The Importance of Both Opportunity and Social Disorganization Theory in a Future Research Agenda to Advance Criminological Theory and Crime Prevention at Places

Criminology has been primarily focused on people and why they commit crime. In an article in Crime and Justice that examined quantitative studies of criminological theory in Criminology (1968–2005), 60 percent of the articles identified focused on individuals (Weisburd and Piquero 2008). Just 15 (8 percent) of the studies examined communities or neighborhoods. Only one study examined a micro geographic unit such as that studied in our book, The Criminology of Place. While certainly the interest in small geographic units within neighborhoods and communities has grown over the last decade, the exploration of criminological theory at the level of small geographic units is new, and we would argue that it is an important area for criminological study. Braga and Clarke (Forthcoming) agree with us, and we think their thoughtful essay examining the implications of our work raises important issues regarding how to proceed with a research agenda on the criminology of place.

Behind recent interest in the criminology of place lies an empirical reality that fundamentally alters the ways in which we understand crime in urban areas. Starting with early studies of crime concentrations at addresses and extending to other micro geographic units (e.g., see Crow and Bull 1975; Pierce, Spaar, and Briggs 1986; Roncek 2000; Sherman, Bueger, and Gartin 1989; Weisburd et al. 2004; Weisburd and Green 1994), scholars began to identify what we call in our book a law of crime concentrations at places (Weisburd, Groff, and Yang 2012). Using the street segment as a micro geographic unit, there appears to be a consistent level of concentration across cities and across time (e.g., see Telep, Mitchell, and Weisburd Forthcoming; Weisburd and Amram 2014; Weisburd et al. 2004; Weisburd, Telep, and Lawton Forthcoming). About 5 percent of streets account for 50 percent of crime and about 1 percent of streets accounts for 25 percent of crimes. In The Criminology of Place, we not only illustrate that concentration; we also show that there is strong street-to-street variability in crime levels. Our key concern was to answer the question of what factors seem to explain such variability.

We wanted our study to be theoretically informed, so we began by identifying the key theoretical perspectives that have been identified earlier by scholars to understand crime at place. Opportunity theories including routine activities theory (Cohen and Felson 1979), situational prevention (Clarke 1980, 1983), and crime pattern theory (Brantingham and Brantingham 1984, 1981 [1991]) have been seen by most of those who study micro geographic units as the key factors explaining crime patterns. But we also thought that it is important to consider the relevance of social disorganization theories (Bursik and Grasmick 1993; Sampson 2012; Sampson, Raudenbush, and Earls 1997; Shaw and McKay 1942), which have been dominant in understanding crime in studies of neighborhoods and communities.

Braga and Clarke (Forthcoming) question why we took this latter approach and did not stick to the core theories that have motivated study in this area. Our answer is that we began with a theory of life on street segments that seemed consistent with core elements of social disorganization theories at a higher geographic level. Our work was informed by behavior setting theory (Barker 1968; Wicker 1987:614) and its applications to criminology in earlier work (Taylor 1997; Taylor and Harrell 1996). Street segments in this context can be seen as micro communities. People who frequent a street segment get to know one another and become familiar with each other’s routines. Residents develop certain roles they play in the life of the street segment (e.g., the busybody and the organizer). Norms about acceptable behavior develop and are generally shared. Blocks have “standing patterns of behavior” (Barker 1968, p. 18), for example, people whose routines are regular, like the mail carrier or the shop owner. Together, these elements support the cumulative familiarity that is the basis for the development of mutual trust, which supports the willingness to intervene and is necessary to the ability of a street segment’s users to achieve their shared goals. In this context, street segments have many traits of communities that are prominent in social disorganization theory. These small spatial units function also as social units with specific routines. This led us to consider the implications of social disorganization theory at a micro level.

Reinforcing the potential importance of such social structural variables is the fact that they were concentrated in micro geographic hot spots and often varied tremendously street by street, just as crime varied across streets. For example, 50 percent of housing assistance (a measure of social disadvantage) is consistently found on about 0.4 percent of the street segments in Seattle. Within 800 feet of these public assistance hot spots, 84 percent of street segments do not have any public housing assistance recipients. The question is whether housing assistance and other possible indicators of social disorganization were strongly related to crime hot spots.

Braga and Clarke (Forthcoming) note that many of the most important variables we identify are associated with opportunity theories, and this is consistent with prior theorizing in the criminology of place. We acknowledge that our work provides confirmation of many of the elements of opportunity theories. But we also found that measures of poverty and collective efficacy were also key factors in understanding variability of crime at street segments. Braga and Clarke raise questions about our use of voting behavior as an indicator of collective efficacy. They argue that voting behavior is not a good measure of collective efficacy, pointing to Sampson’s (2012) recent findings that voting patterns across neighborhoods were not strongly correlated with measures of collective efficacy. Importantly, our measure and Sampson’s measure are substantially different. Sampson examined the proportion of residents in the neighborhood who reported voting in the last mayoral election. We measure the proportion of active voters on a street, defined by voting patterns over two years. Voting once does not necessarily show strong commitment to involvement in public affairs, but voting consistently over time says more about an individual’s commitment (see Coleman 2002; Putnam 2001). More important, it reflects a general propensity toward civic engagement that is likely to be even stronger on their home street segment. In turn, we found the same variability across street segments here as in many other of our measures. Within 800 feet of the hot spots of active voters (the top 10 percent), only 25 percent of neighboring street segments also evidenced such high levels of active voting.

