drug users, prostitutes, loitering youth — claim public spaces for their own use. In the fifth stage, the neighborhood’s reputation as a disordered, unpoliced area attracts serious criminal offenders — burglars, robbers, drug dealers, violent predators — who recognize that the breakdown of informal social control creates an environment in which their activities are unlikely to be challenged, witnessed, or reported to police.

This developmental sequence — from minor disorder through fear and community withdrawal to the collapse of informal social control and the invasion of serious crime — is the core theoretical proposition of broken windows theory. It implies that the most effective point of intervention is at the beginning of the sequence: if minor disorders are addressed before they can trigger the cascade, the entire developmental process can be arrested and serious crime can be prevented. This implication provided the intellectual foundation for the order-maintenance policing strategies that would transform American law enforcement in the 1990s.

Key Concepts

Physical and Social Disorder

Broken windows theory distinguishes between two forms of disorder that contribute to the developmental sequence. Physical disorder refers to the visible deterioration of the built environment: broken windows, abandoned buildings, graffiti, litter, abandoned vehicles, overgrown lots, and other signs of neglect and decay. Social disorder refers to the visible presence of individuals engaged in disruptive, threatening, or norm-violating behavior: panhandlers, public drinkers, drug users, loiterers, prostitutes, rowdy groups of youth, and homeless individuals sleeping in public spaces.

Both forms of disorder are theorized to operate through the same mechanism: they signal to observers — both law-abiding residents and potential offenders — that the community has lost its capacity to regulate behavior in public spaces. Physical disorder communicates that property is not being maintained and that owners and authorities have abandoned their responsibility for the built environment. Social disorder communicates that behavioral norms are not being enforced and that individuals can engage in disruptive conduct without consequence.

The distinction between physical and social disorder has important implications for policy. Physical disorder can be addressed through environmental interventions — repairing broken windows, cleaning graffiti, removing abandoned vehicles, maintaining vacant lots — that do not involve the criminal justice system. Social disorder, by contrast, is typically addressed through policing — arresting or dispersing panhandlers, clearing homeless encampments, enforcing anti-loitering ordinances — interventions that raise significant civil liberties concerns and that have been criticized for disproportionately targeting the poor and racial minorities.

Fear of Crime and Community Withdrawal

Fear of crime is the psychological mechanism that links disorder to the breakdown of informal social control in the broken windows model. Wilson and Kelling argued that visible disorder generates fear — a sense of vulnerability, unease, and danger — among residents who observe it. This fear is distinct from the actual risk of criminal victimization: residents may feel unsafe in a disorderly neighborhood even if the crime rate is no higher than in an orderly one, because the visible signs of disorder communicate that the neighborhood is out of control and that victimization is more likely.

The relationship between disorder and fear has been extensively studied. Research has generally confirmed that residents of disorderly neighborhoods report higher levels of fear of crime, lower levels of neighborhood satisfaction, and weaker attachment to their communities compared to residents of orderly neighborhoods (Skogan, 1990; Taylor, 2001). However, the causal direction of the relationship is contested: disorder may cause fear (as the theory predicts), or fear may cause perceptions of disorder (residents who are already fearful may perceive their neighborhood as more disorderly than it actually is), or both disorder and fear may be caused by a third factor — concentrated disadvantage — that produces both conditions simultaneously.

The community withdrawal that fear produces is theorized to weaken informal social control — the willingness and capacity of residents to monitor public spaces, challenge incivility, intervene in threatening situations, and report suspicious activity to police. This withdrawal creates a vacuum that disorderly individuals and criminal offenders fill, accelerating the developmental sequence from disorder to crime. The withdrawal mechanism connects broken windows theory to social disorganization theory, which identifies the breakdown of informal social control as the primary cause of high crime rates in disadvantaged neighborhoods.

The Role of Informal Social Control

Informal social control — the capacity of community residents to regulate behavior through surveillance, intervention, and the enforcement of shared norms — is the critical mediating variable in the broken windows model. Wilson and Kelling argued that informal social control, not formal police enforcement, is the primary mechanism through which communities maintain order and prevent crime. In well-ordered neighborhoods, residents monitor public spaces, challenge strangers, confront incivility, and communicate to potential offenders that their behavior is being observed and that the community will respond.

This emphasis on informal social control connects broken windows theory to the broader social control tradition in criminology — particularly to Sampson, Raudenbush, and Earls’s (1997) concept of collective efficacy, which defines informal social control as the combination of social cohesion and shared expectations for intervention among neighbors. The relationship between broken windows theory and collective efficacy theory is complex and contested: both identify informal social control as the key factor in preventing crime, but they disagree about the relationship between disorder and informal social control. Broken windows theory proposes that disorder causes the breakdown of informal social control; collective efficacy theory proposes that weak informal social control causes both disorder and crime, and that the apparent relationship between disorder and crime is spurious.

The debate between these two accounts — whether disorder is a cause of weakened social control or a consequence of it — is the central empirical question in the broken windows literature and has generated some of the most sophisticated and consequential research in contemporary criminology.

