Criminal Justice Social Disorganization

American Association for the Advancement of Science is collaborating with JSTOR to digitize, preserve and extend access to Science.

http://www.jstor.org

Neighborhoods and Violent Crime: A Multilevel Study of Collective Efficacy Author(s): Robert J. Sampson, Stephen W. Raudenbush and Felton Earls Source: Science, New Series, Vol. 277, No. 5328 (Aug. 15, 1997), pp. 918-924 Published by: American Association for the Advancement of Science Stable URL: http://www.jstor.org/stable/2892902 Accessed: 26-01-2016 18:42 UTC

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at http://www.jstor.org/page/ info/about/policies/terms.jsp

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

This content downloaded from 129.137.5.42 on Tue, 26 Jan 2016 18:42:44 UTC All use subject to JSTOR Terms and Conditions

-U---

the north central Pacific, the estimated residence time of 21Pb at mid-depths is 200 to 400 years, decreasing to 80 to 100 years toward the Pacific margins [Y. Nozaki , K. K. Turekian, K. von Damm, Earth Planet. Sci. Lett. 49, 393 (1980); H. Craig, S. Krishnaswami, B. L. K. Somayajulu, ibid. 17, 295 (1973)]. Given the observed differences in the Pb isotopic composition of Mn crusts and nodules from different ocean basins (18), the residence time must be less than the -103-year mixing time of the oceans [W. S. Broecker and T.-H. Peng, Tracers in the Sea (Eldigio Press, Columbia Univ., Palisades, NY, 1982)] but in the central Pacific may be suffi- ciently long to mix and integrate differing inputs from incoming water masses and the basin margins. Studies of 210Pb indicate residence times in the up- per ocean of --10 years [R. M. Sherrell, E. A. Boyle, B. Hamelin, J. Geophys. Res. 97, 11257 (1992)], which are much shorter than the residence time in deep water. The most important mechanism for Pb transport to the deep sea is scavenging by par- ticulates, particularly organic particulates [A. R. Fle- gal and C. C. Patterson, Earth Planet. Sci. Lett. 64, 19 (1983)], which may have varied considerably in the geologic past because of changes in biologic productivity.

17. T. J. Chow and C. C. Paterson, Geochim. Cosmo- chim. Acta 26, 263 (1962).

18. W. Abouchami and S. L. Goldstein, ibid. 59, 1809 (1995).

19. Y. Asmerom and S. B. Jacobsen, Earth Planet. Sci. Lett. 115, 245 (1993).

20. C. E. Jones, A. N. Halliday, D. K. Rea, R. M. Owen, Eos 77, 320 (1996).

21. M. Tatsumoto, Earth Planet. Sci. Lett. 38, 63 (1978). 22. T. J. Barrett, P. N. Taylor, J. Lugowski, Geochim.

Cosmochim. Acta 51, 2241 (1987). 23. L. V. Godfrey, R. A. Mills, H. Elderfield, E. Gurvich,

Mar. Chem. 46, 237 (1994); L. V. Godfrey et al., Earth Planet. Sci. Lett., in press.

24. D. K. Rea, Rev. Geophys. 32, 159 (1994). 25. T. Pettke, D. K.-Rea, A. N. Halliday, in Seventh An-

nual V. M. Goldschmidt Conference Abstracts, LPI contribution 921, 163 (Lunar and Planetary Institute, Houston, TX, 1997).

26. T. H. van Andel, Earth Planet. Sci. Lett. 26, 187 (1975).

27. D. K. Rea, GSA Today 3, 207 (1993). 28. , AGU Monogr. 70, 387 (1992). 29. J. D. Blum, Y. Erel, K. Brown, Geochim. Cosmo-

chim. Acta 57, 5019 (1994); J. D. Blum and Y. Erel, Nature 373, 415 (1995).

