Collective Efficacy Theory and Broken Windows examines the relationship between Robert Sampson and Stephen Raudenbush’s collective efficacy theory and Wilson and Kelling’s broken windows theory, two of the most influential neighborhood-level explanations of urban crime developed in the late twentieth century. Collective efficacy theory, developed through the Project on Human Development in Chicago Neighborhoods, proposed that a neighborhood’s crime rate depends less on visible physical or social disorder than on residents’ shared willingness to intervene on behalf of the collective good, combined with mutual trust among neighbors. This article traces how collective efficacy research emerged partly as an empirical test of broken windows theory, examines the extent to which its findings support, qualify, or challenge the disorder-crime causal chain, and situates the two theories’ subsequent, sometimes uneasy coexistence within contemporary criminological explanation of neighborhood crime.
Collective Efficacy Theory and Broken Windows addresses what has become one of the most consequential empirical episodes in the broken windows literature. When Sampson, Raudenbush, and Earls (1997) and later Sampson and Raudenbush (1999) tested disorder’s relationship to crime using more rigorous measurement than Wilson and Kelling’s original essay had employed, they found that collective efficacy, not disorder itself, carried the stronger and more consistent independent association with neighborhood violent crime, a finding widely interpreted as a significant qualification, though not an outright refutation, of the original broken windows causal chain discussed in Wilson and Kelling’s Broken Windows Article and Core Principles of Broken Windows Theory.
This article should be read alongside Criticisms of Broken Windows Theory, which situates the collective efficacy findings within the broader empirical critique of the theory, and Physical Disorder and Crime and Social Disorder and Crime, which examine how subsequent researchers have attempted to reconcile disorder-based and collective-efficacy-based explanations of neighborhood crime rather than treating them as strictly competing accounts.
Origins of Collective Efficacy Theory
The Project on Human Development in Chicago Neighborhoods
Collective efficacy theory emerged from the Project on Human Development in Chicago Neighborhoods, a large-scale, multi-method longitudinal study of Chicago neighborhoods launched in the early 1990s and designed to examine how neighborhood social processes shaped a range of developmental and behavioral outcomes, including criminal offending and victimization (Sampson, Raudenbush, & Earls, 1997). The project’s scale, encompassing systematic social observation of thousands of city blocks, resident surveys across hundreds of neighborhood clusters, and longitudinal crime data, gave researchers measurement resources considerably beyond what Wilson and Kelling’s original impressionistic account, or even Skogan’s (1990) earlier survey-based research, had employed.
Sampson and his collaborators designed the project partly in explicit response to the broken windows hypothesis, seeking to test whether disorder’s association with crime, documented in earlier survey research, would hold once more rigorous, observationally grounded measures of both disorder and neighborhood social processes were employed simultaneously in the same statistical models (Sampson & Raudenbush, 1999). This design choice reflected a broader methodological ambition within 1990s criminology to move beyond single-variable explanations of neighborhood crime toward multivariate models capable of adjudicating between competing theoretical accounts.
The project also drew on an earlier tradition of research examining informal surveillance and street-level guardianship, including Bellair’s (1997) analysis of how the frequency of informal social interaction among neighbors, rather than formal organizational participation, predicted neighborhood crime rates, and Morenoff, Sampson, and Raudenbush’s (2001) subsequent extension of the collective efficacy framework to explicitly model spatial dependence, finding that collective efficacy in surrounding neighborhoods, not merely a given neighborhood’s own collective efficacy, predicted local homicide rates. This spatial extension proved theoretically significant because it implied that neighborhood crime could not be fully explained, or effectively intervened upon, by treating each neighborhood as a bounded, independent unit, a limitation that had also constrained much of the earlier disorder-focused research reviewed in Physical Disorder and Crime.
Defining Collective Efficacy
Sampson, Raudenbush, and Earls (1997) defined collective efficacy as the combination of social cohesion among neighbors and their shared willingness to intervene on behalf of the common good, particularly in response to observed deviance such as children skipping school, spray-painting a building, or disrespecting adults. This definition drew on Albert Bandura’s (1997) individual-level concept of self-efficacy, extending it to the neighborhood level as a property of collective social relations rather than individual psychological disposition, and measured it through validated survey scales asking residents about their willingness to intervene in specific hypothetical scenarios and their trust in neighbors.