Implementation: The New York City Model

Bratton, CompStat, and Quality-of-Life Policing

The most famous and influential implementation of broken windows theory occurred in New York City under Mayor Rudolph Giuliani and Police Commissioner William Bratton, beginning in 1994. Bratton, who had previously implemented order-maintenance policing on the New York City subway system, adopted broken windows as the guiding philosophy of the New York City Police Department and launched a “quality-of-life” policing initiative that aggressively targeted minor offenses — turnstile jumping, public urination, aggressive panhandling, unlicensed street vending, public drinking, and marijuana possession — that had previously been treated as low-priority nuisances.

The quality-of-life initiative was supported by CompStat (Computer Statistics) — a data-driven management system that tracked crime patterns at the precinct level, held precinct commanders accountable for crime reductions in their areas, and provided the analytical infrastructure for deploying police resources to the specific locations and offense categories that CompStat data identified as priorities. CompStat transformed police management by making crime data available in near-real-time and by creating institutional incentives for precinct commanders to produce measurable crime reductions — incentives that drove the aggressive enforcement of minor offenses as a strategy for preventing serious crime.

The combination of broken windows theory, quality-of-life policing, and CompStat represented a fundamental transformation in American policing — from a reactive model focused on responding to serious crimes after they occurred to a proactive model focused on preventing serious crimes by addressing the minor disorders that were theorized to precede them. The NYPD’s approach was widely publicized, enthusiastically endorsed by political leaders, and adopted (in various forms) by police departments across the United States and internationally.

The New York Crime Decline

The implementation of broken windows policing in New York coincided with one of the most dramatic crime declines in American history. Between 1990 and 2000, murder declined by approximately 73 percent, robbery by 67 percent, and burglary by 66 percent — reductions that transformed New York from one of the most dangerous major cities in the United States to one of the safest. Advocates of broken windows policing — including Bratton, Kelling, and Mayor Giuliani — attributed the crime decline to the quality-of-life initiative, arguing that the aggressive enforcement of minor offenses had interrupted the developmental sequence from disorder to serious crime.

However, the attribution of the New York crime decline to broken windows policing was contested by researchers who noted that crime declined across virtually all major American cities during the same period — including cities that did not adopt order-maintenance policing. Levitt (2004) identified several factors that contributed to the nationwide crime decline of the 1990s, including the waning of the crack cocaine epidemic, increases in the incarceration rate, growth in the number of police officers, and the legalization of abortion in the 1970s (which, he argued, reduced the cohort of unwanted children who would have reached peak offending age in the 1990s). None of these factors was specific to New York, and the crime decline in New York was roughly comparable in magnitude to declines in cities with very different policing strategies.

Harcourt and Ludwig (2006) conducted one of the most rigorous tests of the broken windows hypothesis using data from New York City and from the Moving to Opportunity (MTO) experiment — a randomized housing mobility program that moved families from high-poverty, high-disorder neighborhoods to lower-poverty, lower-disorder neighborhoods. If broken windows theory is correct, the families that moved to lower-disorder neighborhoods should have experienced lower victimization rates. Harcourt and Ludwig found no evidence of such an effect, concluding that the evidence provides “no support for a simple first-order disorder-crime relationship as hypothesized by Wilson and Kelling.”

Stop, Question, and Frisk

The most controversial application of broken windows policing in New York was the dramatic expansion of the NYPD’s stop, question, and frisk (SQF) program. Under the logic of broken windows theory, police encounters with individuals engaged in minor disorder or suspicious behavior were expected to deter serious crime by increasing the perceived certainty of detection. SQF stops increased from approximately 97,000 in 2002 to a peak of over 685,000 in 2011 — with the overwhelming majority of stops targeting Black and Latino males.

The racial dimensions of SQF generated fierce controversy and ultimately led to the landmark federal court decision Floyd v. City of New York (2013), discussed below. The expansion of SQF also raised the question of whether the broken windows model — originally framed as a theory about community order and informal social control — had been transformed in practice into a justification for aggressive, racially targeted policing that bore little resemblance to the community-oriented approach that Wilson and Kelling had originally envisioned.

Kelling himself expressed reservations about the way broken windows theory had been implemented, arguing that zero-tolerance policing — the indiscriminate enforcement of all minor offenses regardless of community context — was a distortion of the theory, which called for police to exercise discretion and to work with communities to identify the specific forms of disorder that residents found most threatening. This distinction between the theory as originally proposed and the policing strategies that were implemented in its name has become an important theme in the contemporary assessment of broken windows.

Empirical Evidence

Skogan’s Support and Its Reanalysis

Wesley Skogan’s (1990) Disorder and Decline: Crime and the Spiral of Decay in American Neighborhoods provided the first major empirical support for the broken windows thesis. Using survey data from 40 neighborhoods in six American cities, Skogan found a significant positive relationship between disorder and crime — neighborhoods with higher levels of disorder had higher levels of robbery victimization, even after controlling for poverty, residential instability, and racial composition. Skogan interpreted this finding as supporting the broken windows hypothesis that disorder has an independent causal effect on crime.

However, Harcourt (2001) reanalyzed Skogan’s data and found that the relationship between disorder and robbery was entirely driven by five of the 40 neighborhoods — and that when these outlier neighborhoods were removed, the relationship disappeared. Harcourt also found that Skogan’s results were sensitive to the specific control variables included in the analysis: when additional controls for poverty and minority concentration were included, the disorder-crime relationship was no longer statistically significant. Xu, Fiedler, and Flaming (2005) conducted a further reanalysis that partially supported Skogan’s findings but also identified significant methodological limitations.

