Measuring community social organization: Sense of community as a mediator in social disorganization theory

Journal of Criminal Justice 31 (2003) 321 – 339

Measuring community social organization: Sense of community as a mediator in social disorganization theory Dan Cantillona,*, William S. Davidsonb, John H. Schweitzerc a

Health Research and Policy Centers (MC 275), School of Public Health, University of Illinois at Chicago, 850 West Jackson Boulevard, Suite 400, Chicago, IL 60607, USA b Department of Psychology, Michigan State University, 58 Baker Hall, East Lansing, MI 48824, USA c Urban Affairs Department, Michigan State University, W-30 Owen Graduate Hall, East Lansing, MI 48824, USA

Abstract The current study utilized an updated systemic model of social disorganization to investigate neighborhood effects on both positive and negative youth outcomes. Although empirical support for updated social disorganization models has increased in recent years, the field continues to rely too heavily on behavioral indicators of community social organization. Unfortunately, these measures do not assess the truly important social processes and dynamics that result in cohesive and supportive neighborhoods. It was proposed that sense of community (SOC) was a more valid, comprehensive, and applicable measure for the mediating variables in social disorganization theory. Results supported the hypothesis that SOC mediates the effect of neighborhood disadvantage on youth outcomes and implications for the field are discussed. D 2003 Elsevier Science Ltd. All rights reserved.

Introduction Social disorganization theory re-emerged and is once again one of the dominant social theories utilized to explain the influence of neighborhood characteristics on variations in crime and delinquency rates (Bursik, 1988; Bursik & Grasmick, 1993; Elliott et al., 1996; Sampson, 1997; Sampson & Groves, 1989; Simcha-Fagan & Schwartz, 1986). The resurgence of social disorganization theory was accompanied by the corresponding need to accurately and validly measure the level of community social organization. The goal of the current study was to assess the viability of using sense of community (SOC) as a measure of community social organization and to * Corresponding author. E-mail address: [email protected] (D. Cantillon).

evaluate if it mediated the impact of neighborhood disadvantage on the youth outcomes of delinquency, conventional activity, and grade point average. First, this article reviews social disorganization theory and the advances made in the field of neighborhood effects, with an emphasis on research that focused on the outcomes of delinquency and violence. Specifically, this review explicated the multiple ways the proposed mediating variables of social disorganization theory were measured to date. Second, a brief review of the literature on SOC illustrated the strengths of this construct, how it was measured in the past, and why it deserved inclusion in social disorganization theory. Third, this study demonstrated that SOC mediated the influence of neighborhood disadvantage on youth outcomes and implications for future research within the field of neighborhood effects are discussed.

0047-2352/03/$ – see front matter D 2003 Elsevier Science Ltd. All rights reserved. doi:10.1016/S0047-2352(03)00026-6

322

D. Cantillon et al. / Journal of Criminal Justice 31 (2003) 321–339

Social disorganization theory One of the earliest and most influential social ecological theories of delinquency, social disorganization, was based on an ecological theory of urban dynamics that originated at the University of Chicago (Park, Burgess, & McKenzie, 1967). The old Chicago school examined both the physical expansion of the city of Chicago and the concomitant changes in social life for its residents. Shaw and McKay (1942) extended this ecological approach in social disorganization theory to account for the greater proportion of delinquency that occurred in certain neighborhoods in Chicago. In general, social disorganization theory was defined as the inability of local communities to realize the common values of their residents or solve commonly experienced problems (Kornhauser, 1978). Shaw and McKay’s main finding was that the association between neighborhood residence and delinquency decreased as the distance from the central city increased, and that this association remained despite the rapid changes in the ethnic and racial characteristics of neighborhood residents. Thus, rather than blaming the higher rates of crime and delinquency on the individuals who resided in these urban neighborhoods (i.e., individual-level theories), social disorganization theory looked at how macrolevel structural processes such as poverty and residential instability detrimentally impacted community social organization (social networks, norms, informal social control) and resulted in increased rates of crime and delinquency. In its severe form, social disorganization leaves residents isolated from one another and from the social institutions that are supposed to provide basic services and a reasonable quality of life. The lack of residential stability and the presence of a heterogeneous population make it extremely difficult to establish a strong network of relations or ‘‘weak ties’’ within the community that serve to establish norms and a supportive context for youth development. In terms of formal community structures, the lack of neighborhood stability interacts with low socioeconomic composition and results in a paucity of highquality neighborhood institutions (e.g., schools, social service agencies), which serve to bind residents together and provide services to residents, particularly youth. In sum, social disorganization is a systemic theoretical model, which incorporates formal associations as well as the informal networks within a community that arise through friendship and kinship ties (Kasarda & Janowitz, 1974). Social disorganization theory posited that it was the level of community social organization which mediated the relationship between a neighborhood’s

compositional or structural characteristics and delinquency rates. Thus, social disorganization and social organization are on opposite ends of a continuum and describe a neighborhood’s capacity to exert control over inappropriate or illegal behavior within its domain. This distinction is important because confusion has arisen due to the misuse of the theoretical framework in the past. For instance, social disorganization was used as a descriptor of disadvantaged environments (e.g., structural disadvantage), the mediating social process (e.g., social organization/disorganization), and even its ultimate outcomes (e.g., delinquency and crime; Bursik, 1988). Rather than solely describing the structural characteristics of communities or community outcomes, social disorganization theory should be viewed as an explanatory framework that links how disadvantaged neighborhood contexts impact the level and extent of community social organization, which impacts youth outcomes such as delinquency.

Updated systemic social disorganization models Despite its widespread theoretical acceptance, one of the main and certainly valid criticisms of the social disorganization model was the lack of empirical support for the proposed intervening relationship between neighborhood disadvantage and delinquency. This was due, in large part, to the overreliance on census data. While the important independent variables of social disorganization theory (e.g., low SES, population turnover, cultural heterogeneity, family composition) were easily quantified through census data, the measurement of the mediating variable (e.g., community social organization) required costly and time-consuming survey and interview collection (Bursik, 1988; Heitgerd & Bursik, 1987; Sampson, 1993a). Thus, a large number of studies hypothesized this mediating process of community social organization without actually measuring it (Brooks-Gunn, Duncan, Klebanov, & Sealand, 1993; Simons, Johnson, Beaman, Conger, & Whitbeck, 1996). Recent studies addressed this limitation though and resurrected social disorganization theory to be one of the major conceptual frameworks through which to investigate the impact of neighborhood disadvantage on youth outcomes, particularly in the field of delinquency research (Bursik & Grasmick, 1993; Levanthal & Brooks-Gunn, 2000). Simcha-Fagan and Schwartz’s (1986) study marked the return of social disorganization theory by empirically demonstrating that the level of community social organization mediated much of the

D. Cantillon et al. / Journal of Criminal Justice 31 (2003) 321–339

effects of disadvantaged structural characteristics on self-reported and officially recorded delinquency. Since the publication of this seminal article, a number of other studies utilizing the social disorganization framework confirmed that the level of community social organization mediated a significant amount of the effects of negative structural characteristics on delinquency, victimization, and general levels of violence (Elliott et al., 1996; Krivo & Peterson, 1996; Sampson & Groves, 1989; Sampson, Morenoff, & Earls, 1999). Unfortunately, as the hypothesized theoretical link between neighborhood disadvantage and delinquency continued to receive empirical support, the field’s sophistication in conceptualizing and operationalizing the mediating variable of community

323

social organization was incomplete. As Table 1 indicates, there were two dominant methods of measuring community social organization that had been empirically supported. One method was to measure the level of friendship networks or contacts in the neighborhood, which emerged from the theoretical tradition of emphasizing the importance of social networks and weak ties (Bellair, 1997; Elliott et al., 1996; Granovetter, 1973; Sampson & Groves, 1989; Warner & Rountree, 1997; Warren, 1978). The second approach was to examine the informal social control processes that operated at the neighborhood level (Elliott et al., 1996; Sampson, 1997; Sampson & Groves, 1989). In this case, informal social control was usually measured by the likelihood that neighborhood residents would intervene

Table 1 Significant findings from recent social disorganization studies Study

Mediating variables

Outcome variables

Simcha-Fagan and Schwartz (1986)

. Organizational participation . Community disorder—criminal subculture

. Self-reported delinquency . Severe self-reported delinquency . Officially recorded delinquency

Sampson and Groves (1989)

. Organizational participation . Local friendship networks . Control of street corner youtha

. Criminal offending rates (personal violence and property/vandalism) . Criminal victimization rates (mugging/robbery, stranger violence, total crime, etc.)