The collective efficacy construct was explicitly designed to capture the informal social control mechanism that broken windows theory had invoked only implicitly, giving researchers, for the first time, a directly measured variable representing the theoretical mechanism, weakened informal control, that Wilson and Kelling’s essay had proposed disorder erodes, rather than requiring researchers to infer informal control’s presence or absence indirectly from disorder levels themselves.
Subsequent psychometric work validated the collective efficacy scale across diverse urban contexts beyond Chicago, including Browning, Feinberg, and Dietz’s (2004) research examining how collective efficacy interacted with neighborhood social networks in Columbus, Ohio, and finding that dense local networks did not automatically translate into higher collective efficacy, and in some circumstances, particularly where local networks connected residents to offenders or facilitated illicit exchange, could instead undermine it, a nuance Browning and colleagues termed the “paradox of social organization.” This finding qualified the collective efficacy framework’s initial, comparatively optimistic assumption that stronger neighborhood social ties would uniformly predict lower crime, showing instead that the content and orientation of social ties, not merely their density, determined their crime-relevant consequences.
Empirical Findings and Their Relationship to Broken Windows Theory
Collective Efficacy as the Stronger Predictor
Sampson, Raudenbush, and Earls’s (1997) headline finding, based on multilevel models spanning 343 Chicago neighborhood clusters, was that collective efficacy was strongly and independently associated with lower rates of neighborhood violence, and that this association held after controlling for a range of structural neighborhood characteristics, including concentrated disadvantage, residential instability, and immigrant concentration. When Sampson and Raudenbush (1999) subsequently incorporated systematically observed physical and social disorder into comparable models, they found that disorder’s association with violent crime was substantially weaker than collective efficacy’s association and became statistically insignificant in several model specifications once collective efficacy was included as a covariate.
This pattern of findings has been widely interpreted, including by Sampson and Raudenbush themselves, as evidence that collective efficacy represents the more fundamental causal mechanism linking neighborhood social conditions to crime, with disorder functioning, at most, as a correlate or partial indicator of weak collective efficacy rather than as an independent causal force in its own right (Sampson & Raudenbush, 1999). This interpretation directly challenges the specific causal sequence Wilson and Kelling’s original essay proposed, in which disorder itself, rather than an underlying deficit in collective efficacy, was posited as the proximate trigger of neighborhood decline.
Disorder as Indicator Versus Disorder as Cause
A central interpretive question raised by the collective efficacy findings concerns whether disorder should be understood as a cause of weakened informal control, as Wilson and Kelling proposed, or instead as a symptom or indicator of pre-existing weak collective efficacy, a reversal of the original theory’s causal direction. Sampson and Raudenbush’s (1999) findings are consistent with either interpretation, since their cross-sectional and short-term longitudinal data could not fully establish whether disorder preceded declines in collective efficacy or instead reflected collective efficacy deficits that predated the observed disorder.
Subsequent researchers examining this ambiguity have generally favored the indicator interpretation, arguing that disorder and collective efficacy likely arise jointly from shared underlying structural conditions, particularly concentrated poverty and residential instability, rather than standing in a simple one-directional causal relationship in either direction (Sampson, 2012). This joint-causation interpretation, elaborated most fully in Sampson’s (2012) Great American City, represents a significant theoretical departure from Wilson and Kelling’s original, more linear disorder-to-crime causal narrative.
Resolving this directional ambiguity has proven difficult even with the Chicago project’s relatively rich longitudinal data, since collective efficacy itself changes slowly over time within neighborhoods, limiting researchers’ ability to detect short-term reciprocal effects between disorder and collective efficacy using the multi-year survey intervals available in the original study design. Some researchers have proposed that the disorder-collective efficacy relationship may operate as a genuinely reciprocal, self-reinforcing cycle rather than a one-directional process in either direction, with modest initial declines in collective efficacy permitting some disorder to accumulate, and that accumulated disorder then further eroding collective efficacy through the resident-withdrawal mechanism discussed in Core Principles of Broken Windows Theory, producing a downward spiral that neither purely structural nor purely disorder-focused interventions could fully interrupt on their own.