The Skogan-Harcourt exchange illustrates the fragility of the empirical evidence for broken windows theory. The existence of a disorder-crime relationship is not in dispute — neighborhoods with more disorder tend to have more crime. What is disputed is the nature of this relationship: whether disorder causes crime (as broken windows theory claims), whether disorder and crime are both caused by underlying structural factors (as the collective efficacy framework suggests), or whether the observed relationship is an artifact of methodological choices and outlier observations.

Sampson and Raudenbush: The Collective Efficacy Challenge

The most influential empirical challenge to broken windows theory came from Sampson and Raudenbush (1999), who used systematic social observation — video recording of 23,816 street segments in Chicago — combined with survey data and official crime statistics to test the relationships among disorder, collective efficacy, and crime. Their analysis found that both disorder and crime were concentrated in neighborhoods characterized by disadvantage and low collective efficacy — but that the effect of disorder on crime was substantially mediated by collective efficacy. When collective efficacy was controlled, the direct effect of disorder on most types of crime was small and often nonsignificant.

Sampson and Raudenbush concluded that disorder and crime are both manifestations of the same underlying condition — weak collective efficacy — rather than standing in a direct causal relationship with each other. In neighborhoods with strong collective efficacy (high social cohesion and shared willingness to intervene), disorder is kept in check and crime rates are low. In neighborhoods with weak collective efficacy, both disorder and crime flourish — not because disorder causes crime but because the same structural conditions (poverty, residential instability, concentrated disadvantage) that produce disorder also produce crime by undermining the community’s capacity for informal social control.

This finding challenged the foundational premise of broken windows policing: if disorder does not independently cause crime, then addressing disorder through aggressive enforcement should not be expected to reduce serious crime. The policy implication of the collective efficacy framework is that effective crime prevention requires strengthening the social cohesion and shared expectations for intervention that constitute collective efficacy — a goal that is more likely to be achieved through community development, institutional investment, and relationship-building than through aggressive order-maintenance policing.

Sampson and Raudenbush (2004) extended their analysis by demonstrating that residents’ perceptions of disorder are influenced not only by the actual level of disorder in their neighborhood but also by the neighborhood’s racial composition and socioeconomic status. Neighborhoods with larger proportions of minority residents were perceived as more disorderly than their actual observed disorder levels warranted — a finding that suggested the disorder concept itself is contaminated by racial stereotypes and that broken windows policing may target neighborhoods for enforcement based on their racial composition rather than their actual disorder levels.

Experimental Tests: Keizer, Lindenberg, and Steg

Keizer, Lindenberg, and Steg (2008) conducted the first experimental test of the broken windows thesis in a series of six field experiments in the Netherlands. They manipulated the level of disorder in controlled environments — placing graffiti on a wall near a mailbox, scattering litter in a parking garage, removing a sign prohibiting entry to an alley — and observed whether the experimentally induced disorder increased norm-violating behavior among passersby (littering, trespassing, stealing).

The results provided support for the spreading effect that broken windows theory describes: individuals were more likely to litter, trespass, and steal in environments where visible disorder was present than in environments where it was absent. The effect sizes were substantial and consistent across all six experiments. The Keizer et al. study is frequently cited as the strongest experimental evidence for the broken windows hypothesis.

However, the external validity of the Keizer et al. findings is limited. The experiments tested whether visible disorder increases minor norm violations (littering, trespassing, petty theft) in controlled laboratory-like settings — not whether disorder causes serious crime in real neighborhoods over time. The leap from demonstrating that graffiti increases littering to concluding that graffiti causes robbery requires a chain of causal reasoning that the experiments did not test. The experiments provide evidence for a “spreading of disorder” effect but do not demonstrate the full developmental sequence — from minor disorder through community withdrawal and the collapse of informal social control to serious crime — that constitutes the broken windows thesis.

Taylor’s Baltimore Longitudinal Study

Ralph Taylor’s (2001) Breaking Away from Broken Windows: Baltimore Neighborhoods and the Nationwide Fight Against Crime, Grime, and Fear represents the most thorough longitudinal test of the broken windows thesis. Using data from 66 Baltimore neighborhoods tracked over a 12-year period (1981–1993), Taylor examined whether changes in physical and social disorder predicted subsequent changes in crime, fear of crime, and neighborhood decline.

Taylor’s results provided mixed and largely unsupportive evidence for the broken windows thesis. Physical deterioration predicted some increases in fear and some increases in certain types of crime, but the relationships were modest and inconsistent across different crime types and different measures of disorder. The longitudinal design allowed Taylor to test the causal direction of the disorder-crime relationship, and his results suggested that the relationship is bidirectional: disorder may contribute to subsequent crime in some circumstances, but crime also contributes to subsequent disorder, and both are substantially driven by underlying structural conditions (poverty, residential instability, racial segregation) that produce both disorder and crime.

