Ecology of Crime

Within cities, there are safe neighborhoods and crime-ridden ones; even within crime-ridden neighborhoods, there are safe street-blocks—the two sides of the street bounded by the two cross streets—and dangerous ones. What do these patterns look like, and how are they to be explained? Most of the work in this area has examined neighborhood-to-neighborhood variations, although some have considered block-to-block differences (Taylor et al., 1984). Further, the bulk of the work has focused on social, economic, and cultural factors, although physical design features including landuse mix and design features linked to territorial functioning are relevant as well.

Many of the explanatory models used here have relied on a family of loosely associated perspectives on attributes of community including social problems, called human ecology (Hawley, 1981). These views seek to explain geographic variation in those attributes by concentrating on features of the immediate surround, whether that be streetblock, neighborhood, or city sector, and the connections between that surround and the broader geographic arena. Three fundamental premises of this family of perspectives are that place-to-place differences in racial and ethnic composition, socioeconomic status, and stability and family structure arise from broader dynamics at work in the larger spatial context; that those place-to-place differences in turn simultaneously instigate and reflect local, face-to-face and small group social dynamics; and that those local dynamics link to a wide range of crime-related outcomes such as delinquency rates, local crime rates, and fear of crime and related reactions to crime.

In 1925 Sir Cyril Burt, a British psychologist, published The Young Delinquent. He looked up the addresses of boys and girls reported as "industrial school cases" in London. Then he looked up where they lived, and made up a delinquency rate. Delinquency rates were highest in the areas right near the central business district (CBD), and declined as one moved outward. In addition, the areas of highest delinquency were also the areas of highest poverty. Burt concluded a relationship existed between social class and delinquency. Furthermore, even though his data were cross-sectional, he concluded that the relationship was causal. Later research of individuals continues to find connections between delinquency and social class (e.g., Hindelang et al.). But this does not mean the relationship holds at the individual level—to presume so is to commit the ecological fallacy.

Sociologists at the University of Chicago in the first half of this century investigated a wide array of urban social problems: delinquency, petty theft, dance halls, gambling, and immigrants' "culture shock," to name a few.

Two of these sociologists, Clifford Shaw and Henry McKay, investigated delinquency. They collected data not only from Chicago, but from other cities as well: Philadelphia, Richmond, Cleveland, Birmingham, Denver, and Seattle. Shaw and McKay went to juvenile courts and collected data about the number of juveniles who had been adjudicated delinquent. They were able to construct delinquency rates by posing the question, for every one thousand youths living in the community between the ages of nine and fifteen, how many had officially been adjudicated delinquent by the court? They also constructed rates using other spatial units, such as one square mile areas.

As had Burt, Shaw and McKay found higher delinquency rates closer to the center of the city, the central business district (CBD), than they did further away from the center city. Indeed, they observed that the further away a community was from the center city, the lower its delinquency rate. This pattern appeared not just in Chicago, but in each of the other cities they examined as well.

As is often time in cities, spatial differences link to social and economic differences. At the time Shaw and McKay were writing, populations were increasing in older cities. This "engine" of city growth led to economic differences across communities at varying distances from the city center. More specifically, because of city growth the CBD was expanding to keep up and "serve" the growth in the broader city. This, of course, had happened in the past as well. Given this historical and ongoing pattern, more desirable locations were always at the outer edge of the expanding city. Land use closer to the city center was often converted to nonresidential land uses such as large industries, stockyards (in the case of both Chicago and Baltimore), and large commercial concerns.

Not only were more central locations less desirable per se, they also were the sites of older housing. For the most part, older housing is also more worn-out housing. Given these less desirable locations, and the more dilapidated housing stock, housing in these areas tended to be cheaper. As prices shifted so too did the types of households living there. Poorer households were more likely to locate close to the city center, where housing was cheapest. Further away one would find housing occupied by low wage or blue-collar workers. More distant, one would find middle-income households. And finally, even further away, in an outer-city or perhaps in a more distant suburban location, one would find the highest income households.