Reconciling the Two Theories
Complementary Rather Than Competing Accounts
Despite the apparent tension between collective efficacy’s empirical dominance and broken windows theory’s original disorder-centered claims, several scholars have argued that the two theories are better understood as complementary rather than strictly competing accounts of neighborhood crime (Sampson, 2012; Taylor, 2001). On this reading, collective efficacy represents the underlying social-structural capacity for informal control, while disorder represents one observable manifestation, among several, of that capacity’s presence or absence, meaning that policies targeting disorder reduction might still contribute to crime prevention insofar as they support or signal a neighborhood’s collective efficacy, even if disorder is not itself the primary causal engine Wilson and Kelling’s essay proposed.
This complementary framing has practical implications for policy design, suggesting that interventions combining physical disorder remediation with deliberate efforts to build resident social cohesion and collective intervention capacity, such as community organizing and resident association support, may prove more effective than disorder-focused policing alone, a possibility examined further in Alternatives to Broken Windows Policing.
Warner’s (2003) research on the role of what she termed “attenuated culture,” weakened normative consensus about appropriate behavior within disadvantaged neighborhoods, offered a further reconciliation pathway, proposing that concentrated disadvantage undermines both residents’ shared normative expectations and their collective efficacy simultaneously, with visible disorder emerging as one observable consequence of this joint erosion rather than as its independent cause. This account, like Sampson’s (2012) joint-causation framework, treats disorder as diagnostically useful, a visible marker practitioners can use to identify neighborhoods likely suffering broader social-organizational deficits, while stopping short of endorsing disorder itself as the primary causal lever broken windows theory’s original formulation proposed.
Persistent Points of Theoretical Tension
Notwithstanding these reconciliation efforts, genuine theoretical tension remains between the two frameworks regarding the appropriate target of policy intervention. Broken windows theory’s original formulation directs attention and resources toward visible disorder itself, readily addressed through policing and municipal remediation, while collective efficacy theory directs attention toward the more diffuse, harder-to-engineer social relations of trust and mutual obligation among neighbors, a target considerably less amenable to direct police or municipal intervention (Sampson, 2012). This difference in actionability has practical consequences: order-maintenance policing strategies, discussed in Order Maintenance Policing, could be implemented relatively quickly following the 1982 essay’s publication, while collective-efficacy-building interventions have generally required longer-term, more resource-intensive community organizing efforts with less immediate, visible results.
This tension has also shaped how the two theories have fared in municipal budget and political priority-setting, since order-maintenance policing draws on existing police department budgets and personnel and produces readily quantifiable output measures, such as citations issued or arrests made, whereas collective-efficacy-building interventions typically require new, non-policing funding streams and produce outcomes, changes in resident trust and willingness to intervene, that are considerably harder to measure and report to funders or elected officials on a comparable timescale. This asymmetry in political and administrative convenience, independent of the underlying evidence regarding each approach’s effectiveness, has plausibly contributed to order-maintenance policing’s more rapid and widespread municipal adoption relative to collective-efficacy-oriented interventions, despite the latter’s stronger empirical support in the research reviewed above.
Measurement and Methodological Legacy
The Systematic Social Observation Toolkit
Beyond its substantive findings, the Project on Human Development in Chicago Neighborhoods left a lasting methodological legacy through the systematic social observation protocol Sampson and Raudenbush (1999) developed to measure physical and social disorder independently of resident perception. This protocol, involving trained observers coding standardized checklists while traversing every face block within sampled neighborhood clusters, has since been adapted and redeployed in numerous subsequent studies across American and international cities seeking to measure disorder without relying on potentially biased resident self-report, addressing the perceptual confound discussed in Physical Disorder and Crime.
The observation toolkit’s influence extends beyond disorder measurement specifically; its emphasis on directly observed, geographically fine-grained neighborhood data helped establish a broader methodological standard in urban criminology favoring observational and administrative measures over survey-based proxies wherever feasible, a standard reflected in subsequent large-scale administrative data research such as O’Brien, Sampson, and Winship’s (2015) analysis of Boston’s municipal service request records.