Elliott et al. (1996)

. Social integration . Informal networks . Informal controlb

. Prosocial competence . Conventional friends . Problem behavior

Bellair (1997)

. Social interaction

. Burglary . Motor vehicle theft . Robbery

Sampson et al. (1997)

. Collective efficacyc

. Violence . Household victimization . Homicide rates

Warner and Rountree (1997)

. Local social tiesd

. Assault . Burglary

Bellair (2000)

. Informal surveillancee . Informal control

. Burglary . Assault/robbery

The reviewed studies all utilized a social disorganization framework with similar independent variables such as socioeconomic status, population turnover, cultural heterogeneity, and family disruption variables. Only the mediating variables that proved significant with at least one of the outcome variables are listed. a Control of street corner youth was the most consistent significant mediating variable across the criminal offending and criminal victimization outcomes. b Informal control was the only mediating variable to reach significance with all three outcome variables. c Collective efficacy combined five items on social cohesiveness and trust and five items on informal social control. d Local social ties significantly reduced assault but had little mediating effect between community structure and crime rate. Local social ties was also insignificant and in the opposite direction as expected for burglary. e Results suggested that some forms of crime (burglary) increased surveillance and informal control, while others reduced it (robbery/stranger assault), possibly out of fear of victimization.

324

D. Cantillon et al. / Journal of Criminal Justice 31 (2003) 321–339

if children and teenagers were engaging in ‘‘delinquent’’ behaviors. Table 1 displays that informal social control received more empirical support, especially as of late, and is being utilized to a greater degree than friendship networks to measure the level of community social organization. Logically, informal social control taps into the ability of a community to realize its common values and regulate behavior that would be harmful to the collective and, in fact, there was extensive discussion over the years on the important role of informal social control in controlling crime and delinquency (Greenberg & Rohe, 1986; Sampson et al., 1999). At this point, only a minor distinction was drawn between these two methods of measuring community social organization in recent literature; however, it seemed readily apparent that the development of social networks, weak ties, and community norms were a prerequisite for informal social control and intervention behaviors to occur in the community. Specifically, the degree to which families, neighbors, community members, and local institutions interact with and socialize youth is part of the process of a community realizing its common values and developing shared norms. Once this consensus is reached, it is communicated informally from neighbor to neighbor and more formally via a community’s institutions. It is only when such shared norms exist that residents are emboldened to intervene when they are being violated. In other words, informal social control, as measured by intervention behaviors, should be viewed as an outcome of community social organization rather than as a true measure of community social organization itself. The major problem is that such measures do not capture the development and level of social networks, trust, and reciprocity that is necessary for the development of cohesive and supportive communities that enhance youth development. Even though some of the newer studies incorporated items that measured informal social processes of community cohesion in addition to intervention items (Sampson, Raudenbush, & Earls, 1997), there continued to be an overreliance on behavioral indicators. Moreover, the few studies that assessed these complex social processes found inconsistent results (Bellair, 2000; Hackler, Ho, & Urquhart-Ross, 1974; Macoby, Johnson, & Church, 1958; Warner & Rountree, 1997). A critical next step for the field is to develop or utilize a pre-existing measure that assesses the social processes which are required for informal social control and intervention behaviors with youth in the community. Recent studies noted this need and asked researchers to incorporate measures that moved beyond behavioral indices and measured the various

social processes that were inherently embedded in the complex and reciprocal relationship between community social organization, informal social control, and crime and delinquency (Bellair, 2000; Veysey & Messner, 1999). SOC measures such social processes. This construct measures the processes that precede behavioral intervention and are usually enforced through more subtle social interactions among neighborhood residents, such as the withdrawal of sentiment, respect and esteem, and social and instrumental support (Black, 1989; Greenberg & Rohe, 1986; Hunter, 1985).

Sense of community SOC has been defined as ‘‘a feeling that members have of belonging, a feeling that members matter to one another and to the group, and a shared faith that members’ needs will be met by their commitment to be together’’ (italics added; as cited in McMillan & Chavis, 1986, p. 9). As the definition indicates, SOC is an emotion-laden construct that stresses feelings of togetherness and a communal spirit, variables not typical in psychological or criminological research. As Sarason (1974) sardonically noted over two decades ago, ‘‘It [sense of community] is a phrase associated with a kind of maudlin togetherness, a tear-soaked emotional drippiness that misguided dogooders seek to experience’’ (as cited in Chavis, Hogge, McMillan, & Wandersman, 1986, p. 24). Thus, while readily acknowledging its ‘‘nonscientific’’ nature, it is precisely the shared emotional and communal quality of this variable that captures the complex and subtle social processes which lead to cohesive and supportive communities. In conjunction with the above definition, McMillan and Chavis (1986) conceptualized four distinct aspects of SOC: membership, influence, sharing of values with an integration and fulfillment of needs, and a shared emotional connection. Thus, for SOC to exist in a residential neighborhood setting, residents must identify with the community, feel that they matter to the community and that the community matters to them, feel that the community shares their values and meets their needs, and experience affective attachments with other community members.

Measuring SOC While these four proposed theoretical dimensions of SOC have long been discussed, empirical evidence has not fully supported them and this resulted in varying conceptualizations and measures of SOC over the years (Chipuer & Pretty, 1999) (please see

D. Cantillon et al. / Journal of Criminal Justice 31 (2003) 321–339

325

Table 2 Sense of community and related measures for geographic communities Study

Quick description

Doolittle and Macdonald (1978) . PCA found support for six separate factors.

Basic results . Six components: supportive climate, family life cycle, localism, safety, informal interaction, and neighborly interaction . Percent of variance accounted for by each component: 15, 12, 8, 7, 6, and 6 percent

Krupat and Guild (1980)

. Surveyed a number of different sized communities and created a community social climate scale with thirty items and six components.

. Six components: warmth and closeness (a=.83), activity/entertainment (a=.78), alienation/isolation (a=.68), good life (a=.66), privacy (a=.54), and uncaring (a=.53)

Riger and Lavrakas (1981)

. Found two important components in patterns of attachment and interaction of neighborhood residents. . Created a typology of neighborhood attachment for residents: young mobiles, young participants, isolates, and established participants.

. Two components: physical rootedness (a=.59) and socially bonded (a=.56) . Rooted accounted for 38 percent of variance and bonded for 17 percent of variance

Glynn (1981)

. Assessed ideal and actual levels of SOC in three different communities through a 178-item scale. . Decreased instrument to sixty items through analyses.

. Unidimensional – psychological sense of community . Actual SOC (a=.97) . Ideal SOC (a=.92)

Chavis et al. (1986)

. Four components: membership, . Created the Sense of Community influence, integration and fulfillment Index (SCI), which had four components. of needs, and shared emotional . Shortened version of scale connection (SCI—short form) was first widely published in Perkins et al. (1990). It contained twelve . Overall alpha has been found to be around .71 items and was one of the most often SOC scales. It had, however, reliability problems, especially the subscales.

Davidson and Cotter (1986)

. Created Sense of Community Scale (SCS) which originally had seventeen items. . Referred to level of city versus neighborhood.

. Unidimensional . a=.85 in one community and .81 in the other

Buckner (1988)

. Created Neighborhood Cohesion Instrument (NCI). Originally thought to be three separate scales (attraction to neighborhood, neighboring, and psychological sense of community), but analyses indicated it could be one overall construct.

. Unidimensional . Seventeen-item scale entitled NCI . a=.95

Pretty (1990)

. Used short form of SCI in a university residence sample comparing SOC with the University Residence Environment Scale (URES; Moos & Gerst, 1974).

. Unidimensional . PCA found that it was inappropriate to look at the short form as four separate components (continued on next page)

326

D. Cantillon et al. / Journal of Criminal Justice 31 (2003) 321–339

Table 2 (continued) Study

Quick description

Basic results

Davidson and Cotter (1993)

. Used a shortened version of their SCS. This version was reduced to five items.

. Unidimensional . a=.84

Pretty, Andrews, and Collet (1994)

. Used shortened version of SCI to calculate school sense of community and neighborhood sense of community.

. Unidimensional

Hill (1996)

. Reviewed various measures of SOC and discussed implications for the field.

. Found less than thirty published measures of SOC and factor-analytic studies yielded both unidimensional and multidimensional interpretations

Puddifoot (1996)

. Reviewed measures of SOC, community satisfaction and identity, etc. . Argued for more multidisciplinary research and focused on qualitative measures.

