What the Research Says About Broken Windows

What the Research Says About Broken Windows provides the master empirical synthesis for this silo, drawing together the full evidentiary record, survey-based, systematic-observational, quasi-experimental, randomized, city-level, and meta-analytic, that other articles throughout this collection have examined piecemeal within their specific thematic scope. Where Criticisms of Broken Windows Theory organizes the accumulated critical literature by argumentative strand, this article organizes the same underlying body of evidence chronologically and methodologically, tracing how successive generations of research design, each addressing limitations identified in its predecessors, have progressively narrowed and specified what the empirical record actually supports.

What the Research Says About Broken Windows treats the theory’s evidentiary status as a genuinely evolving, rather than settled, question, one whose answer differs considerably depending on which specific component claim, which operational variant, and which outcome measure a given study addresses. This article’s organizing premise, established throughout this silo and stated most explicitly in Core Principles of Broken Windows Theory, is that broken windows theory is not a single, unified, testable hypothesis but a bundle of related propositions, disorder signals weakened control, disorder predicts fear, disorder independently causes serious crime, disorder-focused policing reduces crime, each carrying a distinct and often quite different evidentiary status.

This article proceeds chronologically through four successive waves of research design, survey-based correlational studies, systematic observational studies, quasi-experimental and randomized studies, and meta-analytic syntheses, before concluding with a direct assessment of where the accumulated evidence converges and where genuine scientific disagreement persists.

The First Wave: Survey-Based Correlational Evidence

What Wilson and Kelling Actually Presented

Wilson and Kelling’s (1982) original essay, examined in detail in Wilson and Kelling’s Broken Windows Article, presented remarkably little direct empirical evidence for its central causal claims, relying instead on the Newark Foot Patrol Experiment’s finding that foot patrol improved residents’ perceived safety without measurably reducing recorded crime, a single field demonstration of vandalism escalation from Philip Zimbardo, and narrative illustration through the broken-window metaphor itself. This thin initial evidentiary base meant that the theory’s first genuine empirical tests occurred only after its publication and rapid policy adoption, an unusual sequencing that several critics, examined in Criticisms of Broken Windows Theory, have identified as consequential for how the subsequent research agenda developed.

This sequencing, policy adoption preceding rather than following rigorous empirical validation, shaped the entire subsequent research trajectory this article traces, since researchers testing the theory across the following four decades were not merely evaluating an abstract academic hypothesis but were, in effect, retroactively assessing a policy already deployed at scale across major American cities, a circumstance that gave the research agenda a distinctive urgency and policy relevance while also complicating researchers’ ability to design clean baseline comparisons, since by the time systematic empirical testing began in earnest during the late 1990s, broken windows-informed policing had already been implemented in New York and several other cities for the better part of a decade.

Early Survey-Based Support: Skogan’s Disorder and Decline

Wesley Skogan’s (1990) Disorder and Decline represented the first substantial empirical test, using resident survey data from forty neighborhoods across six American cities to examine associations between perceived disorder and both fear of crime and self-reported victimization. Skogan found generally positive associations supporting the theory’s core predictions, disorder correlated with both heightened fear and, somewhat more weakly, with victimization, findings that became, for nearly a decade, the theory’s primary empirical support base and were cited extensively in policy discussions justifying broken windows policing’s initial municipal adoption.

Skogan’s methodology, however, relied entirely on resident-reported disorder measures rather than independently observed physical conditions, a limitation that later researchers, examined below, identified as a significant source of interpretive ambiguity, since resident perceptions of disorder could themselves be shaped by the very fear and victimization concerns the study sought to explain, raising a circularity concern that systematic observational methods were subsequently developed specifically to address.

Ross and Mirowsky’s (1999) subsequent methodological critique of Skogan’s approach, examined in Physical Disorder and Crime, sharpened this circularity concern by directly comparing resident-reported and independently assessed disorder measures within the same neighborhoods, finding meaningful divergence between the two measurement approaches tied to respondent characteristics including length of residence and homeownership status. This divergence finding, emerging within a decade of Skogan’s original publication, marked the beginning of a sustained methodological reckoning with survey-based disorder measurement that would ultimately motivate the systematic observational research wave discussed below, illustrating how each successive generation of broken windows research has emerged substantially in response to specific, identified limitations in its predecessors rather than as an independent, freestanding line of inquiry.