30. B. Peucker-Ehrenbrink and J. D. Blum, Eos 77, 324 (1996).

31. Y. Erel, Y. Harlavan, J. D. Blum, Geochim. Cosmo- chim. Acta 58, 5299 (1994).

32. M. Andree et al., Clim. Dyn. 1, 53 (1986). 33. J. C. Zachos, L. D. Stott, K. C. Lohmann, Pale-

oceanography 9, 353 (1994). 34. J. P. Kennet, G. Keller, M. S. Srinivason, GSA Mem-

oir No. 163, 197 (1985). 35. C. R. Scotese and J. Golonka, PALEOMAP Paleo-

geographic Atlas (PALEOMAP Progress Report 20, Department of Geology, Univ. of Texas at Arlington, 1992); D. B. Walsh and C. R. Scotese, Plate Tracker version 1.2 (Department of Geology, Univ. of Texas at Arlington, 1995).

36. F. Woodruff and S. Savin, Paleoceanography 4, 87 (1989).

37. L. Keigwin, Science 217, 350 (1982). 38. M. G. Gross, Oceanography (Prentice-Hall, Engle-

wood Cliffs, NJ, ed. 6,1993). 39. D. A. Hodell, J. A. Mueller, J. A. McKenzie, G. A.

Mead, Earth Planet. Sci. Lett. 92, 165 (1989); D. A. Hodell, G. A. Mead, J. A. Mueller, Chem. Geol. 80, 291 (1990); D. A. Hodell, J. A. Mueller, J. R. Garrido, Geology 19, 24 (1991); J. S. Oslick, K. G. Miller, M. D. Feigenson, J. D. Wright, Paleoceanography 9, 427 (1994); K. G. Miller, M. D. Feigenson, D. V. Kent, R. K. Olsson, ibid. 3, 223 (1988), K. G. Miller, M. D. Feigenson, J. D. Wright, B. M. Clement, ibid. 6, 33 (1991).

40. W. M. White, A. W. Hofmann, H. Puchelt, J. Geo- phys. Res. 92, 4881 (1987).

41. This research was supported by a grant from the

U.S. Department of Energy. We thank F. von Blanck- enburg and R. K. O'Nions for sending us a preprint describing their Pb isotopic work; B. Wilkinson, W. Broecker, S. Epstein, S. Savin, and K. Burton for helpful discussions; M. Johnson and C. M. Hall for

technical help; Cetac Technologies for use of one of their laser systems; and F. von Blanckenburg and an anonymous reviewer for their comments.

16 December 1996; accepted 19 June 1997

Neighborhoods and Violent Crime: A Multilevel Study of

Collective Efficacy Robert J. Sampson, Stephen W. Raudenbush, Felton Earls

It is hypothesized that collective efficacy, defined as social cohesion among neighbors combined with their willingness to intervene on behalf of the common good, is linked to reduced violence. This hypothesis was tested on a 1995 survey of 8782 residents of 343 neighborhoods in Chicago, Illinois. Multilevel analyses showed that a measure of col- lective efficacy yields a high between-neighborhood reliability and is negatively asso- ciated with variations in violence, when individual-level characteristics, measurement error, and prior violence are controlled. Associations of concentrated disadvantage and residential instability with violence are largely mediated by collective efficacy.

For most of this century, social scientists have observed marked variations in rates of criminal violence across neighborhoods of U.S. cities. Violence has been associated with the low socioeconomic status (SES) and residential instability of neighborhoods. Although the geographical concentration of violence and its connection with neighbor- hood composition are well established, the question remains: why? What is it, for exam- ple, about the concentration of poverty that accounts for its association with rates of vi- olence? What are the social processes that might explain or mediate this relation (1 -3)? In this article, we report results from a study designed to address these questions about crime and communities.