. Fourteen components—presented an argument for fourteen dimensions of community identity

Skjaeveland et al. (1996)

. Created the Multidimensional Measure of Neighboring (MMN) scale. This scale contained fourteen items and one of few to measure negative neighbor relations.

. Four components: supportive acts of neighboring, neighbor annoyance, neighborhood attachment, and weak social ties . Percent of variance accounted for by each component: 32, 16, 8, and 8 percent . Alphas ranged from .70 to .86 for the four subscales

Barnes (1997)

. Adapted questionnaire from Simcha-Fagan and Schwartz (1986) to use with families with young children. . Named Neighborhood Characteristics Questionnaire (NCQ). . Demonstrated similarity of community social organization in SD theory and measures of SOC.

. Four components: perception of street crime and life quality (a=.85), social relationships and networks among neighbors (a=.82), attachment to neighborhood (a=.81), and neighborhood disorder (a=.77)

Brodsky, O’Campo, and Aronson (1996)

. Used a revised form of the SCI short form.

. Unidimensional . PCA found support for two dimensions, although one was too small, so only kept one dimension for the study . a=.84

Kingston et al. (1999)

. The SOC scale utilized in this study assessed many of the same elements as previous scales although some components were developed for this study.

. Four components: neighborhoodrelated attitude scale (a=.83), neighborhood influence scale (a=.89), neighborhood-related behavior (a=.80), and participation in community organizations (no alpha, one item)

Chavis and Pretty (1999)

. Reviewed theoretical and measurement articles on SOC and served as an introduction to a special issue.

. One major theme of this article was the continued search for measures

Chipuer and Pretty (1999)

. Reviewed the short form of SCI and its proposed four dimensions.

. Found that components of SCI were unreliable (alphas ranged from .16 to a high of .72) . Suggested using long form of SCI as foundation to create a new measure

D. Cantillon et al. / Journal of Criminal Justice 31 (2003) 321–339

327

Table 2 (continued) Study

Quick description

Basic results

Chipuer et al. (1999)

. Created an SOC scale for youth called the Neighborhood Youth Inventory. Twenty-two-item measure which had four components. . Used Buckner’s NCI and short form of SCI to create inventory with input from youth.

. Four components: support, safety, activity, and friendships. Alphas ranged from .64 to .94

Crew et al. (1999)

. Four items measuring SOC at block level.

. Unidimensional . a=.81

Schweitzer et al. (1999)

. SOC measure was comprised of sixteen items related to connection, belonging, and support.

. Alpha not reported, but assessed as a unidimensional construct

Prezza, Amici, Roberti, and Tedeschi (2001)

. Translated Davidson and Cotter’s SCS into Italian. Eighteen items. . Found support for five components, but also found that it could be looked at as one dimension.

. Unidimensional . While five factors had eigenvalues over 1, all items loaded >.36 on the first factor, so interpreted unidimensionally. Overall a=.82

Zani, Cicognani, and Albanesi (2001)

. Translated Davidson and Cotter’s SCS into Italian and measured adolescent’s SOC. . Items loaded differently from adults. So, concept of SOC played out differently for youth.

. Four components: opportunity for participation and fulfillment of needs, pleasantness of living area, social climate, and membership . Percent of variance explained by each component: 15, 14, 12, and 12 percent . a=.85

Martinez, Black, and Starr (2002)

. Created Perceived Neighborhood Scale (PNS), a scale with items from previous published items on SOC, neighborhood satisfaction, etc. Intended for parents of young children. Total thirty-four items and a CFA found four components.

. Four components: social embeddedness, sense of community, satisfaction with neighborhood, fear of crime

Obst, Smith, and Zinkiewicz (2002)

. Looked at numerous scales over the years (Bardo & Bardo, 1983; Buckner, 1988; Glynn, 1981; Lalli, 1992; Nasar & Julian, 1995; Skjaeveland et al., 1996). . Performed a PCA on all ninety-five items from these scales and found support for five components of SOC.

. Five components: ties and friendship, influence, support, belonging, and conscious identification . Percent of variance accounted for by each component: 24, 13, 10, 7, and 4 percent

An attempt was made to review the vast majority of published studies on sense of community and related variables. Since the concept spanned many disciplines and was labeled differently across these content areas, some studies might have been missed. Also, studies which investigated nonterritorial communities (e.g., Bishop, Chertok, & Jason, 1997) were excluded from presentation.

Table 2 for a brief description of prior measures). There has even been considerable disagreement over whether SOC is a unidimensional or multidimensional construct (Hill, 1996). While evidence seems to be in favor of a multidimensional construct, the exact factor structure of SOC has not been confirmed. Thus, the dimensions of this important construct remain elusive, yet rather than deterring research, it

should simply indicate the diversity of the community experience. Simply put, like people, all communities are distinctive. Thus, it is unlikely that any one measure of SOC is going to capture all the important dynamics across various communities (e.g., urban/rural, poor/ wealthy, stable/transitory, territorial/nonterritorial). Given the debate over what exactly constitutes the

328

D. Cantillon et al. / Journal of Criminal Justice 31 (2003) 321–339

important elements of SOC, the conceptualization used for this study attempted to assess its most basic and vital components: a sense of physical safety, emotional connections, and attachment, and an empowering or action-oriented component. It is argued that healthy communities first require an adequate level of physical security and safety, and unfortunately, many researchers neglected to include this dimension in their conceptualization and measurement of SOC. Since the field of neighborhood effects usually attempts to identify how poor urban environments affect adolescent development, it seems ironic that more emphasis was not placed on basic physical safety issues, especially since many of these researchers noted how dangerous some of these neighborhoods were to live in. One recent study (McGuire, 1997), which included a safety measure, summed up its importance by stating, ‘‘It appears in this community residents’ judgments about the neighborhood as a worthwhile place to live may be most heavily influenced by immediate danger and crime, because the ratings of quality of life loaded on to the street crime scale’’ (p. 562). Healthy communities are also often described as areas where people know each other and experience a feeling of togetherness. As articulated by McMillan and Chavis (1986), the shared emotional connection among neighborhood residents ‘‘. . . seems to be the definitive element for true community’’ (p. 14). Thus, while safety is necessary for the mere existence and development of SOC, its defining element is the camaraderie and connection neighborhood residents feel for each other. As extensively noted by researchers, authors, and lay people, communities can be extremely safe, and yet at the same time, also extremely alienating with little to no feelings of community togetherness. In safe communities where there is a strong emotional attachment among residents, this is usually witnessed by neighborhood celebrations such as block parties, summer festivals, and other occasions, where, whether consciously noted or not, community itself is celebrated. It is in such communities that neighbors can come together to also address public problems and it is argued that this action orientation or ‘‘neighborhood empowerment’’ is vital for a healthy community. In fact, it is this element of SOC that most closely parallels the definition of social disorganization as laid out by Kornhauser (1978) (i.e., the inability of local communities to realize the common values of their residents or solve commonly experienced problems). Thus, as conceptualized and measured in this study, SOC corresponded to the original formulation laid out by Chavis et al. (1986), except that more emphasis was placed on physical safety and their construct of

influence and fulfillment of needs was collapsed into one component and labeled action.

Research findings on SOC Since SOC’s conceptualization, research demonstrated its importance to both neighborhood-level and individual-level outcomes. For instance, SOC was found to relate to the amount of emotional and instrumental support one provided to neighbors (Unger & Wandersman, 1982, 1983), and communities with greater degrees of SOC were able to communicate important information about neighborhood resources for children (McGuire, 1997). SOC was also found to have a positive effect on psychological health in adults (Davidson & Cotter, 1991). In accordance with social disorganization theory, SOC was found to influence the degree to which residents worked together on common public problems and even participated in the political process (Chavis & Wandersman, 1990; Davidson & Cotter, 1989, 1993). In youth, SOC was found to significantly reduce adolescent loneliness and to be more important in this respect than levels of social support (Pretty, Conroy, Dugay, Fowler, & Williams, 1996). Research also documented the strong role perceived safety played in the development of adolescent SOC (Chipuer et al., 1999). Most significant in terms of using SOC as a mediating variable in social disorganization theory, prior research documented SOC as a quantifiable neighborhood-level construct which could be targeted in designing intervention and rehabilitation programs for disadvantaged urban neighborhoods (Buckner, 1988; Chavis et al., 1986; Glynn, 1981; Kingston, Mitchell, Florin, & Stevenson, 1999). In sum, the re-emergence of social disorganization theory and its reformulated systemic models remains in an embryonic state. While conceptual and methodological improvements advanced theory and empirical support, the field is still struggling for a complete and adequate measure of community social organization. The main argument of the current study was that SOC was a more valid, comprehensive, and applicable construct than prior mediating variables used in social disorganization studies. While strong theoretical arguments were made that SOC improved upon the current measures of community social organization, the fact remained that it was never empirically evaluated within the social disorganization framework. Thus, the main research goal of the current study was to test SOC as a mediator in social disorganization theory. Specifically, it was hypothesized that SOC would mediate the effect of neighborhood disadvantage on both positive and negative youth outcomes.