The Second Wave: Systematic Observational Studies

The Chicago Studies and Collective Efficacy

Sampson, Raudenbush, and Earls’s (1997) Project on Human Development in Chicago Neighborhoods, discussed extensively in Collective Efficacy Theory and Broken Windows, represented a methodological turning point, introducing both validated collective efficacy survey measures and, in Sampson and Raudenbush’s (1999) companion study, trained-observer systematic social observation protocols capable of measuring physical and social disorder independently of resident perception for the first time. This methodological advance allowed researchers to test the theory’s core claims with considerably greater precision than Skogan’s earlier survey-based approach had permitted.

The Chicago project’s scale, spanning 343 neighborhood clusters and combining resident surveys, systematic block-level observation, and longitudinal crime data collected over multiple years, gave researchers analytical resources considerably exceeding what any single prior broken windows study had employed, allowing the project’s investigators to test disorder’s association with crime while simultaneously controlling for the structural neighborhood characteristics, concentrated poverty, residential instability, and immigrant concentration, that earlier, more limited studies could not adequately incorporate into their statistical models.

The Chicago findings proved decisive for the field’s subsequent trajectory: Sampson and Raudenbush (1999) found that systematically observed disorder’s association with subsequent violent crime became statistically weak once collective efficacy and structural neighborhood characteristics were included in the same models, a finding examined at length in Criticisms of Broken Windows Theory that substantially qualified, without eliminating, support for the theory’s central causal claim.

Attenuation Once Structural Controls Are Applied

This attenuation finding has been substantially, though not perfectly, replicated in subsequent research employing comparable systematic observation methodology in other cities, establishing what has become the field’s most consistent generalization: disorder’s association with serious crime, while frequently detectable in simple bivariate analysis, consistently weakens once concentrated disadvantage, residential instability, and collective efficacy are properly accounted for, a pattern holding across the American cities where comparable systematic observation research has been conducted. Sampson and Raudenbush’s (2004) subsequent finding that disorder perception itself carries racialized bias, discussed in Broken Windows Theory and Race, added a further layer of interpretive complexity, suggesting that even the attenuated associations detected in systematic observation studies may partly reflect residual measurement confounds rather than a fully purified environmental signal.

The systematic observation wave’s broader methodological legacy extends beyond its specific attenuation findings to the measurement infrastructure it established for the entire field, since the trained-observer protocols Sampson and Raudenbush developed have since been adapted, refined, and redeployed across dozens of subsequent studies spanning multiple American and international cities, establishing systematic observation as the field’s default methodological standard for disorder measurement in a manner that Skogan’s earlier survey-based approach, despite its own considerable influence, never fully achieved. This methodological standardization has made cross-study comparison within the systematic observation wave considerably more tractable than comparison across the survey-based and systematic-observational waves, since studies employing comparable observation protocols can be more confidently pooled and compared than studies employing the more varied and idiosyncratic survey instruments the earlier research wave relied upon.

The Third Wave: Quasi-Experimental and Randomized Evidence

The Lowell Hot Spots Experiment

Braga and Bond’s (2008) randomized controlled trial in Lowell, Massachusetts, discussed in Broken Windows Theory and Policing, addressed the causal-inference limitations of both the survey-based and systematic-observational research waves by randomly assigning matched disorder hot spots to either intensified order-maintenance intervention or standard patrol, finding statistically significant reductions in both disorder and crime calls for service at treated locations. This randomized design represented a genuine methodological advance, since random assignment eliminates the confounding concern that has complicated interpretation of both earlier research waves, though the study’s geographic scope, individual hot spots within a single city, limits its generalizability to broader, city-wide disorder-policing claims.

The Lowell study’s design also permitted researchers to disaggregate intervention types, distinguishing situational and social-service interventions from conventional enforcement activity, finding that the former produced larger disorder reductions than pure enforcement, an early and methodologically credible signal that would later be substantially reinforced by the meta-analytic evidence discussed in the fourth research wave below, suggesting that the specific mechanism through which disorder-focused intervention operates, problem-solving versus enforcement volume, may matter more for outcomes than the underlying disorder-focused premise itself.