Taylor concluded that broken windows theory overstates the causal significance of disorder by treating it as the primary driver of neighborhood decline when it is, in fact, one of many symptoms of the structural disadvantage that produces both disorder and crime. His analysis supported a more modest version of the disorder hypothesis — that disorder may contribute to fear and to some forms of crime in some contexts — while rejecting the strong version that disorder is the initiating cause of a developmental sequence leading inevitably to serious crime.

Braga and the Hot Spots Alternative

Anthony Braga and colleagues have contributed to the broken windows debate by examining the effects of disorder reduction within the framework of hot spots policing rather than order-maintenance policing. Braga, Welsh, and Schnell (2015) conducted a meta-analysis of disorder-focused policing interventions and found that policing strategies that targeted disorder at specific high-crime locations — through a combination of environmental improvements, community engagement, and targeted enforcement — produced significant reductions in crime. However, the effective strategies were not the aggressive, city-wide order-maintenance approaches associated with broken windows policing but more focused, problem-oriented interventions that addressed the specific conditions generating disorder and crime at particular locations.

This finding suggests that the relationship between disorder and crime may operate differently at different geographic scales. At the micro-level (specific addresses, street corners, and blocks), addressing disorder through environmental improvements and targeted interventions may reduce crime by removing criminal opportunities and improving natural surveillance. At the neighborhood level, however, the causal significance of disorder is much less clear, and aggressive order-maintenance policing may produce more harm — in the form of community resentment, racial profiling, and civil liberties violations — than benefit.

Weisburd, Hinkle, Braga, and Wooditch (2015) found that broken windows policing at crime hot spots produced crime reductions without the adverse community effects (reduced legitimacy, increased distrust) that critics had predicted — suggesting that the negative consequences of broken windows policing may be attributable to how the strategy is implemented (aggressively, across entire cities, with racial disparities) rather than to the disorder-reduction approach itself.

Table 1. Key Empirical Tests of Broken Windows Theory

Study	Method	Finding	Implication
Skogan (1990)	Survey data, 40 neighborhoods	Disorder predicts robbery after controls	Supports theory; but see Harcourt reanalysis
Harcourt (2001)	Reanalysis of Skogan data	Relationship driven by 5 outlier neighborhoods	Weakens Skogan’s support; findings fragile
Sampson & Raudenbush (1999)	Systematic social observation, Chicago	Disorder-crime link mediated by collective efficacy	Disorder and crime are co-effects, not cause and effect
Harcourt & Ludwig (2006)	MTO experiment + NYC analysis	No support for disorder-crime relationship	Against the theory’s strongest claims
Keizer et al. (2008)	Six field experiments, Netherlands	Disorder increases minor norm violations	Supports spreading effect; limited to minor deviance
Taylor (2001)	Longitudinal, 66 Baltimore neighborhoods	Mixed; modest and inconsistent effects of disorder	Structural factors dominate; disorder is secondary
Braga et al. (2015)	Meta-analysis of disorder policing	Problem-oriented disorder reduction reduces crime at hot spots	Place-based, not citywide, disorder reduction works

The Racial Dimensions of Broken Windows Policing

Disparate Enforcement and Racial Profiling

The implementation of broken windows policing has been characterized by stark racial disparities. In New York City, SQF data revealed that approximately 83 percent of all stops between 2004 and 2012 involved Black and Latino individuals — a proportion far exceeding their share of the city’s population and not fully explained by differences in neighborhood crime rates. Misdemeanor marijuana arrests, which increased dramatically under the quality-of-life initiative, were also racially disproportionate: Black and Latino New Yorkers were arrested for marijuana possession at rates far exceeding those of white New Yorkers, despite comparable rates of marijuana use across racial groups (Harcourt & Ludwig, 2007; Golub, Johnson, & Dunlap, 2007).

These racial disparities reflect the discretionary nature of order-maintenance policing. Unlike serious crimes, which are typically reported by victims and investigated by detectives, minor disorder offenses are almost entirely police-initiated — officers decide where to patrol, whom to stop, and which offenses to enforce. When these discretionary decisions are shaped by racial stereotypes, implicit biases, and the concentration of police resources in minority neighborhoods, the result is a pattern of enforcement that targets Black and Latino communities for a level of surveillance and intervention that white communities are spared.

The racial dimensions of broken windows policing have led critics to argue that the theory functions as a race-neutral justification for racially targeted policing. By framing enforcement in terms of disorder rather than race, broken windows policing allows police departments to concentrate their resources in minority neighborhoods and to target minority individuals without explicitly invoking racial categories — while producing outcomes that are indistinguishable from those produced by overtly racial profiling.

Floyd v. City of New York

The racial dimensions of broken windows policing in New York culminated in the landmark federal court decision Floyd v. City of New York (2013), in which Judge Shira Scheindlin ruled that the NYPD’s SQF program had been conducted in a manner that violated the Fourth and Fourteenth Amendments. Judge Scheindlin found that the NYPD had engaged in a practice of making stops without reasonable suspicion and had disproportionately targeted Black and Latino individuals in a manner that constituted racial profiling.

The Floyd decision required the NYPD to adopt reforms including the appointment of an independent monitor, the development of new training programs, and the implementation of a pilot program for body-worn cameras. The decision effectively ended the mass SQF program: stops declined from over 685,000 in 2011 to approximately 12,000 by 2016. Notably, the dramatic reduction in SQF stops was not accompanied by an increase in serious crime — a finding that undermined the claim that aggressive order-maintenance policing was necessary for public safety.