Our basic premise is that social and or- ganizational characteristics of neighbor- hoods explain variations in crime rates that are not solely attributable to the aggregated demographic characteristics of individuals. We propose that the differential ability of neighborhoods to realize the common val- ues of residents and maintain effective so- cial controls is a major source of neighbor- hood variation in violence (4, 5). Although social control is often a response to deviant behavior, it should not be equated with formal regulation or forced conformity by

R. J. Sampson is in the Department of Sociology, Univer- sity of Chicago, Chicago, IL, 60637 and is a Research Fellow of the American Bar Foundation, Chicago, IL 60611, USA. S. W. Raudenbush is at the College of Education, Michigan State University, East Lansing, Ml 48824, USA. F. Earls is the Principal Investigator of the Project on Human Development in Chicago Neighbor- hoods and is at the School of Public Health, Harvard University, Boston, MA 02115, USA.

institutions such as the police and courts. Rather, social control refers generally to the capacity of a group to regulate its members according to desired principles-to realize collective, as opposed to forced, goals (6). One central goal is the desire of community residents to live in safe and orderly envi- ronments that are free of predatory crime, especially interpersonal violence.

In contrast to formally or externally in- duced actions (for example, a police crack- down), we focus on the effectiveness of informal mechanisms by which residents themselves achieve public order. Examples of informal social control include the mon- itoring of spontaneous play groups among children, a willingness to intervene to pre- vent acts such as truancy and street-corner "hanging" by teenage peer groups, and the confrontation of persons who are exploiting or disturbing public space (5, 7). Even among adults, violence regularly arises in public disputes, in the context of ille- gal markets (for example, prostitution and drugs), and in the company of peers (8). The capacity of residents to control group- level processes and visible signs of social disorder is thus a key mechanism influenc- ing opportunities for interpersonal crime in a neighborhood.

Informal social control also generalizes to broader issues of import to the well-being of neighborhoods. In particular, the differ- ential ability of communities to extract re- sources and respond to cuts in public ser- vices (such as police patrols, fire stations, garbage collection, and housing code en- forcement) looms large when we consider

918 SCIENCE * VOL. 277 * 15 AUGUST 1997 * www.sciencemag.org

This content downloaded from 129.137.5.42 on Tue, 26 Jan 2016 18:42:44 UTC All use subject to JSTOR Terms and Conditions

_l RESEARCH ARTICLES

the known link between public signs of disorder (such as vacant housing, burned- out buildings, vandalism, and litter) and more serious crime (9).

Thus conceived, neighborhoods differen- tially activate informal social control. It is for this reason that we see an analogy between individual efficacy and neighborhood effica- cy: both are activated processes that seek to achieve an intended effect. At the neighbor- hood level, however, the willingness of local residents to intervene for the common good depends in large part on conditions of mu- tual trust and solidarity among neighbors (10). Indeed, one is unlikely to intervene in a neighborhood context in which the rules are unclear and people mistrust or fear one another. It follows that socially cohesive neighborhoods will prove the most fertile contexts for the realization of informal social control. In sum, it is the linkage of mutual trust and the willingness to intervene for the common good that defines the neighbor- hood context of collective efficacy. Just as individuals vary in their capacity for effica- cious action, so too do neighborhoods vary in their capacity to achieve common goals. And just as individual self-efficacy is situated rather than global (one has self-efficacy rel- ative to a particular task or type of task) (1 1 ), in this paper we view neighborhood efficacy as existing relative to the tasks of supervising children and maintaining public order. It follows that the collective efficacy of resi- dents is a critical means by which urban neighborhoods inhibit the occurrence of per- sonal violence, without regard to the demo- graphic composition of the population.

What Influences Collective Efficacy?

As with individual efficacy, collective effi- cacy does not exist in a vacuum. It is em- bedded in structural contexts and a wider political economy that stratifies places of residence by key social characteristics (12). Consider.the destabilizing potential of rapid population change on neighborhood social organization. A high rate of residential mo- bility, especially in areas of decreasing pop- ulation, fosters institutional disruption and weakened social controls over collective life. A major reason is that the formation of social ties takes time. Financial investment also provides homeowners with a vested interest in supporting the commonweal of neighborhood life. We thus hypothesize that residential tenure and homeownership promote collective efforts to maintain so- cial control (13).