Philadelphia Vacant Lot and Building Studies

Branas and colleagues’ (2011, 2018) Philadelphia studies, discussed extensively in Physical Disorder and Crime, employed difference-in-differences and cluster-randomized designs to test physical disorder remediation specifically, finding statistically significant reductions in nearby gun violence following vacant lot greening and abandoned building remediation. These studies, along with the Lowell experiment, represent the methodologically strongest evidence in the entire broken windows literature, sharing the common feature of testing narrowly scoped, geographically bounded interventions rather than the theory’s broadest, city-wide causal claims, a scope limitation that has shaped how contemporary researchers interpret their policy relevance.

Kondo and colleagues’ (2018) systematic review of neighborhood violence-reduction interventions situated these Philadelphia findings within the broader landscape of place-based crime prevention research, concluding that physical environment interventions targeting vacant land and abandoned structures represented among the most consistently replicated crime-reduction findings across the entire violence-prevention literature, a conclusion the review’s authors explicitly distinguished from the more contested, broader broken windows causal narrative from which such interventions originally drew their conceptual justification, treating physical remediation as an independently validated intervention rather than as evidence requiring endorsement of the theory’s full original causal chain.

City-Level Policy Evaluations and Their Contested Attribution

Parallel to this experimental research, city-level policy evaluations, most prominently Kelling and Sousa’s (2001) and Rosenfeld, Fornango, and Rengifo’s (2007) analyses of New York City’s crime decline discussed in Broken Windows Theory in New York City and Zero Tolerance Policing and Broken Windows, attempted to evaluate broken windows policing’s effects using precinct-level variation in enforcement intensity. Harcourt and Ludwig’s (2006) methodological reanalysis found these associations substantially weakened once regression to the mean and broader comparison samples were incorporated, while Zimring’s (2011) comprehensive historical analysis credited multiple concurrent factors, including increased police staffing and targeted gun enforcement, alongside rather than instead of order-maintenance policing specifically.

Messner and colleagues’ (2007) epidemiological analysis of New York’s homicide decline specifically added a further alternative explanatory factor to this city-level evaluation literature, finding that changes in illegal drug markets, particularly the waning of the crack cocaine epidemic that had driven much of the city’s homicide surge during the late 1980s, appeared to explain a comparably large or larger share of the 1990s decline than targeted policing strategies, a finding that, alongside Levitt’s (2004) national-level analysis identifying demographic and incarceration-rate factors, has collectively established that no single-factor explanation, whether policing-centered or otherwise, can adequately account for New York’s crime decline without acknowledging substantial contribution from multiple concurrent causes operating simultaneously.

The Fourth Wave: Meta-Analytic Synthesis and Natural Experiments

Braga, Welsh, and Schnell’s Systematic Review

Braga, Welsh, and Schnell’s (2015) systematic review and meta-analysis, discussed in Broken Windows Theory and Policing, synthesized findings across the accumulated quasi-experimental and experimental literature, finding a statistically significant but modest overall crime-reduction effect associated with disorder policing interventions, with substantially stronger effects for community-oriented and problem-solving variants than for aggressive, arrest-focused zero-tolerance implementations. This meta-analytic finding, that effectiveness varies considerably by operational variant rather than representing a uniform effect of the underlying theory, has become the field’s most influential contemporary synthesis, informing the variant-specific framing this entire silo has adopted throughout its policing-focused articles.

The methodological significance of this meta-analysis extends beyond its specific effect-size findings to its broader contribution in resolving what had, by the early 2010s, become an increasingly polarized field-wide debate between researchers emphasizing the theory’s supportive evidence and researchers emphasizing its critical literature. By systematically coding included studies for operational variant, Braga, Welsh, and Schnell demonstrated that much of the apparent disagreement in the prior literature reflected genuine underlying heterogeneity in what different studies had actually tested, community-oriented problem-solving interventions in some cases, aggressive zero-tolerance enforcement in others, rather than irreconcilable disagreement about a single, unified phenomenon, a methodological contribution that has shaped how this entire silo organizes its treatment of the theory’s policing applications.

