\server05productnCCPP7-2CPP204.txt unknown Seq: 1 24-JUN-08 13:14

POLICY ESSAY

BROKEN WINDOWS: WHY—AND HOW—WESHOULD TAKE THEM SERIOUSLY

WESLEY G. SKOGANNorthwestern University

What is disorder? Wilson and Kelling’s seminal 1982 “broken windows”article introduced the concept, at least in spirit. Although they did not giveit that name, they focused on what we now call social disorder. Theiroriginal list included public gambling, public drinking and urination, streetprostitution, congregations of idle men and bands of youths dressed inapparently gang-related apparel, and activities such as panhandling,disturbing the peace, loitering, and vagrancy. Since then, the list has grownto include truant high-schoolers, squeegee men looking for tips, dumpsterdivers in search of dinner, street preachers armed with bullhorns, “urbancampers” in parks under cardboard tents, people with a “street lifestyle,”the presence of sexually oriented establishments, street harassment ofwomen, open gambling, threatening phone calls, recreational violence inpubs and clubs, and—in the article by Gau and Pratt (2008, this issue)—“noise” and “dogs running at-large.” The title of Wilson and Kelling’sarticle implied that the signs of physical decay needed to be addressed aswell. In various studies, physical disorders have included dilapidation,abandoned buildings, stripped and burned-out cars, collapsing garages,broken streetlights, junk-filled and unmowed vacant lots, litter, garbage-strewn alleys, alcohol and tobacco advertising, graffiti, and the visibleconsequences of vandalism.

Taken as a whole, these items constitute an untidy list. In contrast,criminal codes seem to encompass a more bounded and orderly set ofprohibitions. What the common crimes have in common, however, is onlythat legislators do not like them. Otherwise, they also encompass apotpourri of behaviors, motives, and modalities. The length of thedisorders list is not a problem. The disorder indicators used in anyparticular research or evaluation project should be selected judiciously fortheir relevance to the setting, just as no study includes “all” crimes.Disorders were overlooked for a long time by criminologists because manyhave only a tenuous link to the criminal law or are clearly civil lawmatters. Many evade attention from the police because they lie on thefringes of being “serious enough” to warrant attention; Albert Reiss(1985) dubbed these disorders “soft crimes.” Others items from the listhover in the vicinity of protected rights to speech and assembly; street

VOLUME 7 NUMBER 2 2008 PP 195–202 R

\server05productnCCPP7-2CPP204.txt unknown Seq: 2 24-JUN-08 13:14

196 SKOGAN

preachers, panhandlers, sex shop operators, and the visibly idle fall in thiscategory. Some items were traditionally overlooked because they were notcaptured easily by police information systems. Importantly for the police,the “victims” of many disorders do not experience them as incidents but asconditions. They do not know whether it is a police matter; what theyknow is that they want something done.

Why should we take disorder seriously? Although they have a varyingrelationship with criminality, what disorders have in common is that theyhave demonstrably serious consequences for communities. From thebeginning, the broken windows argument has principally discussed theeffects of disorder at the aggregate level. Wilson and Kelling advanced alist of ideas that researchers have taken up: Disorder undermines thecapacity of neighborhoods to defend themselves; those who can do somove away, many of those remaining withdraw from community publiclife, and everyday uses of public space wither. In their view, the directlycriminogenic effects of disorder stem from the attendant decline ofinformal social control; predatory troublemakers from outside theneighborhood join unruly insiders in creating new possibilities for crime inundefended places. Gambling and drinking can lead to robberies andfights; prostitution and drug sales attract those who prey on the customers.Disorder thus begets an even broader range of problems and can, in shortorder, inundate an area with serious and victimizing crime.