Of course, we would have preferred to have a more direct measure of collective efficacy and indeed many of the other theoretical constructs in our study. A key theme of the article by Braga and Clarke (Forthcoming) relates to the need for more and better data, both on the nature of opportunities for crime at a micro geographic level and indicators of social disorganization. We agree wholeheartedly. The weakness of secondary data studies more generally is that they rely on information that is available. We were able to collect a wealth of data reflecting opportunities for crime and social disorganization at places using archival records for Seattle. Although our data are the most exhaustive available for examining crime trends at the street segment level, we could not measure directly some key dimensions of either opportunity or social disorganization. Like any study that explores new territory, our work is necessarily a first step. Weisburd with Brian Lawton, Justin Ready, and Amelia Haviland have recently embarked on a large study supported by the National Institutes of Health that will allow detailed information to be collected on opportunity, social and demographic factors at the street segment level, including detailed measures of collective efficacy.

We think that Braga and Clarke (Forthcoming) provide important suggestions for how to proceed, not only in terms of the identification of measures but also in examining dynamic models of how variables intersect in time and space. We would certainly have liked to model not just the presence in space and time of specific measures of opportunity and social disorganization but also the extent to which these measures overlap within more specific bands of time. For example, we show that increased opportunities for crime increase the likelihood of a street falling in a chronic crime pattern. This is a confirmation of routine activities theory (Cohen and Felson 1979). But a next generation of studies armed with more robust data will hopefully be able to model in a more dynamic way the specific moments of time when victims and offenders intersect in the absence of a capable guardian.

Finally, we comment on the strong reservations that Braga and Clarke (Forthcoming) have about our policy recommendations in the area of social prevention. We agree that the evidence for situational responses to crime have strong empirical support (e.g., see Guerette and Bowers 2009). This is particularly the case for hot spots policing, where a large number of randomized controlled trials have shown that police can prevent crime (Braga 2005; Braga, Papachristos, and Hureau Forthcoming; Sherman and Weisburd 1995). Our work helps us to understand why such interventions are effective. Opportunities for crime are key to understanding the variability of crime at a micro geographic level, and increasing police guardianship in that context can deter crime (Durlauf and Nagin 2011; Nagin Forthcoming). Such deterrence will not simply shift crime to areas nearby (Weisburd et al. 2006) because, as our study illustrates, areas nearby will often not have the same types of crime opportunities.

There is, in contrast, little evidence of the salience of social prevention at a micro geographic level as Braga and Clarke (Forthcoming) point out. But we disagree strongly that this means that we should not begin to consider social prevention at crime hot spots. When hot spots policing experiments began, there was little basic research literature beyond cross-sectional studies of crime at addresses (e.g., Pierce et al. 1986; Sherman et al. 1989). And much of what we know now about situational prevention comes from applied research (e.g., see Clarke 1980, 1995; Guerette and Bowers 2009). Our work provides empirical support for the idea that street segments are behavior settings and that the characteristics of such behavioral settings vary across streets. Collective efficacy (as we measure it) and social disadvantage vary at a micro geographic level, and they are strongly related to crime at the street segment level. The question this raises is whether by altering these characteristics at a micro geographic level, we can do something about crime. We argue that directing social interventions at this level is more realistic, given the economic realities we face today, than trying to change whole neighborhoods or communities. In turn, just as hot spots policing has shown that the police must focus police patrol at the specific places where crime is concentrated (rather than wandering across large areas), we think that social interventions should be focused as well.

Braga and Clarke (Forthcoming) argue that social disorganization has been too amorphous a concept, seldom leading to direct proposals for action by police or other crime prevention agents. We think that the application of social prevention programs at the street segment level can help frame such approaches more sharply. Our suggested agenda would require clear and precise statements about what police or others might do to increase collective efficacy or reduce social disadvantage at specific places. Importantly, it will also require measurable mechanisms to be defined. For example, Weisburd and Gill are working with police in Brooklyn Park, Minnesota, to increase collective efficacy at hot spot streets (see Davis, Weisburd, and Gill 2013). Police will actively seek to meet with residents and encourage them to become involved with each other and the affairs of the streets more generally. They will try to identify key figures on the hot spot streets and work with them to get people on the block involved in crime prevention. Many community-policing programs have already taken similar approaches, without recognizing concepts of collective efficacy and informal social controls (Gill et al. 2014). Of course, as Braga and Clarke (Forthcoming) suggest, it will be difficult to distinguish these crime prevention gains from those that are due to the development of situational prevention approaches and deterrence through increased guardianship, by both the police and the public. We think that what is new here is the direct recognition that collective efficacy at the micro geographic level is relevant and important.

In some ways, our suggestions are more radical than simply taking into account both the social and physical environment of places, and perhaps this is what raises particular concern on the part of the Braga and Clarke (Forthcoming). We are suggesting that street segment characteristics offer a stronger foundation for the targeting of social programs meant to address structural problems that lead to crime, such as employment and income inequality. We think that we should test the efficacy of economic and employment programs at the micro geographic level. And we argue that economies of scale make it feasible to develop such social interventions. Successful crime prevention can be seen as operating at two levels. One level responds to the immediate situational components of crime and often relies on deterrence and directly reducing crime opportunities as a method of reducing the likelihood of crime. At another, there are long-term social factors often relating to disadvantage and low collective efficacy. Our work suggests that it is time to experiment in this latter area at a micro geographic level. There will often be overlap, and we suspect that the most effective long-term crime prevention will focus on both approaches. Informal social controls that are encouraged by increased collective efficacy will also work to reduce opportunities for crime. Hot spots policing and situational prevention may, by reducing fear of crime and increasing quality of life on streets, also increase collective efficacy and reduce social disadvantage. It is time to examine these possibilities and mechanisms more carefully to develop a fuller repertoire of crime prevention at a micro geographic level. Accordingly, we would add to their proposed research agenda an emphasis on how social prevention at hot spots can be more carefully defined and systematically evaluated.