The Post-Floyd Natural Experiment

New York City’s dramatic post-Floyd reduction in stop-and-frisk volume, examined in Stop-and-Frisk and Broken Windows Policing and Broken Windows Theory in New York City, provided an unusually large-scale natural experiment in enforcement intensity, with MacDonald, Fagan, and Geller’s (2016) quasi-experimental analysis finding no significant relationship between the magnitude of precinct-level stop reduction and subsequent crime trends, evidence that has proven influential in the broader debate over whether aggressive enforcement’s crime-control contribution was as essential as its strongest early proponents claimed. Rosenfeld and Fornango’s (2014) related precinct-level analysis similarly found weak and inconsistent associations between stop intensity and subsequent robbery and burglary trends during the tactic’s peak deployment years.

This natural-experiment evidence carries particular methodological weight within the broader research record because it emerged from an involuntary, court-mandated policy reversal rather than from a voluntarily implemented pilot program, reducing the selection-bias concerns that complicate interpretation of the voluntarily adopted disorder-policing initiatives examined throughout the earlier research waves discussed above. Where the Lowell and Philadelphia experiments demonstrated that disorder-focused intervention could produce localized benefits when implemented, the post-Floyd natural experiment demonstrated the complementary and equally important finding that aggressive, city-wide disorder enforcement’s withdrawal did not produce the crime increase its strongest proponents had predicted, together bracketing the plausible range of the theory’s genuine causal contribution from both directions.

Variant-Specific Effect Sizes Across the Accumulated Literature

Considered together, the accumulated quasi-experimental, randomized, and meta-analytic evidence supports a variant-specific rather than unified conclusion regarding broken windows policing’s crime-control effectiveness: geographically targeted, problem-oriented interventions, particularly physical remediation and hot-spots-focused order maintenance, show the most consistent independent effects, while broad, city-wide, discretion-eliminating enforcement regimes show effects that are smaller, less consistent, and more difficult to isolate from concurrent confounding trends, a pattern holding across nearly every methodologically rigorous study this article has reviewed.

This variant-specific conclusion connects broken windows policing’s evidentiary record to the broader hot-spots policing literature, a research tradition sharing substantial methodological and conceptual overlap with the geographically targeted broken windows applications discussed above. Braga, Papachristos, and Hureau’s (2014) updated systematic review of hot-spots policing broadly, encompassing but not limited to disorder-focused interventions specifically, found a consistently significant overall crime-reduction effect across the accumulated experimental and quasi-experimental literature, reinforcing from an independent research tradition the same geographic-targeting principle that the broken windows-specific evidence reviewed in this article supports. Weisburd’s (2015) related theoretical contribution, the “law of crime concentration,” proposed that crime and disorder cluster with remarkable consistency at a small proportion of urban micro-locations across virtually every American city examined, a foundational empirical regularity that helps explain why geographically targeted interventions, whether framed explicitly as broken windows applications or as hot-spots policing more generally, have consistently outperformed broader, undifferentiated citywide strategies throughout the research record this article has traced.

Weisburd and Eck’s (2004) earlier synthesis, reviewing the accumulated evidence on what police interventions generally can achieve for crime, disorder, and fear across the full range of policing strategies rather than disorder-focused approaches specifically, reached a broadly convergent conclusion: focused, geographically or demographically targeted interventions consistently outperformed generalized, undifferentiated patrol and enforcement strategies across virtually every category of policing intervention the authors examined, situating the broken windows-specific pattern this article documents within an even broader empirical regularity spanning the whole of contemporary policing research rather than representing a finding unique to disorder-focused strategies alone.

Synthesis: Where the Evidence Converges and Diverges

Well-Supported Findings

Several specific findings enjoy strong, consistent, cross-study support: the association between disorder and fear of crime, examined in Fear of Crime and the Incivilities Thesis, has been replicated across virtually every study design and city examined in this literature; targeted physical remediation interventions, examined in Physical Disorder and Crime, have produced methodologically credible evidence of independent, geographically localized crime-reduction effects; and community-oriented, problem-solving policing variants, examined in Order Maintenance Policing, have outperformed aggressive zero-tolerance variants on both crime-control and civil liberties dimensions across the meta-analytic evidence.

These well-supported findings share a common structural feature worth naming explicitly: each addresses a comparatively narrow, well-specified causal question, does disorder predict fear, does targeted remediation reduce crime at treated locations, does one policing variant outperform another, rather than the theory’s broadest, most sweeping original formulation. This pattern, that the most credible evidence supports the theory’s narrower, more precisely bounded component claims while its broadest original claims remain considerably more contested, recurs so consistently across the four research waves this article has traced that it constitutes, in its own right, one of the field’s most important methodological lessons regarding how large, ambitious social-scientific theories are best empirically evaluated over time.