These economic differences in house values and rents were exacerbated by the threat of invasion from the expanding CBD. People were constantly trying to "trade up" in their housing anyway, and move to a slightly better location. But since the CBD was growing at the time, residents from each inner zone would be "invading" the zones just beyond. In the innermost zone, the residential areas were in transition, converting from residential to commercial or industrial. This zone was thus labeled the transition zone. These impending changes led those residents who could get out to do so, those who owned properties there to stop maintaining them, and to maximize their return by converting these units to apartments. Left living in these sites were low-income individuals and households that could not afford housing anywhere else. The residential environment there was rather chaotic.

Linked to the economic differences were ethnic ones. It is generally true, with some exceptions (Massey and Denton), that the newest immigrants to a city make up predominantly lower-income households. This is still true today in large U.S. cities even though the immigrant groups in question are different now than they were then. Consequently, many members of these immigrant groups, when they first arrived in U.S. cities, were limited to central-city, low-income neighborhoods where housing was cheap.

In short, Shaw and McKay's basic model was an economic one; location-based dynamics were set in motion based on the socioeconomic status of the group in question. The physical dilapidation of an area matched the segregation of the population on an economic basis. Given the ethnic heterogeneity in these more dilapidated areas, and shorter tenures, supervision of juveniles was more lax, willingness to reprimand others' children was weaker, and delinquency was higher (Maccoby et al.).

The spatial pattern described above has shifted markedly in large cities in the post–World War II era:

1. Centralized city planning increased in the years following World War II. Urban renewal initiatives destroyed vast tracts of older, worn-out housing in older cities, and replaced them with large numbers of public-housing communities. Many of those displaced from older "slum" locations lost many friends in the process (Frey; Gans). The siting of these communities influenced the surrounding locations, sometimes destabilizing them.
2. Suburbanization increased as federal highway initiatives, especially under President Dwight D. Eisenhower, provided drastically improved road access to cities.
3. But for a number of reasons, the suburbanization of African American households proceeded more slowly than the suburbanization of white households (Massey et al.). Consequently, the larger, older cities themselves became increasingly African American in composition.
4. Passage of various fair housing laws, and related court cases in the 1950s and 1960s increased African Americans' access to housing. In cities where African Americans had historically been limited to specific sections of the city, pent-up demand resulted in rapid racial turnover in large numbers of neighborhoods.

Since about 1970 additional changes in cities have further modified the spatial pattern described above. Most importantly, large numbers of manufacturing jobs have left, migrating from central city locations first to southern locations, then abroad, making it increasingly difficult for those with relatively low education levels to secure employment. Receiving more media attention than has perhaps been warranted given the relatively small number of locations where it has occurred, central city neighborhoods in many urban locations have become partially gentrified. Lower-income households were partially replaced by middle or upper income households that moved in and improved the housing stock.

Given these shifts in cities since around 1950, one would not necessarily expect to see the same spatial pattern for delinquency rates, or crime rates, as were reported for the years prior to World War II. Nonetheless, one still might expect community characteristics to link to these outcomes in a similar way.

At the heart of the human ecological model of offense and delinquency rates is a constellation of processes: social disorganization. Its opposite is collective efficacy. A locale is socially disorganized if several things are true: residents do not get along with one another; residents do not belong to local organizations geared to bettering the community and thus cannot work together effectively to address common problems; residents hold different values about what is and what is not acceptable behavior on the street; and residents are unlikely to interfere when they see other youths or adults engaged in wrongdoing (Bursik, 1988).

By contrast, if collective efficacy is high in a locale, residents will work together on common, neighborhood-wide issues, will get along somewhat with one another, and will take steps to supervise activities of youth or teens taking place in the immediate locale. These outcomes link to organizational participation ("Do you belong to the local improvement association? Does your neighbor?"); informal social control ("If your neighbor saw a young teen spray painting the side of a building about midnight, would he do something about it?"); and local social ties based on propinquity ("How many of the people living on your block do you know by name? How many can you recognize when you see them? If you needed to borrow a tool, could you do so from a close neighbor?").