The Floyd decision and its aftermath provided a natural experiment in the effects of reducing broken windows policing. If the theory were correct — if aggressive enforcement of minor disorders is necessary to prevent serious crime — then the near-elimination of SQF should have produced significant increases in crime. The fact that crime continued to decline in New York City after SQF was curtailed suggests that the relationship between order-maintenance policing and serious crime is considerably weaker than broken windows advocates claimed.

The Social Construction of Disorder

Sampson and Raudenbush’s (2004) finding that perceptions of disorder are influenced by racial composition — that residents perceive neighborhoods with more minority residents as more disorderly, even after accounting for actual observed disorder levels — raises fundamental questions about the objectivity of the disorder concept on which broken windows theory depends. If disorder is not an objective feature of the environment but a socially constructed perception shaped by racial stereotypes, then policing strategies based on disorder are, at least in part, policing strategies based on race.

Sampson (2012) extended this analysis in Great American City, arguing that the relationship between perceived disorder and racial composition reflects a broader process of racial-spatial stigmatization — the association of minority neighborhoods with danger, decay, and social problems that is embedded in American culture and reproduced through media representations, political discourse, and institutional practices. Broken windows policing, in this analysis, does not merely respond to disorder — it participates in the social construction of disorder by deploying police resources to neighborhoods whose residents are perceived as disorderly because of their racial and socioeconomic characteristics.

This critique strikes at the theoretical foundations of broken windows theory. If the categories of “order” and “disorder” on which the theory depends are not objective descriptions of neighborhood conditions but socially constructed perceptions shaped by race, class, and power, then the theory’s claim to identify a universal causal mechanism — disorder causes crime — is fundamentally undermined. What the theory identifies as a relationship between disorder and crime may, in fact, be a relationship between racial stigmatization and the concentration of punitive state intervention in minority communities.

Table 2. Broken Windows Theory: The Debate

Proposition	Supporting Evidence	Challenging Evidence
Disorder causes fear	Survey studies find disorder-fear correlation (Skogan 1990; Taylor 2001)	Fear may cause disorder perceptions; both may reflect structural conditions
Fear causes community withdrawal	Theoretical proposition; limited direct testing	Withdrawal may reflect rational risk assessment rather than disorder-driven fear
Disorder causes serious crime	Skogan (1990); Keizer et al. (2008) for minor norm violations	Sampson & Raudenbush (1999); Harcourt (2001); Harcourt & Ludwig (2006); Taylor (2001)
Order-maintenance policing reduces crime	NYC crime decline; Braga et al. (2015) for hot spots	Crime declined nationwide; SQF reduction did not increase crime; MTO found no effect
Disorder perceptions are objective	Wilson & Kelling’s implicit assumption	Sampson & Raudenbush (2004): racial composition shapes disorder perceptions

Criticisms and Limitations

The Causal Arrow: Does Disorder Cause Crime?

The most fundamental criticism of broken windows theory concerns the direction of the causal arrow: does disorder cause crime, or are disorder and crime both caused by the same underlying structural conditions? The empirical evidence, reviewed above, provides far stronger support for the second interpretation than for the first. Sampson and Raudenbush’s (1999) finding that the disorder-crime relationship is substantially mediated by collective efficacy, Taylor’s (2001) finding that the longitudinal relationship between disorder and crime is modest and bidirectional, and Harcourt and Ludwig’s (2006) finding of no disorder-crime effect in the MTO experiment all suggest that the simple causal claim at the heart of broken windows theory — that disorder independently causes serious crime — is not supported by the available evidence.

This does not mean that disorder is irrelevant to crime. Disorder may contribute to fear, reduce quality of life, accelerate neighborhood decline, and create opportunities for certain types of crime — particularly at specific microgeographic locations where disorder reduces natural surveillance and signals the absence of guardianship. But these effects are more modest, more conditional, and more context-dependent than the strong version of broken windows theory implies, and they do not support the conclusion that aggressive, citywide order-maintenance policing is the optimal strategy for preventing serious crime.

The causal ambiguity has important policy implications. If disorder causes crime, then addressing disorder through aggressive policing is a rational and potentially effective crime prevention strategy. If disorder and crime are both caused by underlying structural conditions, then addressing disorder through policing treats a symptom rather than a cause — and may produce harmful side effects (racial profiling, community resentment, civil liberties violations) that outweigh any crime-reduction benefits.

Harcourt’s Critique: The Illusion of Order

Bernard Harcourt’s (2001) Illusion of Order: The False Promise of Broken Windows Policing provided the most sustained and systematic critique of broken windows theory. Harcourt argued that the theory rests on unexamined categories of “order” and “disorder,” “law-abiding citizens” and “disorderly people” that have no intrinsic reality independent of the techniques of punishment implemented to enforce them. The boundary between order and disorder is not a natural feature of the social environment but a product of political decisions about which behaviors and which populations will be subjected to police intervention — decisions that are shaped by race, class, and power.