D. Cantillon et al. / Journal of Criminal Justice 31 (2003) 321–339

Methodology Procedures Setting This research study was conducted in a mediumsized midwestern city with a population of approximately 127,000. The city was the state capital and had a large presence of automobile and related manufacturing companies. All interviews were conducted either in the homes or on the front porches of the respondents from June 1999 to early January 2000. Participants A total of 103 tenth-grade males participated in the study along with one of their parents and one of their neighbors. Therefore, for each student there were a total of three interviews (student, parent, and neighbor) for a combined total of 309. Of the 103 youth, 41 percent were White, 40 percent African American, 11 percent Hispanic, 6 percent Asian, and 3 percent mixed race. As can be seen in Table 3, student respondents closely matched the racial/ethnic percentages of the city’s high school students. Primary caretakers and neighbors had to be over the age of eighteen. Of the 206 participating adults, 67 percent were female and 67 percent owned their own homes. The average length of time respondents lived on the block was approximately ten years. An interpreter was utilized to increase the participation of Asian residents since a large percentage were Hmong refugees and many of these youths’ parents spoke little or no English. Recruitment procedures A list of 300 tenth-grade male youth was randomly computer generated from the city’s school district office and provided to the research project.

Table 3 Race/ethnic background of city’s high school students and research participants Race/ethnic category

High school students (%)

Research participants (%)

White African-American Hispanic Asian Mixed race Native American

45 36 12 7 N.A.a 1

41 40 11 6 3 0

Race/ethnic background of students was obtained from the city’s school district office of research and evaluation services. Percentages were rounded so sum could be over 100 percent. a N.A. = not available. Mixed race/ethnicity was not an option on school demographic forms.

329

Explanatory letters were sent out to every youth’s guardian with a stamped return address postcard for parents who wanted their teenager to be removed from a list of potential study participants. This process resulted in a 22 percent reduction in the sample to a list of 235 potential study participants. Of the 235, 21 percent were dropped because of school database errors (e.g., wrong address, wrong age) or if the family moved, and an additional 17 percent refused to participate when approached for inclusion in the study. There was a three-step consent procedure employed to assure that no parent, youth, or neighbor felt coerced to participate. Parents needed to sign a consent form allowing project staff to interview them, their tenth-grade youth, and another resident on the block. Even if parents consented, youth were also asked to sign an assent form prior to completing the youth survey. While referred to as neighbors, the third individual interviewed was not required to live next door to the youth, but had to fulfill three requirements. They had to be at least eighteen years old, agree to participate, and had to live on the same face block or block as the youth. Face blocks or blocks were operationally defined as two sides of a street intersected by cross streets, a dead end, or a similar demarcation. Face-block measures The utilization of the face block or urban block as a conceptualization of ‘‘neighborhood’’ had a long history and empirical research documented the importance of this contextual setting in understanding neighboring behaviors, SOC, crime, and fear of crime (Appleyard, 1981; Perkins, Florin, Rich, Wandersman, & Chavis, 1990; Perkins & Taylor, 1996; Schweitzer, Kim, & Mackin, 1999; Taylor, 1997; Unger & Wandersman, 1985). Unlike the majority of the research to date, this study also directly measured the proposed independent variables of income, residential turnover, cultural heterogeneity, and family composition at the block or face-block level. Previous studies utilized some combination of census indicators or other administrative unit data to derive these measures. Income or SES was typically measured by summing and standardizing a number of indicators within a census tract such as: percentage of families below the poverty line, percentage of residents employed in professional or managerial positions, percentage of residents who completed college, and average income or housing value. Residential turnover was most often calculated as the proportion of families that moved in the last five years in a census tract. Cultural heterogeneity usually was measured by the presence and number of racial/ ethnic groups greater than 10 percent in a census

330

D. Cantillon et al. / Journal of Criminal Justice 31 (2003) 321–339

tract. Finally, family composition was measured in the past by the proportion of single-parent families within a census tract. The direct assessment of neighborhood structural characteristics used in the current study allowed for much more accurate estimates since data were collected at the local level (block), and more importantly, since the census data would have been nearly a decade old. Another major difference in the current study was the way in which the independent variables were operationalized. The independent variables were conceptualized and measured on a level of advantage versus the traditional method of focusing on levels of disadvantage. While low levels of neighborhood advantage correspond to neighborhood disadvantage, this modification is more than merely a matter of semantics. It is important to be sensitive to the communities under study because what is being investigated at the community level ultimately affected individual lives in the areas of safety, quality of life, and even the possible life outcomes of youth and adults in these communities. By conceptualizing and measuring these neighborhoods in a positive manner, the dialogue can change from one focused on deficits to one focused on investigating and explaining the various strengths of these neighborhoods, and thus, the assets and strengths of the people who live in them. Block income Income was individually calculated by dividing monthly income by the number of residents in the house. To obtain a block income measure, the parent’s and neighbor’s income were averaged. Block stability Individual scale scores (e.g., parent, neighbor) were calculated by summing four Likert-type items which asked respondents to assess the residential stability of their block (e.g., people move in and out of this block a lot). Block scores were obtained by averaging the parent’s and neighbor’s total scale scores. Respondents were asked to rate from 5 (strongly agree) to 1 (strongly disagree) their endorsement of four items. Reliability was calculated at the block level by including both parent’s and neighbor’s responses in the analysis. Alpha was .84. Block homogeneity Individual scale scores were calculated by summing two Likert-type items on a five-point scale from 5 (strongly agree) to 1 (strongly disagree). Parent’s and neighbor’s scores were averaged for the faceblock score. Items included: (1) The residents on this block are from my racial/ethnic group, and (2) This block is racially/ethnically diverse.

Block composition Block composition was calculated as the percentage of two-parent households based on a family structure item. The item asked the parent and neighbor if they had a spouse or live-in partner. Thus, the variable of interest in this study was if there were two adults living in the house, regardless of whether these individuals were married or if they were the birthparents. Each block could fall into one of three categories: 100 percent two-parent households, 50 percent two-parent households, or 0 percent twoparent households. Sense of community This scale was created to measure the SOC that existed on residential face blocks within an urban community. The scale was created by the SOC in Lansing Neighborhoods project team and incorporated items from previously published scales of SOC, community attachment, neighborhood attachment, and neighborhood cohesion (Crew, Kim, & Schweitzer, 1999). Parents and neighbors of the targeted youth were asked to respond to this measure and their scores were again averaged to construct an overall SOC measure for the face block. Respondents were asked to rate from 5 (strongly agree) to 1 (strongly disagree) their endorsement of each of the SOC items. Examples included: (1) People on this block socialize with each other, (2) If faced with a problem on the block, residents would be unable to create a solution—reverse coded, and (3) People on this block feel they belong here. Alpha was .93 for the block. Principal components and reliability analyses indicated that there were three dimensions to the SOC construct: safety, emotion, and action (please see Appendix A). Emotion The emotion component of the SOC construct contained eight items related to connection, belonging, and support among residents on the face block. This construct measured the emotional connection between residents on the block and was representative of the affective component of SOC. Reliability was .79 for the block. Action The action component of the SOC construct contained seven items pertaining to residents’ feelings of influence, empowerment, and actual participation in block-related activities. Reliability was .84 for the face block. Safety The safety component contained two items regarding safety issues on the block. One asked about residents’ perceptions of safety during the day and

D. Cantillon et al. / Journal of Criminal Justice 31 (2003) 321–339

one asked about resident perceptions of safety at night. Youth outcome measures There were three measures in the youth outcome survey: (1) self-reported delinquency, (2) conventional activity, and (3) grade point average. The vast majority of the items from the youth survey were adopted from Elliot et al.’s (1996) youth outcome measure, although there were some modifications to the delinquency section. Self-reported delinquency This was a basic self-report delinquency scale with seventeen items. The questions surveyed a range of delinquent behaviors from minor offenses (e.g., stole or tried to steal something worth US$5 or less) to major infractions of the juvenile code (e.g., attacked someone with the idea of seriously hurting him or her). Similar to SOC, a principal components analysis was conducted and found empirical support for four separate dimensions to this variable (please see Appendix B). Due to severe skewness and kurtosis problems, all of the delinquency scales were log transformed [LN(Scale + 1)]. The following section provides a brief description of the items that comprise each construct along with their reliabilities.