Consider next patterns of resource dis- tribution and racial segregation in the Unit- ed States. Recent decades have witnessed an increasing geographical concentration of

lower income residents, especially minority groups and female-headed families. This neighborhood concentration stems in part from macroeconomic changes related to the deindustrialization of central cities, along with the out-migration of middle-class resi- dents (14). In addition, the greater the race and class segregation in a metropolitan area, the smaller the number of neighborhoods absorbing economic shocks and the more severe the resulting concentration of poverty will be (15). Economic stratification by race and place thus fuels the neighborhood con- centration of cumulative forms of disadvan- tage, intensifying the social isolation of low- er income, minority, and single-parent resi- dents from key resources supporting collec- tive social control (1, 16).

Perhaps more salient is the influence of racial and economic exclusion on perceived powerlessness. Social science research has demonstrated, at the individual level, the direct role of SES in promoting a sense of control, efficacy, and even biological health itself (17). An analogous process may work at the community level. The alienation, ex- ploitation, and dependency wrought by re- source deprivation act as a centrifugal force that stymies collective efficacy. Even if per- sonal ties are strong in areas of concentrated disadvantage, they may be weakly tethered to collective actions.

We therefore test the hypothesis that concentrated disadvantage decreases and residential stability increases collective effi- cacy. In tum, we assess whether collective efficacy explains the association of neigh- borhood disadvantage and residential insta- bility with rates of interpersonal violence. It is our hypothesis that collective efficacy mediates a substantial portion of the effects of neighborhood stratification.

Research Design

This article examines data from the Project on Human Development in Chicago Neigh- borhoods (PHDCN). Applying a spatial def- inition of neighborhood-a collection of people and institutions occupying a subsec- tion of a larger community-we combined 847 census tracts in the city of Chicago to create 343 "neighborhood clusters" (NCs). The overriding consideration in formation of NCs was that they should be as ecologically meaningful as possible, composed of geo- graphically contiguous census tracts, and in- temally homogeneous on key census indica- tors. We settled on an ecological unit of about 8000 people, which is smaller than the 77 established community areas in Chicago (the average size is almost 40,000 people) but large enough to approximate local neigh- borhoods. Geographic boundaries (for exam- ple, railroad tracks, parks, and freeways) and

knowledge of Chicago's neighborhoods guid- ed this process (18).

The extensive racial, ethnic, and so- cial-class diversity of Chicago's population was a major criterion in its selection as a research site. At present, whites, blacks, and Latinos each represent about a third of the city's population. Table 1 classifies the 343 NCs according to race or ethnicity and a trichotomized measure of SES from the 1990 census (19). Although there are no low-SES white neighborhoods and no high-SES Latino neighborhoods, there are black neighborhoods in all three cells of SES, and many heterogeneous neighbor- hoods vary in SES. Table 1 at once thus confirms the racial and ethnic segregation and yet rejects the common stereotype that minority neighborhoods in the Unit- ed States are homogeneous.

To gain a complete picture of the citiy's neighborhoods, 8782 Chicago residents rep- resenting all 343 NCs were interviewed in their homes as part of the community sur- vey (CS). The CS was designed to yield a representative sample of households within each NC, with sample sizes large enough to create reliable NC measures (20). Hence- forth, we refer to NCs as "neighborhoods," keeping in mind that other operational def- initions might have been used.

Measures

"Informal social control" was represented by a five-item Likert-type scale. Residents were asked about the likelihood ("Would you say it is very likely, likely, neither likely nor unlikely, unlikely, or very unlikely?") that their neighbors could be counted on to intervene in various ways if (i) children were skipping school and hanging out on a street corner, (ii) children were spray-paint- ing graffiti on a local building, (iii) children were showing disrespect to an adult, (iv) a

Table 1. Racial and ethnic composition by SES strata: Distribution of 343 Chicago NCs in the PHDCN design.