Contested and Unresolved Questions

Other components remain genuinely contested: disorder’s independent, city-wide causal contribution to serious violent crime, once structural confounds are properly controlled, remains weakly and inconsistently supported at best; broad, aggressive enforcement’s independent crime-control contribution, once national trends and concurrent policy changes are accounted for, remains difficult to isolate with confidence; and the precise causal direction between disorder and collective efficacy, whether disorder erodes control or merely indicates its prior absence, remains an open question that even the most sophisticated longitudinal designs available have not fully resolved.

A further genuinely unresolved question concerns generalizability across differing urban contexts: nearly all of the methodologically strongest evidence reviewed in this article derives from a comparatively small set of American cities, Chicago, New York, Philadelphia, and Lowell prominent among them, raising the question of whether findings from these specific, extensively studied cities generalize to the much larger population of American cities and neighborhoods that have never been subject to comparably rigorous evaluation. This generalizability gap represents a significant, though rarely emphasized, limitation on how confidently this article’s synthesis can be extended beyond the specific research settings where the underlying studies were actually conducted, a caution worth bearing in mind when translating this article’s conclusions into policy guidance for jurisdictions considerably different from the handful of extensively studied cities this literature has concentrated upon.

Analytical Tables

Table 1. Four Waves of Broken Windows Research by Methodology

Wave	Representative Studies	Methodological Strength	Key Limitation
Survey-based correlational	Skogan (1990)	First systematic multi-city test	Resident-reported measures, circularity risk
Systematic observational	Sampson & Raudenbush (1999, 2004)	Observer-independent disorder measurement	Structural confounds, racialized perception
Quasi-experimental/randomized	Braga & Bond (2008); Branas et al. (2011, 2018)	Strong causal inference at treated sites	Limited geographic generalizability
Meta-analytic/natural experiment	Braga, Welsh & Schnell (2015); MacDonald, Fagan & Geller (2016)	Synthesizes across studies; large-scale natural variation	Effect estimates sensitive to included studies

Table 2. Evidentiary Status of Core Theory Components

Component Claim	Evidentiary Status	Key Supporting/Challenging Studies
Disorder predicts fear of crime	Strong, consistent support	Skogan (1990); Ferraro (1995)
Disorder independently causes serious crime	Weak, inconsistent support	Sampson & Raudenbush (1999); Harcourt (2001)
Physical remediation reduces crime locally	Strong support	Branas et al. (2011, 2018)
Community-oriented policing reduces crime	Moderate to strong support	Braga, Welsh & Schnell (2015)
Aggressive zero-tolerance enforcement reduces crime	Weak, contested support	Harcourt & Ludwig (2006); MacDonald, Fagan & Geller (2016)
Disorder perception is racially neutral	Not supported	Sampson & Raudenbush (2004); Quillian & Pager (2001)

Conclusion

The accumulated empirical record on broken windows theory, spanning four successive waves of increasingly rigorous research design, does not support either the confident causal narrative Wilson and Kelling’s original 1982 essay advanced or the wholesale rejection some of the theory’s most severe critics have proposed. Instead, four decades of research have progressively specified a considerably more precise, variant-specific, and component-specific picture: certain claims, disorder’s association with fear, physical remediation’s localized crime-reduction effect, community-oriented policing’s comparative advantage, have accumulated strong and consistent support, while others, disorder’s independent city-wide causal contribution to serious crime and aggressive enforcement’s necessity for crime control, remain weakly supported or actively contradicted by the strongest available evidence.

This variant-specific and component-specific conclusion, rather than any single verdict on broken windows theory as a whole, represents the field’s most defensible current summary of what the evidence actually shows, a summary this silo’s remaining article, Broken Windows Theory Today, traces into its implications for how the theory’s legacy is understood and applied in contemporary criminal justice policy.