331

Steal and deal delinquency This component contained five items. Four of these items related to stealing and one asked if the participant had sold drugs. Alpha was .69. School delinquency This component contained three items. One item asked about skipping classes while in school and another item asked about skipping a full day of school. The final item in this component asked about carrying a weapon outside of school. The first two items were status offenses and indica-tive of ‘‘minor’’ delinquency while the third would be a crime if committed by an adult. Alpha was .87. Steal and fight delinquency This component was comprised of three items. Two of these items related to stealing and the final item asked if the respondent had fought someone physically in the past year. Alpha was .48. Severe delinquency Two items constituted this component of selfreported delinquency. These two items were on the severe end of the scale and asked the participant if they had attacked someone in the past year, while the other item asked if they used force (may have included a weapon) to take money or things from someone else.

Fig. 1. Measurement model.

332

D. Cantillon et al. / Journal of Criminal Justice 31 (2003) 321–339

Table 4 Correlation matrix of independent and mediating variables

also had three components: emotion, action, and safety. The dependent variable consisted of the positive youth outcomes of conventional activity and grade point average, as well as the negative outcome of selfreported delinquency, which had four components. For the first regression equation, the independent variable must significantly predict the mediating variable. Second, the mediating variable must be a significant predictor of the dependent variable. Finally, when the dependent variable was regressed on both the independent and mediating variables, the relationship between the dependent variable and the mediating variable (e.g., emotion) must remain significant while the relationship between the dependent variable and the independent variable (e.g., block stability) must reduce to a nonsignificant level. Therefore, for organizational purposes, results will be explained in a stepwise fashion following Baron and Kenny’s method. Finally, to reduce the number of regression equations required to establish a mediating model, correlation analyses were first conducted to assess where it was necessary to proceed to regression analyses (please see Fig. 1).

Mediating Independent variables variables Block Block Block Block income homogeneity stability composition SOC Emotion Action Safety

.18* .06 .15 .28**

.28** .25** .21* .26**

.49** .39** .37** .57**

.15 .02 .18* .09

SOC = total score across the three components of emotion, action, and safety. * P < .05. ** P < .01, one-tailed test.

Conventional activity This outcome variable was measured with one item which directly asked the youth for the total number (if any) of school activities, clubs, sport teams, etc., of which he was a member in the past year. Grade point average This outcome variable was measured by a question which directly asked for the youth’s current grade point average.

Step 1: relationship between the independent and mediating variables As can be seen from the correlation table (please see Table 4), the majority of the independent variables significantly correlated with the majority of the mediating variables. Block homogeneity and block stability significantly correlated with all components of SOC, while block income only significantly correlated with safety. Block composition was only significantly correlated with action, although overall, strong patterns of relationships among the independent and mediating variables emerged during this first step in evaluating a mediating model.

Results Data analytic strategy According to Baron and Kenny (1986), three regression equations are required to test for mediating relationships. In the current study, neighborhood advantage was the independent variable and was comprised of block stability, block income, block homogeneity, and block composition. The mediating variable, SOC,

Table 5 Correlation matrix of independent and mediating variables with dependent variables Variables

Dependent variables Steal and deal delinquency

School delinquency

Steal and fight delinquency

Severe delinquency

Conventional activity

Average grade

Independent variables Income .05 Stability .02 Homogeneity .15 Composition .08

.25** .16 .06 .02

.10 .25** .12 .08

.02 .31** .05 .07

.22** .19* .07 .01

.12 .08 .02 .10

Mediating variables Emotion .14 Action .23** Safety .14

.08 .04 .03

.01 .07 .05

.19* .08 .11

.20* .23** .20*

.05 .23** .02

Bolded correlations indicate potential mediating models which require further regression analyses. * P < .05. ** P < .01, one-tailed test.

D. Cantillon et al. / Journal of Criminal Justice 31 (2003) 321–339

Step 2: relationship between the independent, mediating, and dependent variables SOC and its components significantly correlated with many of the dependent variables as can be seen from Table 5. Overall, patterns among these constructs conformed to the prediction that high levels of neighborhood advantage and SOC would be related, in direction and magnitude, to positive youth outcomes, while low levels of these variables would be related to negative outcomes. For instance, all but two of the independent and mediating variables positively correlated with conventional activity and many of these relationships were significant. The highlighted correlation coefficients indicated where possible mediating relationships might have existed. As can be seen in this table, there were only two outcome variables, severe delinquency and conventional activity, that corresponded to the directional hypothesis that SOC mediated the effect of neighborhood advantage on youth outcomes. Thus, only for these two variables was it necessary to proceed to regression analyses to assess whether or not SOC mediated the influence of neighborhood advantage on youth outcomes. Step 3: evaluating the mediating role of SOC on youth outcomes For severe delinquency, correlation and regression analyses indicated that the effect of block stability on this outcome might be mediated by the emotion component of SOC. In the combined regression equation, block stability (b = .26) retained its significant effect, while emotion dropped to an insignificant level (see Table 6). This result disconfirmed the possibility of a mediating relationship. For conventional activity, the independent variable of block stability and all three components of SOC were significantly related to this outcome variable. The independent variable of block income also significantly correlated with conventional activity, but there was only one path to test since block income was only significantly related to the safety component of SOC. In sum, there was a total of four multiple regression equations performed to evaluate the mediating hypothesis in the case of conventional activity (please see Table 7). Table 6 OLS regression estimates of severe delinquency Regression equation Eq. (1) Block stability Emotion

Severe delinquency B

S.E. B .06 .02

* P < .05, one-tailed test.

.03 .02

B

t ratio .26 .09

2.52* .86

333

Table 7 OLS regression estimates of conventional activity Regression equations Eq. (1) Block stability Emotion Eq. (2) Block stability Action Eq. (3) Block stability Safety Eq. (4) Block income Safety

Conventional activity B

S.E. B

B

t ratio

.06 .05

.04 .03

.14 .15

1.28 1.45

.05 .05

.04 .03

.13 .18

1.23 1.72 *

.05 .13

.05 .12

.12 .13

1.02 1.09

.00 .15

.00 .10

.18 .15

1.81 * 1.47

Eqs. (1) – (4) represent all possible mediating paths for conventional activity. * P < .05, one-tailed test.

For the independent variable of block stability, one of the three components of SOC proved to be a significant mediator (Eqs. (1) – (3)). Specifically, action (b=.18) significantly mediated the effect of block stability on conventional activity. Again, there was only one path to test for an intervening relationship between the SOC construct and block income on conventional activity. The results disconfirmed a possible mediating relationship as block income remained significant (b=.18), while safety decreased to an insignificant level (b=.15).

Discussion and implications The results from the current study demonstrated that SOC mediated the effect of neighborhood advantage on conventional activity. Specifically, the action component of SOC mediated the effect of block stability. In this study, conventional activity was defined as the number of school activities in which students participated. This finding indicated that youth reared in communities characterized by high levels of SOC, especially those with an action-oriented or empowering component, were more likely to participate in prosocial behavior such as school activities. Since participation in school activities was the best predictor of grade point average, on average, youth from high SOC neighborhoods received better grades than youth from low SOC neighborhoods. High schools bring together youth from many surrounding communities and the results from the current study suggested the importance of the local neighborhood environment in determining youths’ bonding to, participation in, and ultimately, success in school. Thus, the current study demonstrated that close-knit neighborhoods with high