SES Race or ethnicity

Low Medium High

?75% black 77 37 11 ?75% white 0 5 69 ?75% Latino 12 9 0 ?20% Latino and 6 40 12

?20% white ?20% Latino and 9 4 0

?20% black ?20% blackand 2 4 11

?20% white NCs not classified 8 15 12

above Total 114 114 115

www.sciencemag.org * SCIENCE * VOL. 277 * 15 AUGUST 1997 919

This content downloaded from 129.137.5.42 on Tue, 26 Jan 2016 18:42:44 UTC All use subject to JSTOR Terms and Conditions

fight broke out in front of their house, and (v) the fire station closest to their home was threatened with budget cuts. "Social cohe- sion and truLst" were also represented by five conceptually related items. Respondents were asked how strongly they agreed (on a five-point scale) that "people around here are willing to help their neighbors," "this is a close-knit neighborhood," "people in this neighborhood can be trusted," "people in this neighborhood generally don't get along with each other," and "people in this neigh- borhood do not share the same values" (the last two statements were reverse coded).

Responses to the five-point Likert scales were aggregated to the neighborhood level as initial measures. Social cohesion and in- formal social control were closely associated across neighborhoods (r = 0.80, P < 0.001), which suggests that the two mea- sures were tapping aspects of the same la- tent construct. Because we also expected that the willingness and intention to inter- vene on behalf of the neighborhood would be enhanced under conditions of mutual trust and cohesion, we combined the two scales into a summary measure labeled col- lective efficacy (21).

The measurement of violence was achieved in three ways. First, respondents were asked how often each of the follow- ing had occurred in the neighborhood dur- ing the past 6 months: (i) a fight in which a weapon was used, (ii) a violent argument between neighbors, (iii) a gang fight, (iv) a sexual assault or rape, and (v) a robbery or mugging. The scale construction for perceived neighborhood violence mir- rored that for social control and cohesion. Second, to assess personal victimization, each respondent was asked "While you have lived in this neighborhood, has any- one ever used violence, suLch as in a mug- ging, fight, or sexual assault, against you or any member of your household anywhere in your neighborhood?" (22). Third, we

Table 2. Oblique rotated factor pattern (Loadings 2 0.60) in 343 Chicago neighborhoods. (Data are from the 1990 census.)

Variable Factor loading

Concentrated disadvantage Below poverty line 0.93 On public assistance 0.94 Female-headed families 0.93 Unemployed 0.86 Less than age 18 0.94 Black 0.60

Immigrant concentration Latino 0.88 Foreign-born 0.70

Residential stability Same house as in 1 985 0.77 Owner-occupied house 0.86

tested both survey measures against inde- pendently recorded incidents of homicide aggregated to the NC level (23). Homi- cide is one of the most reliably measuLred crimes by the police and does not suffer the reporting limitations associated with other violent crimes, such as assault and rape.

Ten variables were constructed from the 1990 decennial census of the popula- tion to reflect neighborhood differences in poverty, race and ethnicity, immigration, the labor market, age composition, family structure, homeownership, and residential stability (see Table 2). The census was independent of the PHDCN CS; more- over, the census data were collected 5 years earlier, which permitted temporal sequencing. To assess whether a smaller number of linear combinations of census characteristics describe the structure of the 343 Chicago neighborhoods, we con- ducted a factor analysis (24).

Consistent with theories and research on U.S. cities, the poverty-related vari- ables given in Table 2 are highly associat- ed and load on the same factor. With an eigenvalue greater than 5, the first factor is dominated by high loadings (>0.85) for poverty, receipt of public assistance, un- employment, female headed-families, and density of children, followed by, to a lesser extent, percentage of black residents. Hence, the predominant interpretation re- volves around concentrated disadvan- tage-African Americans, children, and single-parent families are differentially found in neighborhoods with high con- centrations of poverty (25). To represent this dimension parsimoniously, we calcu- lated a factor regression score that weight- ed each variable by its factor loading.