Harcourt also challenged the empirical foundations of the theory, demonstrating (through his reanalysis of Skogan’s data and his analysis of the MTO experiment) that the evidence for a disorder-crime causal relationship is weak at best and nonexistent at worst. He argued that the popularity of broken windows policing was driven not by its empirical support — which was minimal — but by its political appeal: it offered a simple, intuitively appealing story about crime that justified aggressive policing without requiring the politically difficult investments in poverty reduction, education, and social services that structural accounts of crime demanded.

Harcourt’s critique also addressed the civil liberties implications of order-maintenance policing. Aggressive enforcement of minor offenses — particularly when directed at constitutionally protected activities such as panhandling and loitering — raises serious First Amendment and Fourth Amendment concerns. The expansion of police discretion to define and enforce “disorder” creates opportunities for racial profiling, selective enforcement, and the punishment of poverty and homelessness rather than criminal behavior. The result, Harcourt argued, is a policing regime that is both empirically unjustified and constitutionally problematic.

The Displacement and Net-Widening Problems

Broken windows policing faces two additional practical problems. The displacement problem concerns the possibility that aggressive enforcement of disorder in one location simply displaces disorderly individuals and criminal activity to other locations — a concern that is particularly relevant to the policing of homeless populations, whose removal from one public space typically results in their appearance in another. If disorder-focused policing merely moves disorder around rather than reducing it, then its crime-prevention effects may be minimal.

The net-widening problem concerns the expansion of the criminal justice system’s reach that order-maintenance policing produces. Aggressive enforcement of minor offenses generates large numbers of misdemeanor arrests, court appearances, fines, and criminal records — burdening the court system, overwhelming public defender offices, and saddling individuals with criminal records that reduce their employment prospects and increase their likelihood of future criminal justice involvement. The collateral consequences of misdemeanor arrest — including the disruption of employment, housing, family relationships, and educational pursuits — may produce more harm than the minor disorders they are intended to address.

Research on the consequences of misdemeanor arrest has found that even brief encounters with the criminal justice system can have lasting negative effects on employment, earnings, and social integration — effects that are concentrated among the poor and racial minorities who are disproportionately targeted by order-maintenance policing (Kohler-Hausmann, 2018). The net-widening critique suggests that broken windows policing may undermine the informal social controls it seeks to protect by damaging the social bonds and economic prospects of the very community members whose conformity depends on their stake in conventional society.

Alternatives and Contemporary Assessment

From Order Maintenance to Problem-Oriented Policing

The contemporary assessment of broken windows theory has increasingly converged on the conclusion that disorder matters — but not in the simple, direct, causal way that Wilson and Kelling proposed. Disorder is associated with fear, reduced quality of life, and some forms of crime — particularly at specific locations where it creates criminal opportunities and signals the absence of guardianship. But the relationship is mediated by structural conditions and collective efficacy, and aggressive citywide order-maintenance policing is not the most effective or the most just strategy for addressing it.

Problem-oriented policing (Goldstein, 1990) offers an alternative approach that retains broken windows theory’s insight about the importance of disorder while avoiding its most problematic policy implications. Problem-oriented policing identifies specific problems generating crime and disorder at particular locations, diagnoses the underlying conditions producing those problems, and develops tailored interventions that address those conditions — interventions that may include environmental improvements, community engagement, social services, and targeted enforcement, but that are designed to address the specific conditions at each location rather than to impose a uniform order-maintenance regime across the entire city.

The Braga et al. (2015) meta-analysis supports this approach, finding that problem-oriented disorder interventions at crime hot spots produce significant crime reductions without the adverse community effects associated with aggressive citywide order-maintenance policing. The evidence suggests that disorder reduction is most effective when it is targeted, place-based, and integrated with broader community development strategies — rather than implemented as a citywide enforcement mandate.

Community-Based Approaches

Community-based approaches to addressing disorder — including community clean-up programs, vacant lot remediation, community development corporations, and resident-led neighborhood improvement initiatives — offer alternatives that address physical disorder without the civil liberties concerns and racial disparities associated with order-maintenance policing. Branas, South, Kondo, Hohl, Bourgois, Wiebe, and MacDonald (2018) conducted a randomized controlled trial of a vacant lot remediation program in Philadelphia and found that the greening and maintenance of abandoned lots was associated with significant reductions in gun violence, burglary, and nuisance complaints — providing experimental evidence that environmental improvement can reduce crime without police intervention.

These community-based approaches are consistent with the collective efficacy framework: by improving the physical environment and demonstrating community investment, they strengthen the social cohesion and shared expectations for intervention that constitute collective efficacy — addressing the underlying cause of both disorder and crime rather than merely treating the symptoms through enforcement. The evidence supports a shift from policing disorder to building community capacity — a shift that retains the valid insights of broken windows theory while avoiding its most harmful consequences.

Community-based approaches also address the legitimate concerns about quality of life and neighborhood livability that motivated Wilson and Kelling’s original essay. Residents of disorderly neighborhoods suffer real consequences — fear, reduced mobility, diminished quality of life, lower property values — and their concerns deserve serious attention. The question is not whether disorder matters but how it should be addressed: through aggressive policing that produces racial disparities and civil liberties violations, or through community investment that strengthens the social and environmental conditions that prevent both disorder and crime.