334

D. Cantillon et al. / Journal of Criminal Justice 31 (2003) 321–339

levels of SOC had important spillover effects in other youth developmental contexts. The results, however, also indicated that SOC did not significantly mediate the relationship between neighborhood advantage and self-reported delinquency. As demonstrated by Simcha-Fagan and Schwartz (1986) though, the social disorganization framework explained the relationship between neighborhood disadvantage and crime and delinquency better when official data were analyzed. Unfortunately, since the current study did not have access to juvenile records, this line of inquiry could not be investigated. There were also a couple of other possible reasons for this null finding in regard to the selfreported delinquency variables. First, the sample size of the current study was small (N = 103) and this decreased the power to detect significant effects. Second, this was a city-based study and the most extensive review of the literature to date on neighborhood effects found that national and multisite studies had a much better probability of finding significant effects for the neighborhood context (Levanthal & Brooks-Gunn, 2000). This was due to larger sample sizes and greater sampling variability, particularly since these large-scale studies were purposely designed to include neighborhoods on opposite extremes in terms of structural conditions (poverty, residential turnover; Sampson, 1993b). In a city-based study such as this one, higher intercorrelations among neighborhood dimensions lead to a higher probability of a null finding (Duncan & Raudenbush, 1999). The current study was also conducted in a mid-size industrial city, which did not contain pockets of extreme advantaged or disadvantaged neighborhoods typical in cities where neighborhood effects research has been usually conducted. In this manner, the neighborhoods from the current study are probably more representative of the majority of cities across the United States. In sum, SOC was found to be a reliable and valid construct that could be used to measure the mediating variables of social disorganization theory. Importantly, SOC goes beyond measuring the behavioral outcomes of community social organization such as intervention behaviors and taps into the important and subtle social processes that precede such collective behavior. Accordingly, it has been argued that this construct provides a more valid and comprehensive method in which to measure the proposed mediating variables of social disorganization theory. Another advantage of the SOC construct is the applicability of this construct. For instance, if the only outcome variable of a delinquency prevention initiative is informal social control, as measured by intervention behaviors in the community, results could show no change in such behavior, and hence, no program effect when a considerable amount of social

change has actually occurred in the community. Improved measures of community social organization, such as SOC, would be better able to capture this change and even suggest specific areas that need to be targeted to ultimately increase informal social control in the neighborhood. For instance, neighborhoods could be assessed on the three dimensions of SOC (emotion, action, and safety) and specific dimensions could then be targeted for different neighborhoods. While one neighborhood may need to increase general interaction and emotional connection among residents, another neighborhood may have this connection but needs assistance in developing neighborhood leadership to become more empowered and action oriented in addressing their problems. Further, tracking the perceived safety levels of neighborhoods is also vital because fear of crime can decrease the willingness of residents to come together and to intervene when community norms are being violated. For instance, as Korbin and Coulton (1997) documented, the biggest reason for not ‘‘getting involved’’ was fear of retaliation. Thus, if used as an outcome variable in community development and related initiatives, program personnel could gauge progress and even identify important and specific community dimensions which need to be increased in order for the community to reduce local problems and enhance positive youth development. The three-step procedure to establish a mediating relationship also displayed a number of interesting results. First, block income was only significantly related to safety. Thus, in the current study, while lower-income neighborhoods were not perceived to be as safe as other neighborhoods, income level did not have a significant impact on the emotional attachment residents felt toward their neighborhood or on residents’ participation in neighborhood organizations and activities. Meanwhile, block stability was significantly related to all dimensions of SOC—emotion, action, and safety. Correlations ranged from .37 to .57 and thus dwarfed the effects of income in this study. This finding points to the importance of not limiting analyses of neighborhood effects to just the income level or socioeconomic status (SES) of the neighborhood. In this study, block instability exerted considerably more deleterious effects than low income. Nevertheless, congruent with social disorganization theory, the main point is that it is not simply residential stability or income level, or any other independent variable, but the combination and confluence of these variables that usually leads to the extreme disparity between disadvantaged and advantaged neighborhoods in the United States. Second, the current study found that block composition only significantly correlated with the action component of SOC. The sample for this study had a large percentage of dual-parent families in contrast to some neighborhood studies where single-parent fam-

D. Cantillon et al. / Journal of Criminal Justice 31 (2003) 321–339

ilies predominated within a community. Perhaps, as others hypothesized, there is a tipping point or threshold that must be surpassed before negative structural characteristics exert a detrimental impact on neighborhood-level and individual-level youth outcomes (Aneshensel & Sucoff, 1996; Crane, 1991; Krivo & Peterson, 1996; Simons et al., 1996). Another intriguing finding was how the more minor components of self-reported delinquency were related to the independent and mediating variables. For instance, both steal and deal delinquency and steal and fight delinquency were positively related to all three components of SOC. In the case of steal and deal delinquency and the action component of SOC, this relationship was not only in an unanticipated direction, but it was also significant. While somewhat alarming and counterintuitive, this finding is not too surprising when one considers the perpetual debate in the field of criminology as to what self-report scales and official records of delinquency measured (Hagan, Gillis, & Chan, 1978; Sampson, 1986). Since steal and deal delinquency, relatively speaking, was on the minor side of the delinquency components, it certainly could follow that neighborhood structural advantage and SOC levels had little to no influence on this type of offending pattern compared to its role in more severe delinquency. There were some limitations to this study that warrant attention. First, the cross-sectional nature of the study provided only a snapshot of community dynamics and needs to be replicated with a longitudinal design, particularly since the social disorganization framework was conceptualized to explain the impact of structural conditions (poverty, residential turnover) on neighborhood-level social processes, both of which were dynamic and constantly changing. Second, results from the principal component analysis of the SOC construct were somewhat inconclusive, and thus, reliability analyses were also conducted at the individual and block levels. Yet, this finding was in line with current research and indicative of the controversy in the field regarding the dimensionality of the SOC construct (Chipuer & Pretty, 1999; Hill, 1996; Skjaeveland, Garling, & Maeland, 1996). The SOC scale utilized in the present study included three theoretically driven and empirically supported components of emotion, action, and safety. While some previous SOC scales included items or constructs such as community participation and safety dimensions (Davidson & Cotter, 1986; Doolittle & Macdonald, 1978; Kingston et al., 1999; McGuire, 1997), not all SOC scales included these dimensions as integral to the concept (Buckner, 1988; Chavis et al., 1986; Riger & Lavrakas, 1981). Also, while the importance of emotional safety and security

335

was noted in prior conceptualizations (McMillan & Chavis, 1986), it seemed that physical security and safety must be included in delinquency and related research, particularly since youth were at the greatest risk of victimization (Snyder & Sickmund, 1999), and because most delinquency prevention and community development programs were ultimately aimed at in-creasing positive youth outcomes. A limitation of the safety measure used in this study was that there were only two items, and the principal components analysis displayed that these items also loaded on the emotion and action components. Future research should include more safety items and investigate if it is truly a dimension of SOC, or if it should be conceptualized and measured as a distinct variable, albeit a necessary precursor for the development of SOC. Both steal and fight and steal and deal delinquency suffered from reliability problems. A strength of the self-reported delinquency construct, however, was that the principal components analysis demonstrated that minor and severe delinquency represented separate dimensions, thus, minor infractions did not drive any significant results on the more severe delinquency measures. Finally, while the names of the various components of delinquency were descriptive, they should not be interpreted too strictly. For instance, school delinquency only contained three items: two of which asked about cutting classes and school, and a final item that was not school related. Thus, this was not a comprehensive measure of school delinquency per se. The sole purpose for these labels was to attempt to distinguish between the selfreported delinquency components by naming each variable by its major determinant. Since block income was utilized versus the traditional method of measuring SES in social disorganization studies, caution must be taken in the interpretation of the role SES exerted on the level of SOC within a neighborhood and its influence on youth outcomes such as delinquency. Finally, due to cost and time considerations, only two surveys were collected and aggregated to construct summary measures for the face block. Future studies should incorporate more respondents per block in order to utilize multilevel modeling procedures to obtain more reliable estimates and to be able to differentiate or partial individual- from neighborhood-level effects. In closing, as updated systemic models of social disorganization have stressed, the social organization of a community is dependent upon ‘‘. . . the extensiveness and density of the formal and informal networks within the neighborhood that bind the residents together as a social community’’ (italics added; Bursik & Grasmick, 1993, p. 4). SOC measures this affective component, which differentiates

336

D. Cantillon et al. / Journal of Criminal Justice 31 (2003) 321–339

whether someone lives in a neighborhood or is truly part of a community. As displayed by the results of this study, this qualitative difference not only had a substantial impact on residents’ quality of life, but also on both positive and negative youth outcomes.

Acknowledgements This research was funded by the Institute for Children, Youth, and Families at Michigan State University.