Extending Collective Efficacy Measurement Beyond Chicago
Following the original Chicago findings, researchers replicated and extended collective efficacy measurement to numerous other cities and, eventually, other countries, generally finding the construct’s basic predictive relationship with crime to hold, though with some variation in effect size across contexts. Wickes, Hipp, Sargeant, and Homel’s (2013) study of collective efficacy in Brisbane, Australia, found a broadly similar negative association between collective efficacy and neighborhood crime, while also identifying cross-national differences in how collective efficacy related to residential stability, suggesting that the construct’s antecedents, though not necessarily its core crime-suppressing function, vary somewhat by national housing and welfare context.
Silver and Miller’s (2004) extension of collective efficacy research to individual-level violence outcomes among a sample of psychiatric patients further demonstrated the construct’s applicability beyond aggregate neighborhood crime rates, finding that individuals residing in low-collective-efficacy neighborhoods exhibited higher rates of violent behavior independent of their individual clinical risk factors, reinforcing collective efficacy’s status as a genuinely contextual, neighborhood-level effect rather than merely an aggregation of individual-level risk factors correlated with neighborhood of residence.
Contemporary Research Integrating Both Frameworks
Combined Models of Disorder and Collective Efficacy
More recent research has moved toward integrated statistical models examining disorder and collective efficacy simultaneously across larger, more geographically diverse samples than the original Chicago-based studies. These integrated models have generally replicated collective efficacy’s stronger independent association with serious violent crime while finding more modest but still detectable independent associations for disorder, particularly when disorder is measured through administrative or systematic observational data rather than resident survey reports vulnerable to the perceptual biases discussed in Physical Disorder and Crime (Sampson & Raudenbush, 2004; O’Brien, Sampson, & Winship, 2015).
Wickes and colleagues’ (2013) Brisbane research contributed a further refinement to these integrated models by proposing that collective efficacy itself should be understood as task-specific rather than as a single undifferentiated capacity, distinguishing neighborhoods’ willingness to intervene in violence and serious crime from their willingness to intervene in more mundane civic problems such as littering or unkempt properties, and finding that these two forms of task-specific efficacy did not always move together within the same neighborhood. This task-specificity finding suggests that a neighborhood strong in general social cohesion may nonetheless underperform on the specific dimension of efficacy most relevant to violent crime prevention, a nuance with direct implications for how community-based violence-prevention programs, discussed in Alternatives to Broken Windows Policing, should be designed and targeted.
Implications for Contemporary Policy Debate
The accumulated evidence bearing on the relationship between collective efficacy and disorder has informed contemporary policy debates favoring investment in community-based violence-prevention strategies, such as Cure Violence and related community mediator models, alongside or in place of disorder-focused policing strategies, on the theory that such investments more directly target the collective efficacy mechanism that current evidence identifies as the stronger causal driver of neighborhood violence (Sampson, 2012). These alternative approaches are examined in greater detail in Alternatives to Broken Windows Policing, which surveys the broader landscape of non-policing responses to neighborhood disorder and violence that have gained policy traction partly in response to the collective efficacy findings discussed in this article.
The collective efficacy framework has also informed how funders and municipal agencies evaluate community-based violence-prevention proposals, with several major philanthropic and federal grant programs since the 2010s explicitly requiring applicants to articulate a collective-efficacy-building theory of change, distinct from a purely disorder-remediation or enforcement rationale, as a condition of funding. This shift in grant-making criteria reflects the collective efficacy research program’s broader influence on how the criminal justice policy field conceptualizes neighborhood-level violence prevention, moving institutional investment gradually away from the environmental and enforcement-focused interventions Wilson and Kelling’s original essay emphasized and toward the social-relational interventions the collective efficacy research identifies as more fundamental to sustained crime reduction.