The Current Consensus

The current scholarly consensus on broken windows theory can be summarized in several propositions. First, disorder and crime are correlated, but the relationship is substantially mediated by structural disadvantage and collective efficacy; the evidence for a simple, direct causal effect of disorder on serious crime is weak. Second, order-maintenance policing as implemented in New York City and other jurisdictions has been characterized by significant racial disparities and civil liberties concerns that undermine its legitimacy and effectiveness. Third, place-based disorder reduction — through environmental improvements, community engagement, and targeted problem-solving — produces more consistent crime-reduction effects than aggressive citywide enforcement. Fourth, perceptions of disorder are shaped by racial stereotypes and socioeconomic biases, complicating the use of disorder as an objective criterion for police deployment.

These conclusions do not amount to a complete rejection of broken windows theory. The theory correctly identified disorder as a quality-of-life concern worthy of attention, highlighted the importance of informal social control in preventing crime, and stimulated a productive body of research on the relationships among disorder, fear, community organization, and crime. But the strong version of the theory — that disorder directly causes serious crime and that aggressive order-maintenance policing is the optimal strategy for preventing it — has not been supported by the accumulated evidence.

The legacy of broken windows theory illustrates both the power and the danger of criminological ideas. A theory that was presented as an essay in a popular magazine — without formal empirical testing — transformed American policing, justified the aggressive enforcement of minor offenses against millions of predominantly minority citizens, and produced a body of empirical research that ultimately failed to support its core claims. The lesson is that criminal justice policies should be grounded in rigorous evidence rather than in appealing theories, and that the costs of policing strategies — including their effects on racial equity, civil liberties, and community trust — must be weighed against their benefits.

Conclusion

Broken windows theory has been one of the most consequential criminological ideas of the past half century — consequential not primarily for its empirical validity, which remains contested, but for its practical impact on the design and implementation of policing strategies across the United States and internationally. Wilson and Kelling’s 1982 essay provided the intellectual framework for a transformation in American policing that shifted the focus from serious crime to minor disorder, from reactive response to proactive prevention, and from crime fighting to order maintenance — a transformation that produced dramatic increases in misdemeanor enforcement, stop-and-frisk encounters, and the policing of public behavior, with consequences that fell disproportionately on Black, Latino, and economically disadvantaged communities.

The empirical evidence that has accumulated over four decades does not support the theory’s central claim that disorder independently causes serious crime. The weight of the evidence — from Sampson and Raudenbush’s collective efficacy research, Taylor’s longitudinal Baltimore study, Harcourt and Ludwig’s MTO analysis, and the post-SQF natural experiment in New York — points to the conclusion that disorder and crime are co-effects of underlying structural conditions rather than causally linked stages in a developmental sequence. Disorder matters — it affects fear, quality of life, and community well-being — but it does not appear to be the initiating cause of the cascading decline that broken windows theory describes.

The future of disorder-focused crime prevention lies not in the aggressive, citywide order-maintenance policing that broken windows theory inspired but in more targeted, community-engaged, and evidence-based approaches — including problem-oriented policing at hot spots, environmental design improvements, vacant lot remediation, and community development investments that strengthen the collective efficacy of disadvantaged neighborhoods. These approaches retain the valid insight that the physical and social environment affects crime while avoiding the racial disparities, civil liberties violations, and community harm that have characterized the broken windows era of American policing.