Appendix A. Principle component analysis of SOC construct

SOC items

Emotion

People on this block. . . Know each other Socialize with each other Think of themselves as a community Feel connected to each other Feel isolated from each other Watch out for each other Would give rides to each other Take care of each others’ plants, etc. Participate in community improvement Have a voice . . . community issues Participate in neighborhood organizations Participate in social activities Never do things to improve the block Feeling of community spirit exists Talk to each other . . . community Persuade the city to respond Don’t trust each other Don’t care about the block’s future Feel they belong here Make it a safer place to live Hard to get help from neighbors Unable to create a solution Fairly safe to walk on block at night Eigenvalues

.93 .78 .71 .70 .62 .44 .40 .31 .02 .04 .03 .32 .09 .26 .15 .15 .25 .03 .15 .08 .30 .02 .36 8.72

Action

Safety

h2

.05 .09 .09 .09 .08 .08 .05 .23 .89 .73 .70 .59 .54 .51 .48 .46 .46 .17 .06 .23 .11 .33 .08 2.19

.36 .18 .03 .14 .19 .37 .23 .16 .22 .04 .15 .29 .27 .19 .04 .39 .80 .76 .66 .59 .59 .46 .40 1.60

.99 .65 .51 .52 .42 .34 .22 .17 .84 .54 .51 .53 .37 .36 .25 .39 .91 .61 .47 .41 .45 .32 .30

Principal component analysis with promax rotation. h2 equals the sum of squared component loadings for each item.

Appendix B. Principle component analysis of self-reported delinquency construct Self-reported delinquency items Stolen >US$50 but < US$100 Taken some part of a car Sold or dealt drugs Stolen >US$100 Stolen something US$5 or less Skipped a full day of school Skipped class while in school Carried weapon outside of school Fought someone physically Bought, sold, or held stolen goods Gone into building to steal Used force to take US$ or things Attacked someone Eigenvalues

Steal and deal .84 .81 .79 .60 .60 .05 .02 .04 .04 .49 .03 .04 .08 3.56

School .04 .17 .07 .06 .07 .93 .90 .78 .09 .07 .10 .02 .15 2.55

Steal and fight .17 .13 .09 .16 .34 .02 .14 .05 .86 .78 .71 .02 .05 1.81

Severe

h2

.05 .02 .23 .09 .09 .03 .02 .03 .17 .01 .07 .92 .90 1.71

.74 .70 .69 .39 .49 .87 .83 .61 .78 .85 .52 .85 .84

Principal component analysis with varimax rotation. h2 equals the sum of squared component loadings for each item.

D. Cantillon et al. / Journal of Criminal Justice 31 (2003) 321–339

References Aneshensel, C. S., & Sucoff, C. A. (1996). The neighborhood context of adolescent mental health. Journal of Health and Social Behavior, 37, 293 – 310. Appleyard, D. (1981). Livable streets. Berkeley: University of California Press. Bardo, J., & Bardo, D. (1983). A re-examination of subjective components of community satisfaction in a British new town. Journal of Social Psychology, 120, 35 – 43. Barnes, J. B. (1997). The reliability and validity of a questionnaire describing neighborhood characteristics relevant to families and young children living in urban areas. Journal of Community Psychology, 25(6), 551 – 566. Baron, R., & Kenny, D. A. (1986). The moderator – mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51(6), 1173 – 1182. Bellair, P. E. (1997). Social interaction and community crime: Examining the importance of neighbor networks. Criminology, 35(4), 677 – 701. Bellair, P. E. (2000). Informal surveillance and street crime: A complex relationship. Criminology, 38(1), 137 – 168. Bishop, P. D., Chertok, R., & Jason, L. A. (1997). Measuring sense of community: Beyond local boundaries. Journal of Primary Prevention, 18(2), 193 – 212. Black, D. (1989). Toward a general theory of social control (Vol. 4). Orlando, FL: Academic Press. Brodsky, A. E., O’Campo, P. J., & Aronson, R. E. (1996). PSOC in community context: Multi-level correlates of a measure of psychological sense of community in lowincome, urban neighborhoods. Journal of Community Psychology, 27(6), 659 – 679. Brooks-Gunn, J., Duncan, G. J., Klebanov, P. K., & Sealand, N. (1993). Do neighborhoods influence child and adolescent development? American Journal of Sociology, 99 (2), 353 – 395. Buckner, J. C. (1988). The development of an instrument to measure neighborhood cohesion. American Journal of Community Psychology, 16(6), 771 – 791. Bursik, R. J., Jr. (1988). Social disorganization and theories of crime and delinquency: Problems and prospects. Criminology, 26(4), 519 – 551. Bursik, R. J., Jr., & Grasmick, H. G. (1993). Neighborhoods and crime: The dimensions of effective community control. New York: Lexington Books. Chavis, D. M., Hogge, J. H., McMillan, D. W., & Wandersman, A. (1986). Sense of community through Brunswick’s lens: A first look. Journal of Community Psychology, 14, 24 – 40. Chavis, D. M., & Pretty, G. M. H. (1999). Sense of community: Advances in measurement and application. Journal of Community Psychology, 27(6), 635 – 642. Chavis, D. M., & Wandersman, A. (1990). Sense of community in the urban environment: A catalyst for participation and community development. American Journal of Community Psychology, 18(1), 55 – 81. Chipuer, H. M., & Pretty, G. M. H. (1999). A review of the sense of community index: Current uses, factor struc-

337

ture, reliability, and further development. Journal of Community Psychology, 27(6), 643 – 658. Chipuer, H. M., Pretty, G. H., Delorey, E., Miller, M., Powers, T., Rumstein, O., Barnes, A., Cordasic, N., & Laurent, K. (1999). The neighbourhood youth inventory: Development and validation. Journal of Community and Applied Social Psychology, 9, 355 – 368. Crane, J. (1991). The epidemic theory of ghetto and neighborhood effects on dropping out and teenage childbearing. American Journal of Sociology, 95, 1226 – 1259. Crew, T. B., Kim, J. W., & Schweitzer, J. H. (1999). Caring for our neighborhoods: Exploring leadership and participation potential in urban communities in Lansing, Michigan. Journal of the Community Development Society, 30(1), 15 – 28. Davidson, W. B., & Cotter, P. R. (1986). Measurement of sense of community within the sphere of city. Journal of Applied Social Psychology, 16(7), 608 – 619. Davidson, W. B., & Cotter, P. R. (1989). Sense of community and political participation. Journal of Community Psychology, 17, 119 – 125. Davidson, W. B., & Cotter, P. R. (1991). The relationship between sense of community and subjective well-being: A first look. Journal of Community Psychology, 19, 246 – 253. Davidson, W. B., & Cotter, P. R. (1993). Psychological sense of community and support for public school taxes. American Journal of Community Psychology, 21(1), 59 – 66. Doolittle, R. J., & Macdonald, D. (1978). Communication and a sense of community in a metropolitan neighborhood: A factor analytic examination. Communication Quarterly, 26(3), 2 – 8. Duncan, G. J., & Raudenbush, S. W. (1999). Assessing the effects of context in studies of child and youth development. Educational Psychologist, 34(1), 29 – 41. Elliott, D. S., Wilson, W. J., Huizinga, D., Sampson, R. J., Elliott, A., & Rankin, B. (1996). The effects of neighborhood disadvantage on adolescent development. Journal of Research in Crime and Delinquency, 33(4), 389 – 426. Glynn, T. J. (1981). Psychological sense of community: Measurement and application. Human Relations, 34(7), 789 – 818. Granovetter, M. (1973). The strength of weak ties hypothesis. American Journal of Sociology, 78(6), 1360 – 1380. Greenberg, S. W., & Rohe, W. M. (1986). Informal social control and crime prevention in modern urban neighborhoods. In R. B. Taylor (Ed.), Urban neighborhoods: Research and policy ( pp. 79 – 118). New York: Praeger. Hackler, J. C., Ho, K. Y., & Urquhart-Ross, C. (1974). The willingness to intervene: Differing community characteristics. Social Problems, 21(3), 328 – 344. Hagan, J., Gillis, A. R., & Chan, J. (1978). Explaining official delinquency: A spatial study of class, conflict, and control. Sociological Quarterly, 12, 386 – 398. Heitgerd, J. L., & Bursik, R. J., Jr. (1987). Extracommunity dynamics and the ecology of delinquency. American Journal of Sociology, 92, 775 – 787. Hill, J. L. (1996). Psychological sense of community: Suggestions for future research. Journal of Community Psychology, 24(4), 431 – 438.