Not all aspects of this broad view have been tested robustly, and littleresearch has been directed at the neighborhood level where it belongs. Ifound that a broad, neighborhood-level disorder index constructed from40 separate community surveys was linked to declining confidence inneighbors, a diminished capacity for collective action, lower levels ofneighborhood satisfaction and a desire to move to the suburbs,perceptions of levels and trends in crime, and fear of crime and robberyvictimization (Skogan, 1990). I was actually agnostic about what pathmodels would reveal about the last possibility because, in my view, theother things on the list of disorder’s consequences would amply justifyspending time on them. Others (see Sampson and Raudenbush, 2004)have pointed out that perceived disorders are linked in complex ways torace and class and that structural features of neighborhoods explain whythey have disorder problems. This finding is not surprising—almosteverything in criminology is correlated strongly with neighborhoodstructural factors. About half the between-neighborhood variance of mydisorder measure could be attributed to poverty, instability, and racialdiversity, but where disorder comes from was not the focus of the project.An independent relationship exists between disorder and the outcomes,including robbery victimization.

\server05productnCCPP7-2CPP204.txt unknown Seq: 3 24-JUN-08 13:14

POLICY ESSAY 197

Whether disorder is somehow “real” or just “in the eyes of thebeholder” and perhaps a measure of intolerance has been the subject ofconsiderable discussion. Gau and Pratt’s article (2008) sidesteps thisquestion by positing that disorders are “real in their consequences”(nothing in the study explained consequences), but the issue cannot beignored. Observational studies of selected disorders find high interraterreliabilities, so disorders are “really there” in the empirical tradition; I amnot sure why we should think that pairs of students who come in for anhour or so are more accurate raters of local conditions than many pairs ofpeople who live there. One limit of observational studies is that they tendto be conducted during daylight hours, which is when things are visibleand it is safer to be looking; thus, observers count many conditions thatfall in the physical disorder category (for an example, see Sampson andRaudenbush, 1999). Residents are present after dark and on Saturdaynights; what we know about the temporal distribution of 911 “disorder”calls (see Weisburd et al., 2006) indicates that they are positioned toobserve wild and wooly events much more often. The match betweensurvey and observational measures—a “multi-method” correlation—isanother test of the “reality” of disorder, which is perhaps the toughestmethodological test in social science (it is a validity test; see Campbell andStanley, 1963). Do the research observers see the same things as residentobservers? To examine this topic, we would like the survey andobservational measures to focus on the same specific disorders, but it isnot always the case. In Sampson and Raudenbush’s (2004) Chicago study,observers looked for many more and really different conditions (including“alcohol/tobacco advertising”) than were included in the survey, and thisinformation was placed in the observational index. The neighborhood-level correlation between the two indices still turned out to be +0.62. In aBaltimore study, the correlation between differing survey andobservational indices of disorder was again +0.62 (Jang and Johnson,2001). Some multi-method studies (not that many exist) report weakerlinks between survey and observational reports of disorder, but more workis called for on this topic. A thorough study would take into account thatthe upper bound on a validity coefficient is the product of the reliabilitiesof the measures being cross-validated, for example.

What can be done about disorder? One unfortunate aspect of theconversation about dealing with disorder has been its stilted character.Discussion of policy alternatives by criminologists has revolved around“disorder policing” or (as in Gau and Pratt’s article [2008]) “brokenwindows policing” concepts that are often conflated with “zero tolerancepolicing.” This information is surprising, for at the same time, mostcriminologists and many sophisticated practitioners would agree thatenforcement-oriented policing is not always the most effective strategy for

\server05productnCCPP7-2CPP204.txt unknown Seq: 4 24-JUN-08 13:14

198 SKOGAN

addressing common crimes. That “we cannot arrest our way out of crimeproblems” is widely understood in this community. The same is certainlytrue when it comes to conditions and events that fall in the disordercategory. Disorder is addressable by the same kitbag of policy tools thatcrime experts have lauded. Besides enforcement, the toolbox includesinteragency coordination; regulatory leverage created by civil statutes; andthe involvement of organized community residents, nonprofit serviceproviders, and the commercial security sector. Like common crimes, thesources and solutions to particular disorders are diverse and highlysituational, and tailored, problem-solving approaches to disorderreduction should be the order of the day. Problem solving is acounterpoint to the traditional model of police work, which usually entailsresponding sequentially to individual events as they are phoned in byvictims. Problem solving, on the other hand, calls for examining patternsof incidents to reveal their causes and to help plan how to deal with themproactively. It is most easily applicable to disorders, many of which—as Inoted above—are best characterized as conditions rather than as events.Perhaps common crimes and disorders fit the same perceptual factorstructure, but it does not mean that their solutions are likely to be “onesize fits all.”