The second dimension captures areas of the city undergoing immigration, especial- ly from Mexico. The two variables that define this dimension are the percentage of Latinos (approximately 70% of Latinos in Chicago are of Mexican descent) and the percentage of foreign-born persons. Similar to the procedures for concentrated disadvantage, a weighted factor score was created to reflect immigrant concentra- tion. Because it describes neighborhoods of ethnic and linguistic heterogeneity, there is reason to believe that immigrant concentration may impede the capacity of residents to realize common values and to achieve informal social controls, which in turn explains an increased risk of violence (1-5, 7).

The third factor score is dominated by two variables with high (>0.75) loadings: the percentage of persons living in the same house as 5 years earlier and the percentage of owner-occupied homes. The clear emer-

gence of a residential stability factor is con- sistent with much past research (13).

Analytic Models

The internal consistency of a person measure will depend on the intercorrelation among items and the number of items in a scale. The intemal consistency of a neighborhood measure will depend in part on these factors, but it will hinge more on the degree of intersubjective agreement among informants in their ratings of the neighborhood in which they share membership and on the sample size of informants per neighborhood (26). To study reliability, we therefore for- mulated a hierarchical statistical model rep- resenting item variation within persons, per- son variation within neighborhoods, and variation between neighborhoods. Compli- cating the analysis is the problem of missing data: inevitably, some persons will fail to respond to some questions in an interview. We present our hierarchical model as a series of nested models, one for each level in the hierarchy (27).

Level 1 model. Within each person, Yijk' the ith response of person j in neighborhood k, depends on the person's latent perception of collective efficacy plus error:

9

Yijk = wjrp + I otpDpijk+ ejk, (1) p=l

Here DP,., is an indicator variable taking on a value of unity if response i is to item p in the 10-item scale intended to measure col- lective efficacy and zero if response i is to some other item. Thus, aUp represents the "difficulty" of item p, and TFjk iS the "true score" for person jk and is adjusted for the difficulty level of the items to which that person responded (28). The errors of mea- surement, ey11e, are assumed to be indepen- dent and homoscedastic (that is, to have equal standard deviations).

Level 2 model. Across informants within neighborhoods, the latent true scores vary randomly around the neighborhood mean:

Tjk = '1k + 'jk, rjk- N(O, T) (2) Here rk iS the neighborhood mean collec- tive efficacy, and random effects rik associ- ated with each person are independently, normally distributed with variance T,, that is, the "within-neighborhood variance."

Level 3 model. Across neighborhoods, each neighborhood's mean collective effi- cacy -lk varies randomly about a grand mean:

-qk = Y + Uk, Uk N(O,TQ) (3) where y is the grand mean collective effica- cy, Uk iS a normally distributed random effect associated with neighborhood k, and vn, is

920 SCIENCE * VOL. 277 ? 15 AUGUST 1997 ? www.sciencemag.org

This content downloaded from 129.137.5.42 on Tue, 26 Jan 2016 18:42:44 UTC All use subject to JSTOR Terms and Conditions

- RESEARCH ARTICLES the between-neighborhood variance. Ac- cording to this setup, the object of measure- ment is nk. The degree of intersubjective agreement among raters is the intraneigh- borhood correlation, p = /(T, + T'). The reliability of measurement of -lk depends primarily on p and on the sample size per neighborhood. The entire three-level model is estimated simultaneously via maximum likelihood (26).

The results showed that 21% of the vari- ation in perceptions of collective efficacy lies between the 343 neighborhoods (29). The reliability with which neighborhoods can be distinguished on collective efficacy ranges between 0.80 for neighborhoods with a sam- ple size of 20 raters to 0.91 for neighborhoods with a sample size of 50 raters.

Contr-olling r-esponse biases. Suppose, however, that informant responses to the collective efficacy questions vary system- atically within neighborhoods as a func- tion of demographic background (such as age, gender, SES, and ethnicity), as well as homeownership, marital status, and so on. Then variation across neighborhoods in the composition of the sample of respon- dents along these lines could masquerade as variation in collective efficacy. To con- trol for such possible biases, we expanded the level 2 model (Eq. 2) by incorporating 11 characteristics of respondents as covari- ates. Equation 2 becomes

11

1Tr; qk + E 6XpVk + rjk, rjk -

N(0,T.)