References

Ansfield, B. (2020). The broken windows of the Bronx: Putting the theory in its place. American Quarterly, 72(1), 103–127. https://doi.org/10.1353/aq.2020.0005
Braga, A. A., Welsh, B. C., & Schnell, C. (2015). Can policing disorder reduce crime? A systematic review and meta-analysis. Journal of Research in Crime and Delinquency, 52(4), 567–588. https://doi.org/10.1177/0022427815576576
Branas, C. C., South, E., Kondo, M. C., Hohl, B. C., Bourgois, P., Wiebe, D. J., & MacDonald, J. M. (2018). Citywide cluster randomized trial to restore blighted vacant land and its effects on violence, crime, and safety. Proceedings of the National Academy of Sciences, 115(12), 2946–2951. https://doi.org/10.1073/pnas.1718503115
Bratton, W. J., & Kelling, G. L. (2006). There are no cracks in the broken windows. National Review, 58(3), 30–32.
Corman, H., & Mocan, N. (2005). Carrots, sticks, and broken windows. Journal of Law and Economics, 48(1), 235–266. https://doi.org/10.1086/425594
Gau, J. M., & Pratt, T. C. (2008). Broken windows or window dressing? Citizens’ (in)ability to tell the difference between disorder and crime. Criminology and Public Policy, 7(2), 163–194. https://doi.org/10.1111/j.1745-9133.2008.00500.x
Goldstein, H. (1990). Problem-oriented policing. McGraw-Hill.
Golub, A., Johnson, B. D., & Dunlap, E. (2007). The race/ethnicity disparity in misdemeanor marijuana arrests in New York City. Criminology and Public Policy, 6(1), 131–164. https://doi.org/10.1111/j.1745-9133.2007.00426.x
Harcourt, B. E. (1998). Reflecting on the subject: A critique of the social influence conception of deterrence, the broken windows theory, and order-maintenance policing New York style. Michigan Law Review, 97(2), 291–389. https://doi.org/10.2307/1290289
Harcourt, B. E. (2001). Illusion of order: The false promise of broken windows policing. Harvard University Press.
Harcourt, B. E., & Ludwig, J. (2006). Broken windows: New evidence from New York City and a five-city social experiment. University of Chicago Law Review, 73(1), 271–320.
Harcourt, B. E., & Ludwig, J. (2007). Reefer madness: Broken windows policing and misdemeanor marijuana arrests in New York City, 1989–2000. Criminology and Public Policy, 6(1), 165–181. https://doi.org/10.1111/j.1745-9133.2007.00427.x
Keizer, K., Lindenberg, S., & Steg, L. (2008). The spreading of disorder. Science, 322(5908), 1681–1685. https://doi.org/10.1126/science.1161405
Kelling, G. L., & Coles, C. M. (1996). Fixing broken windows: Restoring order and reducing crime in our communities. Free Press.
Kelling, G. L., & Sousa, W. H. (2001). Do police matter? An analysis of the impact of New York City’s police reforms (Civic Report No. 22). Manhattan Institute.
Kohler-Hausmann, I. (2018). Misdemeanorland: Criminal courts and social control in an age of broken windows policing. Princeton University Press.
Levitt, S. D. (2004). Understanding why crime fell in the 1990s: Four factors that explain the decline and six that do not. Journal of Economic Perspectives, 18(1), 163–190. https://doi.org/10.1257/089533004773563485
O’Brien, D. T., Farrell, C., & Welsh, B. C. (2019). Looking through broken windows: The impact of neighborhood disorder on aggression and fear of crime is an artifact of research design. Annual Review of Criminology, 2, 217–236. https://doi.org/10.1146/annurev-criminol-011518-024638
Sampson, R. J. (2012). Great American city: Chicago and the enduring neighborhood effect. University of Chicago Press.
Sampson, R. J., & Raudenbush, S. W. (1999). Systematic social observation of public spaces: A new look at disorder in urban neighborhoods. American Journal of Sociology, 105(3), 603–651. https://doi.org/10.1086/210356
Sampson, R. J., & Raudenbush, S. W. (2004). Seeing disorder: Neighborhood stigma and the social construction of “broken windows.” Social Psychology Quarterly, 67(4), 319–342. https://doi.org/10.1177/019027250406700401
Sampson, R. J., Raudenbush, S. W., & Earls, F. (1997). Neighborhoods and violent crime: A multilevel study of collective efficacy. Science, 277(5328), 918–924. https://doi.org/10.1126/science.277.5328.918
Skogan, W. G. (1990). Disorder and decline: Crime and the spiral of decay in American neighborhoods. Free Press.
Taylor, R. B. (2001). Breaking away from broken windows: Baltimore neighborhoods and the nationwide fight against crime, grime, and fear. Westview Press.
Taylor, R. B. (2015). Community criminology: Fundamentals of spatial and temporal scaling, ecological indicators, and selectivity bias. New York University Press.
Weisburd, D., Hinkle, J. C., Braga, A. A., & Wooditch, A. (2015). Understanding the mechanisms underlying broken windows policing: The need for evaluation evidence. Journal of Research in Crime and Delinquency, 52(4), 589–608. https://doi.org/10.1177/0022427815577837
Wilson, J. Q., & Kelling, G. L. (1982). Broken windows: The police and neighborhood safety. Atlantic Monthly, 249(3), 29–38.
Xu, Y., Fiedler, M. L., & Flaming, K. H. (2005). Discovering the impact of community policing: The broken windows thesis, collective efficacy, and citizens’ judgment. Journal of Research in Crime and Delinquency, 42(2), 147–186. https://doi.org/10.1177/0022427804266544
Zimbardo, P. G. (1969). The human choice: Individuation, reason, and order versus deindividuation, impulse, and chaos. In W. J. Arnold & D. Levine (Eds.), Nebraska Symposium on Motivation (pp. 237–307). University of Nebraska Press.
Dixon, D. (1998). Broken windows, zero tolerance, and the New York miracle. Current Issues in Criminal Justice, 10(1), 96–106. https://doi.org/10.1080/10345329.1998.12036125
Fagan, J., & Davies, G. (2000). Street stops and broken windows: Terry, race, and disorder in New York City. Fordham Urban Law Journal, 28(2), 457–504.
Kubrin, C. E. (2008). Making order of disorder: A call for conceptual clarity. Criminology and Public Policy, 7(2), 203–213. https://doi.org/10.1111/j.1745-9133.2008.00502.x
Reisig, M. D., & Kane, R. J. (Eds.). (2014). The Oxford handbook of police and policing. Oxford University Press.
Sharkey, P., Torrats-Espinosa, G., & Takyar, D. (2017). Community and the crime decline: The causal effect of local nonprofits on violent crime. American Sociological Review, 82(6), 1214–1240. https://doi.org/10.1177/0003122417736289
Welsh, B. C., Braga, A. A., & Bruinsma, G. J. N. (2015). Reimagining broken windows: From theory to policy. Journal of Research in Crime and Delinquency, 52(4), 447–463. https://doi.org/10.1177/0022427815581399