Researchers who observe closely the actual implementation ofcommunity policing understand this dynamic well. When neighborhoodresidents gather to discuss their problems with the police, they inevitablybring a broad spectrum of concerns to the table. They do not make neatlegal or bureaucratic distinctions about who is responsible for what, and itturns out that they are as worried about garbage-strewn alleys and graffition garage doors as they are about burglary and car theft. In 2,500 police-community meetings we examined in Chicago (see Skogan, 2006),residents discussed 11,221 distinct problems. The largest category, whichinvolved 36% of these concerns, was the social disorder category.Residents talked most about loitering, street prostitution, public drinking,and various fears about teenage misconduct. Next on the list (at 24%) wasdiscussion of street drug markets, which is a common crime. But the thirdmost frequent topic of concern, which constituted 12% of the issuesbrought up at beat meetings, was physical decay. The issues that residentstalked about in this category included graffiti, vandalism, abandonedbuildings, abandoned cars, and trash and junk in vacant lots. Even atpolice-sponsored beat meetings, discussion about individually victimizingcrimes like robbery and burglary made up only 9% of the total, which wastied exactly with complaints about parking and traffic problems.

Community policing takes seriously the public’s definition of its ownproblems, which inevitably includes issues that lie outside the traditionalcompetence of the police. An expansion of the police mandate is an

\server05productnCCPP7-2CPP204.txt unknown Seq: 5 24-JUN-08 13:14

POLICY ESSAY 199

almost unavoidable result of involving the public in neighborhood prioritysetting. So when Chicago’s program was developed, planners knew that iftheir officers’ response to community concerns was “that’s not a policematter,” many residents would not show up for the next meeting. Policeknew they needed to have structures in place that would enable them todeal effectively with the issues that concern the public, many of which fallin the disorder category and some of which fall in the traditional domainof other city service agencies. Although at a public meeting officers canlearn that loose garbage and rats in an alley are big issues for the residentswho attend, another agency must deliver the solution to that problem.Interagency collaboration becomes a key component in securingneighborhood safety and security when police organize to address disorderproblems.

As a result of adopting the broken windows theory of neighborhooddecline wholeheartedly (see Skogan, 2006), Chicago police now findthemselves involved in a host of new activities. They orchestrateneighborhood cleanups and graffiti paint-outs by volunteers and cityworkers, distribute bracelets that will identify senior citizens if they falldown, and take note of burned out street lights and trees that needtrimming. Officers drop into stores to ask merchants to display signsrequesting that patrons refrain from giving money to panhandlers. Thepublic steers police toward problems created by the sale of loose cigarettesand individual cans of beer at convenience and grocery stores becausethey encourage loitering and public drinking. Those sales are illegal buttruly soft crimes.

On their side, residents have taken to “positive-loitering” campaigns toretake their streets. These campaigns are efforts to increase the frequencywith which law-abiding residents occupy public spaces to discourage streetprostitutes, loiterers, drinkers, and nascent drug markets. They do so byscheduling dog walks and walking clubs to get out and about during timeswhen these problems are most intense. In a riskier place, police guardresidents at prayer vigils and marches and at barbeque “smoke-outs” inwhich residents congregate in drug and prostitution zones. A popularapproach to addressing troublesome liquor outlets is a “vote dry”referendum. City statutes allow area residents to vote to prohibit the saleof alcohol in a particular electoral precinct (a very small area) or at aparticular address in their precinct. Community groups are well informedabout the mechanics of these referenda, and they constitute such a threatthat they have facilitated the informal but effective resolution of manyproblems with liquor establishments.