Researchers have suggested that three levels of resident-based control shape the level of social disorganization versus collective efficacy in a locale (e.g., Bursik and Grasmick). Private control refers to dynamics within families and between close friends. If Junioretta extorts school lunch money from two other neighbors while walking to school, and her parents find out about it, will they punish her appropriately? Parochial control refers to supervisory efforts made by neighbors and acquaintances. If a neighbor while gardening out back sees Junioretta walking down the alley threatening two other children and demanding their lunch funds, will she grab Junioretta by the ear and walk her home to her dad, or will she, the neighbor, just shrug her shoulders and go about planting her tomatoes? How much parochial control is exercised varies from block to block in a neighborhood. Public control refers to the neighborhood leadership's ability to garner resources from public and private agencies outside the neighborhood. Can the community association's leaders effectively lobby city hall for resources for neighborhood improvements and programs? For example, can they obtain funding for more school crossing guards on well-traveled routes leading to and from the local school? Can they work collaboratively with other neighborhood organizations on issues affecting their part of town?

High delinquency rates occurred in low income, ethnically heterogeneous, unstable locations because those ecological characteristics made social disorganization more likely. In lower income locales residents' concerns are more spatially circumscribed than in higher income locales (Taylor, 1988). In some low-income neighborhoods residents only feel safe within their own dwelling. As ethnic heterogeneity increases, it becomes increasingly difficult for residents to "decode" what other residents are doing. Increasing intercultural distance and perhaps language barriers make it harder to figure out what is going on. As instability increases, residents have less time to get to know their neighbors; it is harder to figure out who "belongs" on the block and who does not belong.

In other words, these structural attributes of the community either increase or decrease the chances that residents would exert some control over what took place in their community; these dynamics in turn would influence outcomes like delinquency, the local offending rate, and local victimization rates. Note that social disorganization mediates the impacts of community structure on the outcomes. It represents a crucial link connecting community fabric with the outcomes. It does appear, however, that community fabric; although it affects social disorganization, continues to exert an independent influence on outcomes like delinquency, victimization, and offending (Veysey and Messner). In short there are structural causes of these community-level differences beyond differences in social disorganization or collective efficacy.

Social disorganization is likely to be strongest, and collective efficacy weakest, when a community is in the midst of an invasion-succession cycle. In such a cycle, a neighborhood "turns over," with one type of resident replacing another. In the midst of such a cycle residents are unlikely to know their neighbors, and the local population will be quite heterogeneous in makeup.

Neighborhood residents are always changing: people move in and people move out. But if the two rates are roughly matched, and if the volume is relatively modest, and if those moving in are sociodemographically similar to those moving out, then the neighborhood is stable (Ahlbrandt and Cunningham). But if the volume of in-movers increases beyond a relatively low rate, and if the in-movers are sociodemographically dissimilar from the current residents, then over time the population in the locale would change. There would be an "invasion" of a new type of resident, and eventually that new type of resident would "succeed" the older type of resident.

Such cycles could be seen most clearly in the 1960s and 1970s in urban neighborhoods where racial succession took place, and white populations were replaced in relatively short order by African American households. Many expected that gentrified neighborhoods would follow the same cycle; but they have not. Even in some of the most reclaimed neighborhoods, higher-income, recent in-migrant owners mingle on the street with lower-income, longer-term, renters (Lee and Mergenhagen). The invasion-succession cycle can "stall" before completion. In these partially gentrified locations violent and property crime rates can be higher (Covington and Taylor).

Shaw and McKay's initial cross-sectional findings have been supported again and again (e.g., Baldwin, 1975). Studies routinely find the following.

Delinquency and offense and offender rates are higher closer to the city center than farther away, although there are exceptions, and although each of these outcomes maps differently onto spatial structure (Baldwin and Bottoms). Delinquency and offense and offender rates are higher in lower income, and/or less stable, and/or more predominantly African American communities (K. D. Harries, 1980), although differences have arisen regarding the relative contribution of each attribute, and the appropriate labels to apply to some of the dimensions of urban community structure examined (Sampson and Lauritsen). For example, some have argued that relative socioeconomic status in a locale—how poor the residents are, or how poor they are relative to those residents in adjoining neighborhoods—is the most important community correlate of high violent crime rates (Land et al.). Others argue that family disruption, and/or family structures that are less stable or provide less supervision of the locale are the most important (Sampson and Lauritsen). This debate is not about to end anytime soon.

In essence, the human ecological theory focuses on a community's position in the larger urban fabric, and how that position changes over time. It is its relative status, stability, and racial composition, and the changes in those features, that determine changes in offense, offender, and delinquency rates.