Reviewing this four-decade research trajectory as a whole offers a broader methodological lesson extending beyond broken windows theory specifically: each successive research wave emerged in direct response to a specific, identified limitation in its predecessor, survey-based circularity motivating systematic observation, systematic observation’s structural confounds motivating quasi-experimental and randomized designs, single-study idiosyncrasy motivating meta-analytic synthesis, and voluntary-adoption selection bias motivating attention to the post-Floyd natural experiment. This progressive methodological refinement, rather than representing failure or repeated reversal, exemplifies how a mature empirical research program is supposed to develop, with each wave narrowing uncertainty and sharpening the field’s understanding of precisely which claims the evidence supports, even where that narrowing process ultimately validates a considerably more modest and qualified picture than the theory’s confident 1982 origins first proposed. Researchers approaching new, unresolved questions within this literature, including the generalizability concerns discussed above and the precise causal direction between disorder and collective efficacy, can reasonably expect this same pattern of progressive methodological refinement to continue shaping the field’s understanding for years to come.

References

Braga, A. A., & Bond, B. J. (2008). Policing crime and disorder hot spots: A randomized controlled trial. Criminology, 46(3), 577–607.
Braga, A. A., Papachristos, A. V., & Hureau, D. M. (2014). The effects of hot spots policing on crime: An updated systematic review and meta-analysis. Justice Quarterly, 31(4), 633–663.
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.
Branas, C. C., Cheney, R. A., MacDonald, J. M., Tam, V. W., Jackson, T. D., & Ten Have, T. R. (2011). A difference-in-differences analysis of health, safety, and greening vacant urban space. American Journal of Epidemiology, 174(11), 1296–1306.
Corman, H., & Mocan, N. (2005). Carrots, sticks, and broken windows. Journal of Law and Economics, 48(1), 235–266.
Ferraro, K. F. (1995). Fear of Crime: Interpreting Victimization Risk. State University of New York Press.
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.
Kelling, G. L., & Sousa, W. H. (2001). Do police matter? An analysis of the impact of New York City’s police reforms. Civic Report, 22, Manhattan Institute.
Kondo, M. C., Andreyeva, E., South, E. C., MacDonald, J. M., & Branas, C. C. (2018). Neighborhood interventions to reduce violence. Annual Review of Public Health, 39, 253–271.
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.
MacDonald, J., Fagan, J., & Geller, A. (2016). The effects of local police surges on crime and arrests in New York City. PLOS ONE, 11(6), e0157223.
Messner, S. F., Galea, S., Tardiff, K. J., Tracy, M., Bucciarelli, A., Piper, T. M., Frye, V., & Vlahov, D. (2007). Policing, drugs, and the homicide decline in New York City in the 1990s. Criminology, 45(2), 385–414.
Quillian, L., & Pager, D. (2001). Black neighbors, higher crime? The role of racial stereotypes in evaluations of neighborhood crime. American Journal of Sociology, 107(3), 717–767.
Rosenfeld, R., & Fornango, R. (2014). The impact of police stops on precinct robbery and burglary rates in New York City, 2003–2010. Justice Quarterly, 31(1), 96–122.
Rosenfeld, R., Fornango, R., & Rengifo, A. F. (2007). The impact of order-maintenance policing on New York City homicide and robbery rates: 1988–2001. Criminology, 45(2), 355–384.
Ross, C. E., & Mirowsky, J. (1999). Disorder and decay: The concept and measurement of perceived neighborhood disorder. Urban Affairs Review, 34(3), 412–432.
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.
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.
Sampson, R. J., Raudenbush, S. W., & Earls, F. (1997). Neighborhoods and violent crime: A multilevel study of collective efficacy. Science, 277(5328), 918–924.
Skogan, W. G. (1990). Disorder and Decline: Crime and the Spiral of Decay in American Neighborhoods. Free Press.
South, E. C., Hohl, B. C., Kondo, M. C., MacDonald, J. M., & Branas, C. C. (2018). Effect of greening vacant land on mental health of community-dwelling adults: A cluster randomized trial. JAMA Network Open, 1(3), e180298.
Weisburd, D. (2015). The law of crime concentration and the criminology of place. Criminology, 53(2), 133–157.
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.
Weisburd, D., & Eck, J. E. (2004). What can police do to reduce crime, disorder, and fear? Annals of the American Academy of Political and Social Science, 593(1), 42–65.
Wilson, J. Q., & Kelling, G. L. (1982). Broken windows: The police and neighborhood safety. The Atlantic Monthly, 249(3), 29–38.
Zimring, F. E. (2011). The City That Became Safe: New York’s Lessons for Urban Crime and Its Control. Oxford University Press.