Police can facilitate the mobilization of city service agencies. We wouldlike to think those bureaucracies could decipher neighborhood prioritieson their own, but in truth it is largely only the police who remain open

\server05productnCCPP7-2CPP204.txt unknown Seq: 6 24-JUN-08 13:14

200 SKOGAN

24/7, have people walking the streets, and have discovered how effectivelythey can build legitimacy and support among the voters and taxpayers byresponding to their definition of their problems. In Chicago, complaints atbeat meetings trigger graffiti removal missions by the city’s Graffiti Blasterteams, who are so named for their high-pressure soda sprays. Likewise,complaints routinely trigger tree trimming, rat poisoning, clearing andmowing vacant lots, repairing streets and sidewalks, new street signs, andpaint jobs for fire hydrants and public structures. Getting seeminglyabandoned cars towed is a logistical problem, but it is not complicated;severely dilapidated or abandoned commercial and residential buildingsare another story, and addressing them involves many agencies andmoney. Police districts have “problem-buildings officers” who inventorydilapidated and abandoned structures and track down property owners forcivil prosecution over building, health, fire, sanitation, zoning, andbusiness-license violations. Police participate in joint multi-agency teamswith representatives of the city’s buildings, fire, police, revenue, and healthdepartments “swatting” prioritized buildings with documented codeviolations that are then taken over by city prosecutors who work out ofdistrict stations. All these actions have nothing to do with putting anyonein jail. Rather, civil courts and administrative hearings can force buildingowners to repair code violations, install security measures, and securelyboard up buildings. The worst buildings can be demolished on a fast trackif they are unsafe. Landlords can be required to post “no loitering” signs,evict problem tenants, and upgrade exterior lighting. In short, effectiveapproaches to the wide range of problems that make up the disordercategory do not just involve arresting people. In Chicago, at least, asolution for broken windows is to fix them.

REFERENCES

Campbell, Donald T. and Julian C. Stanley1963 Experimental and quasi-experimental designs for research on teaching. In

N. L. Gage (ed.), Handbook of Research on Teaching. Chicago, Ill.: RandMcNally.

Gau, Jacinta M. and Travis C. PrattBroken windows or window dressing? Citizens’ (in)ability to tell thedifference between disorder and crime. Criminology & Public Policy. Thisissue.

Jang, Sung and Byron R. Johnson2001 Neighborhood disorder, individual religiosity, and adolescent use of illicit

drugs: A test of multilevel hypotheses. Criminology 39:109–144.

Reiss, Albert J., Jr.1985 Policing a City’s Central District: The Oakland Story. Washington, D.C.:

National Institute of Justice.

\server05productnCCPP7-2CPP204.txt unknown Seq: 7 24-JUN-08 13:14

POLICY ESSAY 201

Sampson, Robert J. and Stephen W. Raudenbush1999 Systematic social observation of public spaces: A new look at disorder in

urban neighborhoods. American Journal of Sociology 105:603–651.2004 Seeing disorder: Neighborhood stigma and the social construction of

“broken windows.” Social Psychological Quarterly 72:319–342.

Skogan, Wesley G.1990 Disorder and Decline: Crime and the Spiral of Decay in American

Neighborhoods. Berkeley, Calif.: University of California Press.2006 Police and Community in Chicago. New York: Oxford University Press.

Weisburd, David, Laura A. Wyckoff, Justin Ready, John E. Eck, Joshua C. Hinkle,and Frank Gajewski2006 Does crime just move around the corner? A controlled study of spatial

displacement and diffusion of crime control benefits. Criminology44:549–592.

Wilson, James Q. and George L. Kelling1982 Broken windows. Atlantic Monthly (March):29–38.

Wesley G. Skogan, Ph.D., has focused his research on victimization, fear of crime, theimpact of crime on communities, public involvement in crime prevention, and policing.His books on Chicago police include Police and Community in Chicago (2006), On theBeat: Police and Community Problem Solving (1999), and Community Policing, ChicagoStyle (1997). In 2003 he edited Community Policing: Can It Work?, which is a collectionof original essays on innovation in policing. His 1990 book, Disorder and Decline, wonthe 1991 Distinguished Scholar Award of the Section on Crime and Deviance of theAmerican Sociological Association. Skogan has been a technical consultant for theHome Office Research Unit of Great Britain and a senior fellow of the Open SocietiesInstitute. Earlier he spent two years at the National Institute of Justice as a visitingfellow. He is also a fellow of the American Society of Criminology. In the 2000s, hechaired a Committee on Police Policies and Practices for the National Research Coun-cil. He was a coauthor of the committee report, which appeared as a book, Fairness andEffectiveness in Policing: The Evidence from National Academies Press, in 2004.