(4) where XLPk is the value of covariate q asso- ciated with respondent j in neighborhood k and bq is the partial effect of that covariate on the expected response of that informant on the collective efficacy items. Thus, -lk iS now the level of efficacy for neighborhood k after adjustment for the composition of the informant sample with respect to 11 char- acteristics: gender (1 = female, 0 = male), marital status (composed of separate indica- tors for married, separated or divorced, and single), homeownership, ethnicity and race (composed of indicators for Latinos and blacks), mobility (number of moves in past 5 years), years in neighborhood, age, and a composite measure of SES (the first princi- pal component of education, income, and occupational prestige).

Association Between Neighborhood Social

Composition and Collective Efficacy

The theory described above led us to ex- pect that neighborhood concentrated dis- advantage (con. dis.) and immigrant con-

centration (imm. con.) would be negative- ly linked to neighborhood collective effi- cacy and residential stability would be positively related to collective efficacy, net of the contributions of the 11 covari- ates defined in the previous paragraph. To test this hypothesis, we expanded the level 3 model (Eq. 3) to

'1k = yo + yi(con. dis.)k + Y2(stability)k

+ y3(imm. con.)k

+ Uk, Uk- N(0,rr,) (5)

where yo is the model intercept and y1, y2, and Y3 are partial regression coefficients.

We found some effects of personal background (Table 3): High SES, home- ownership, and age were associated with elevated levels of collective efficacy, whereas high mobility was negatively as- sociated with collective efficacy. Gender, ethnicity, and years in neighborhood were not associated with collective efficacy.

At the neighborhood level, when these personal background effects were con- trolled, concentrated disadvantage and immigrant concentration were signif- icantly negatively associated with col- lective efficacy, whereas residential stabil- ity was significantly positively associated with collective efficacy (for metric co- efficients and t ratios, see Table 3). The standardized regression coefficients were -0.58 for concentrated disadvantage, -0.13 for immigrant concentration, and 0.25 for residential stability, explain- ing over 70% of the variability across the 343 NCs.

Collective Efficacy as a Mediator of Social Composition

Past research has consistently reported links between neighborhood social composition and crime. We assessed the relation of so- cial composition to neighborhood levels of violence, violent victimization, and homi- cide rates, and asked whether collective efficacy partially mediated these relations.

Perceived violence. Using a model that paralleled that for collective efficacy (Eqs. 1, 4, and 5), we found that reports of neigh- borhood violence depended to some degree on personal background. Higher levels of violence were reported by those who were separated or divorced (as compared with those who were single or married), by whites and blacks (as opposed to Latinos), by younger respondents, and by those with longer tenure in their current neighbor- hood. Gender, homeownership, mobility, and SES were not significantly associated with responses within neighborhoods. When these personal background charac- teristics were controlled, the concentrations of disadvantage (t = 13.30) and immigrants (t = 2.44) were positively associated with the level of violence (see Table 4, model 1). The corresponding standardized regression coefficients are 0.75 and 0.11. Also, as hy- pothesized, residential stability was nega- tively associated with the level of violence (t = -6.95), corresponding to a standard- ized regression coefficient of -0.28. The model accounted for 70.5% of the variation in violence between neighborhoods.

Next, collective efficacy was added as a predictor in the level 3 model (Table 4,

Table 3. Correlates of collective efficacy.