In a series of studies using Shaw and McKay's data on delinquency and census characteristics in Chicago, from the 1930s through the 1960s, more rapid community shifts connected with more rapid changes in the delinquency rate (e.g., Bursik, 1986). The ways in which neighborhoods changed varied across each decade, as did the relative contribution of different types of neighborhood changes to changes in delinquency. What was happening each decade was conditioned by the historical context. But despite these variations in each decade, community changes linked to delinquency changes in the expected ways. For example, increasing unemployment and increasing nonwhite racial composition were both tied to increasing delinquency rates.

Changes in neighborhood fabric are linked not only to changes in delinquency but also to changes in violence. A Baltimore study of changes in the 1970s found that neighborhoods shifting more dramatically on stability or status experienced more sizable shifts in violence as well (Taylor and Covington, 1988). Which particular feature of neighborhood fabric proved important depended on the type of violent crime examined.

Briefly put, one of the major extensions of social disorganization theory in the last two decades has been the application of the model to ecological changes over time. As the theory predicts, neighborhoods whose composition is changing more rapidly, relative to the other neighborhoods in the city, are more likely to experience increasing delinquency or crime problems. Even if the rapid change is in a "positive" direction, such as gentrification, increasing crime may accompany the shift (Covington and Taylor, 1989).

The features of neighborhood structure only predispose a neighborhood to have more or less social disorganization. Key studies, in Britain and in the United States, highlight the central importance of social disorganization versus collective efficacy processes (e.g., Sampson et al.). These processes mediate the impacts of structure on outcomes like offending and victimization, but structural impacts like differences in status and stability continue to exert some impacts on the outcomes separate from these processes (Versey and Messner).

Responses to crime like fear of crime are also ecologically patterned and social disorganization versus collective efficacy processes likewise appear to mediate the impacts of structure—status, race, stability—on the outcomes (Taylor, 1996). Similarly, rapid structural change affects these processes that in turn affect fear of crime (Taylor and Covington, 1993). Generally these studies show that although there are differences from study to study, neighborhood structure—especially status and stability—affects these outcomes in ways anticipated by the human ecological model, and that indicators of social disorganization versus social efficacy at least partially mediate the relationship.

In the last few years a related set of models concentrating on social and physical aspects of disorder in neighborhoods has emerged. The model terms these features incivilities. Physical incivilities include abandoned cars, weed-filled lots, vacant houses, and unkempt properties and yards. Social incivilities, although viewed by some as just misdemeanor crimes, include vandalism, rowdy groups of unsupervised teens, fighting neighbors, public drug use or drug sales, and the like. These models come in different forms, but the version drawing the most attention has suggested that incivilities can contribute independently, over time, to increasing neighborhood crime, neighborhood structural decline, and increased neighborhood fear (Skogan). Longitudinal analyses, however, show that incivilities do not change uniformly in locations (suggesting they are indicative of separate and somewhat unrelated problems) and the independent impacts of incivilities on neighborhood level outcomes are far weaker than the theory anticipates, although some predicted impacts do emerge (Taylor, 2001).

Policy impacts of work on the ecology of crime have been several. Since the early 1900s, city programs have targeted some of the areas where youths are at greatest risk of delinquency. More recently more refined geographic analyses of crime have concentrated enforcement efforts on crime "hot spots"—locations where police are called repeatedly to deal with crimes or disturbances. These targeted interventions can under some conditions have some deterrent or preventive impacts. Concern about incivilities has led to community policing initiatives targeted at these problems in the beliefs that reducing these problems will reduce crime. One recent longitudinal work suggests this enthusiasm may be misplaced (Taylor, 2001).

Interest in this work in the future will increase due in large part to increased availability of mapping software for locating crime and community data geographically, and allowing sophisticated spatial analyses (Weisburd and McEwen). Increasing availability of multilevel models also facilitate work in this area (Bryk and Raudenbush). In addition, after having fallen out of favor in the 1960s and 1970s, interest in the ecology of crime has increased in recent years and sociologists generally are discovering "neighborhood effects" in a range of topic areas. Nonetheless, one of the biggest factors holding work in this area back is the lack of routinely updated data that includes community characteristics, police calls for service and crime data, and social disorganization versus collective efficacy indicators for a number of neighborhoods in a number of cities. Hopefully future efforts of an inter-university consortium will work on such an effort.