338

D. Cantillon et al. / Journal of Criminal Justice 31 (2003) 321–339

Hunter, A. (1985). Private, parochial and public social orders: The problem of crime and incivility in urban communities. In M. N. Zald (Ed.), The challenge of social control: Citizenship and institution building in modern society ( pp. 230 – 242). Norwood, NJ: Ablex Publishing. Kasarda, J. D., & Janowitz, M. (1974). Community attachment in mass society. American Sociological Review, 39, 328 – 339. Kingston, S., Mitchell, R., Florin, P., & Stevenson, J. (1999). Sense of community in neighborhoods as a multi-level construct. Journal of Community Psychology, 27(6), 681 – 694. Korbin, J. E., & Coulton, C. J. (1997). Understanding the neighborhood context for children and families: Combining epidemiological and ethnographic approaches. In J. Brooks-Gunn, G. J. Duncan, & J. L. Aber (Eds.), Neighborhood poverty: Policy implications in studying neighborhoods (Vol. 2, pp. 65 – 80). New York: Russell Sage Foundation. Kornhauser, R. R. (1978). Social sources of delinquency. Chicago: University of Chicago Press. Krivo, L. J., & Peterson, R. D. (1996). Extremely disadvantaged neighborhoods and urban crime. Social Forces, 75(2), 619 – 642. Krupat, E., & Guild, W. (1980). The measurement of community social climate. Environment and Behavior, 12(2), 195 – 206. Lalli, M. (1992). Urban related identity: Theory, measurement, and empirical findings. Journal of Environment Psychology, 12(4), 285 – 303. Levanthal, T., & Brooks-Gunn, J. (2000). The neighborhoods they live in: The effects of neighborhood residence on child and adolescent outcomes. Psychological Bulletin, 126(2), 309 – 337. Macoby, E. E., Johnson, J. P., & Church, R. M. (1958). Community integration and the social control of juvenile delinquency. Journal of Social Issues, 14, 38 – 51. Martinez, M. L., Black, M., & Starr, R. H. (2002). Factorial structure of the perceived neighborhood scale (PNS): A test of longitudinal invariance. Journal of Community Psychology, 30(1), 23 – 43. McGuire, J. B. (1997). The reliability and validity of a questionnaire describing neighborhood characteristics relevant to families and young children living in urban areas. Journal of Community Psychology, 25(6), 551 – 566. McMillan, D. W., & Chavis, D. M. (1986). Sense of community: A definition and theory. Journal of Community Psychology, 14, 6 – 23. Moos, R., & Gerst, M. (1974). University residence environment scale. Palo Alto, CA: Consulting Psychologists Press. Nasar, J. L., & Julian, D. A. (1995). The psychological sense of community in the neighborhood. Journal of the American Planning Association, 61, 178 – 184. Obst, P., Smith, S. G., & Zinkiewicz, L. (2002). An exploration of sense of community: Part 3. Dimensions and predictors of psychological sense of community in geographical communities. Journal of Community Psychology, 30(1), 119 – 133.

Park, R. E., Burgess, E. W., & McKenzie, R. D. (1967). The City (Rev. ed.). Chicago: University of Chicago Press. Perkins, D. D., Florin, P., Rich, R. C., Wandersman, A., & Chavis, D. M. (1990). Participation and the social and physical environment of residential blocks: Crime and community context. American Journal of Community Psychology, 18(1), 83 – 115. Perkins, D. D., & Taylor, R. B. (1996). Ecological assessments of community disorder: Their relationship to fear of crime and theoretical implications. American Journal of Community Psychology, 24(1), 63 – 106. Pretty, G. M. H. (1990). Relating psychological sense of community to social climate characteristics. Journal of Community Psychology, 18, 60 – 65. Pretty, G. M. H., Andrews, L. A., & Collett, C. (1994). Exploring adolescents’ sense of community and its relationship to loneliness. Journal of Community Psychology, 22, 346 – 358. Pretty, G. M. H., Conroy, C., Dugay, J., Fowler, K., & Williams, D. (1996). Sense of community and its relevance to adolescents of all ages. Journal of Community Psychology, 24(4), 365 – 379. Prezza, M., Amici, M., Roberti, T., & Tedeschi, G. (2001). Sense of community referred to the whole town: Its relations with neighboring, loneliness, life satisfaction, and area of residence. Journal of Community Psychology, 29(1), 29 – 52. Puddifoot, J. E. (1996). Some initial considerations in the measurement of community identity. Journal of Community Psychology, 24 (4), 327 – 336. Riger, S., & Lavrakas, P. J. (1981). Community ties: Patterns of attachment and social interaction in urban neighborhoods. American Journal of Community Psychology, 9(1), 55 – 66. Sampson, R. J. (1986). Neighborhood family structure and the risk of personal victimization. In R. J. Sampson (Ed.), The social ecology of crime ( pp. 25 – 46). New York: Springer-Verlag. Sampson, R. J. (1993a). Linking time and place: Dynamic contextualism and the future of criminological theory. Journal of Research in Crime and Delinquency, 30(4), 426 – 444. Sampson, R. J. (1993b). Family and community-level influences on crime: A contextual theory and strategies for research testing. In D. P. Farrington, R. J. Sampson, & P. H. Wikstrom (Eds.), Integrating individual and ecological aspects of crime ( pp. 153 – 168). Stockholm: National Council for Crime Prevention. Sampson, R. J. (1997). Collective regulation of adolescent misbehavior: Validation results from eighty Chicago neighborhoods. Journal of Adolescence Research, 12 (2), 227 – 244. Sampson, R. J., & Groves, W. B. (1989). Community structure and crime: Testing social disorganization theory. American Journal of Sociology, 94(4), 774 – 802. Sampson, R. J., Morenoff, J. D., & Earls, F. (1999). Spatial dynamics of collective efficacy for children. American Sociological Review, 64, 633 – 660. Sampson, R. J., Raudenbush, S. W., & Earls, F. (1997). Neighborhoods and violent crime: A multilevel study of collective efficacy. Science, 277, 918 – 924.

D. Cantillon et al. / Journal of Criminal Justice 31 (2003) 321–339 Sarason, S. B. (1974). The psychological sense of community: Prospects for a community psychology. San Francisco: Jossey-Bass. Schweitzer, J. H., Kim, J. W., & Mackin, J. R. (1999). The impact of the built environment on crime and fear of crime. Journal of Urban Technology, 6(3), 59 – 73. Shaw, C. R., & McKay, H. D. (1942). Juvenile delinquency and urban areas. Chicago: University of Chicago Press. Simcha-Fagan, O., & Schwartz, J. E. (1986). Neighborhood and delinquency: Assessment of contextual effects. Criminology, 24(4), 667 – 698. Simons, R. L., Johnson, C., Beaman, J., Conger, R. D., & Whitbeck, L. B. (1996). Parents and peer group as mediators of the effect of community structure on adolescent problem behavior. American Journal of Community Psychology, 24(1), 145 – 171. Skjaeveland, O., Garling, T., & Maeland, J. G. (1996). A multidimensional measure of neighboring. American Journal of Community Psychology, 24(3), 413 – 435. Snyder, H. N., & Sickmund, M. (1999). Juvenile offenders and victims: 1999 national report. Washington, DC: Office of Juvenile Justice and Delinquency Prevention. Taylor, R. B. (1997). Social order and disorder of street blocks and neighborhoods: Ecology, microecology, and the systemic model of social disorganization. Journal of Research in Crime and Delinquency, 34(1), 113 – 155.

339

Unger, D. G., & Wandersman, A. (1982). Neighboring in an urban environment. American Journal of Community Psychology, 10(5), 493 – 509. Unger, D. G., & Wandersman, A. (1983). Neighboring and its role in block organizations: An exploratory report. American Journal of Community Psychology, 11(3), 291 – 300. Unger, D. G., & Wandersman, A. (1985). The importance of neighbors: The social, cognitive, and affective components of neighboring. American Journal of Community Psychology, 13(2), 139 – 169. Veysey, B. M., & Messner, S. F. (1999). Further testing of social disorganization theory: An elaboration of Sampson & Grove’s ‘‘community structure and crime’’. Journal of Research in Crime and Delinquency, 36(2), 156 – 174. Warner, B. D., & Rountree, P. W. (1997). Local social ties in a community and crime model: Questioning the systemic nature of informal social control. Social Problems, 44 (4), 520 – 536. Warren, D. I. (1978). Explorations in neighborhood differentiation. Sociological Quarterly, 19, 310 – 331. Zani, B., Cicognani, E., & Albanesi, C. (2001). Adolescents’ sense of community and feeling of unsafety in the urban environment. Journal of Community and Applied Psychology, 11, 475 – 489.