Analytical Tables
Table 1. Collective Efficacy Theory and Broken Windows Theory Compared
| Dimension | Broken Windows Theory | Collective Efficacy Theory |
|---|---|---|
| Originating scholars | Wilson & Kelling (1982) | Sampson, Raudenbush & Earls (1997) |
| Primary causal variable | Visible physical and social disorder | Social cohesion and shared willingness to intervene |
| Measurement approach | Impressionistic description; later systematic observation | Validated multi-item resident survey scales |
| Primary policy lever | Police order-maintenance enforcement | Community organizing, resident cohesion-building |
| Empirical support for crime association | Mixed, attenuated by structural controls | Strong, consistently replicated |
| Actionability by police alone | High | Low |
Table 2. Key Studies in the Collective Efficacy and Broken Windows Research Program
| Study | Sample | Key Contribution |
|---|---|---|
| Sampson, Raudenbush & Earls (1997) | 343 Chicago neighborhood clusters | Establishes collective efficacy as a validated, independently measured construct |
| Sampson & Raudenbush (1999) | Chicago blocks, systematic observation | Finds disorder’s crime association attenuated by collective efficacy |
| Sampson & Raudenbush (2004) | Chicago residents and observers | Documents racialized bias in resident-perceived disorder |
| Sampson (2012) | Synthesis across PHDCN studies | Proposes joint-causation reconciliation of disorder and collective efficacy |
| O’Brien, Sampson & Winship (2015) | Boston administrative records | Extends measurement to large-scale administrative data |
Conclusion
Collective efficacy theory represents the most methodologically rigorous and empirically influential challenge to broken windows theory’s original causal claims, not because it denies any relationship between neighborhood disorder and crime, but because it identifies a more fundamental underlying mechanism, residents’ shared capacity and willingness to exercise informal social control, that better predicts serious violent crime than disorder measures alone. The relationship between the two theories remains genuinely contested: some scholars treat collective efficacy as having effectively superseded disorder as the field’s leading neighborhood-level crime explanation, while others treat the two frameworks as complementary accounts of a shared underlying social process operating at different levels of observability.
For practitioners and policymakers, the collective efficacy findings counsel caution against treating disorder reduction as a sufficient crime-prevention strategy in its own right, suggesting instead that disorder-focused interventions are most likely to prove effective when paired with, or understood as contributing to, the harder and slower work of building genuine neighborhood social cohesion, a conclusion that has shaped the broader turn toward community-based and collective-efficacy-oriented crime prevention strategies examined throughout the remainder of this silo.
Related Articles
- Core Principles of Broken Windows Theory
- Fear of Crime and the Incivilities Thesis
- Criticisms of Broken Windows Theory
- Alternatives to Broken Windows Policing
- What the Research Says About Broken Windows
References
- Bandura, A. (1997). Self-Efficacy: The Exercise of Control. W. H. Freeman.
- Bellair, P. E. (1997). Social interaction and community crime: Examining the importance of neighbor networks. Criminology, 35(4), 677–704.
- Browning, C. R., Feinberg, S. L., & Dietz, R. D. (2004). The paradox of social organization: Networks, collective efficacy, and violent crime in urban neighborhoods. Social Forces, 83(2), 503–534.
- Morenoff, J. D., Sampson, R. J., & Raudenbush, S. W. (2001). Neighborhood inequality, collective efficacy, and the spatial dynamics of urban violence. Criminology, 39(3), 517–558.
- O’Brien, D. T., Sampson, R. J., & Winship, C. (2015). Ecometrics in the age of big data: Measuring and assessing “broken windows” using large-scale administrative records. Sociological Methodology, 45(1), 101–147.
- 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.
- 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.
- Silver, E., & Miller, L. L. (2004). Sources of informal social control in Chicago neighborhoods. Criminology, 42(3), 551–583.
- 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, Fear, and Decline. Westview Press.
- Wickes, R., Hipp, J. R., Sargeant, E., & Homel, R. (2013). Collective efficacy as a task specific process: Examining the relationship between social ties, neighborhood cohesion, and the capacity to respond to violence, delinquency, and civic problems. American Journal of Community Psychology, 52(1–2), 115–127.
- Warner, B. D. (2003). The role of attenuated culture in social disorganization theory. Criminology, 41(1), 73–98.
- Wilson, J. Q., & Kelling, G. L. (1982). Broken windows: The police and neighborhood safety. The Atlantic Monthly, 249(3), 29–38.