Variable Coefficient SE t ratio

Intercept 3.523 0.013 263.20 Person-level predictors

Female -0.012 0.015 -0.76 Married -0.005 0.021 -0.25 Separated or divorced -0.045 0.026 -1.72 Single -0.026 0.024 -1.05 Homeowner 0.122 0.020 6.04 Latino 0.042 0.028 1.52 Black -0.029 0.030 -0.98 Mobility -0.025 0.007 -3.71 Age 2.09 x 10-3 0.60 X 10-3 3.47 Years in neighborhood 0.64 x 10-3 0.82 x 10-3 0.78 SES 3.53 x 10-2 0.76 x 10-2 4.64

Neighborhood-level predictors Concentrated disadvantage -0.172 0.016 -10.74 Immigrant concentration -0.037 0.014 -2.66 Residential stability 0.074 0.130 5.61

Variance components Within neighborhoods 0.320 Between neighborhoods 0.026

Percent of variance explained Within neighborhoods 3.2 Between neighborhoods 70.3

www.sciencemag.org ? SCIENCE @ VOL. 277 * 15 AUGUST 1997 921

This content downloaded from 129.137.5.42 on Tue, 26 Jan 2016 18:42:44 UTC All use subject to JSTOR Terms and Conditions

model 2). The analysis built in a correction for errors of measurement in this predictor (30). We found collective efficacy to be negatively related to violence (t = -5.95), net of all other effects, and to correspond to a standardized coefficient of -0.45. Hence, after social composition was controlled, col- lective efficacy was strongly negatively as- sociated with violence. Moreover, the coef- ficients for social composition were substan- tially smaller than they had been without a control for collective efficacy. The coeffi- cient for concentrated disadvantage, al- though still statistically significant, was 0.171 (as compared with 0.277). The differ- ence between these coefficients (0.277 - 0.171 = 0.106) was significant (t = 5.30). Similarly, the coefficients for immigrant concentration and for residential stability were also significantly reduced: The coeffi- cient for immigrant concentration, original- ly 0.041, was now 0.018, a difference of 0.023 (t = 2.42); the coefficient for resi- dential stability, which had been -0.102, was now -0.056, a difference of -0.046 (t = -4.18). The immigrant concentration coefficient was no longer statistically differ- ent from zero. As hypothesized, then, col- lective efficacy appeared to partially medi- ate widely cited relations between neigh- borhood social composition and violence. The model accounted for more than 75% of the variation between neighborhoods in levels of violence.

Violent victimization. Violent victimiza- tion was assessed by a single binary item (Yjk

1 if victimized by violence in the neigh- borhood and Y-k = 0 if not). The latent outcome was the logarithmic odds of vic- timization lTjk. The structural model for pre- dicting

'Ajk had the same form as before

(Eqs. 4 and 5) (31). Social composition, as hypothesized, predicted criminal victimiza- tion, with positive coefficients for concen- trated disadvantage and immigrant concen- tration and a negative coefficient for resi- dential stability (Table 4, model 1). The relative odds of victimization associated with a 2-SD elevation in the predictor were 1.67, 1.33, and 0.750, respectively. These estimates controlled for background charac- teristics associated with the risk of victim- ization. When added to the model, collec- tive efficacy was negatively associated with victimization (Table 4, model 2). A 2-SD elevation in collective efficacy was associ- ated with a relative odds ratio of about 0.70, which indicated a reduction of 30% in the odds of victimization. Moreover, after col- lective efficacy was controlled, the coeffi- cients associated with concentrated disad- vantage and residential stability diminished to nonsignificance, and the coefficient for immigrant concentration was also reduced.

Homicide. To assess the sensitivity of the

findings when the measure of crime was completely independent of the survey, we examined 1995 homicide counts (Yk is the number of homicides in neighborhood k in 1995). A natural model for the expected number of homicides in neighborhood k is E(Yk) = NkXk, where Xk is the homicide rate per 100,000 people in neighborhood k and Nk is the population size of neighborhood k as given by the 1990 census (in hundreds of thousands). Defining nk = log (Xk), we then formulated a regression model for -k of the type in Eq. 5. This is effectively a Poisson regression model with a logarithmic link with extra-Poisson variation represented by between-neighborhood random effects (32).

Although concentrated disadvantage was strongly positively related to homicide, immigrant concentration was unrelated to homicide, and residential stability was weakly positively related to homicide (Ta- ble 4, model 1). However, when social com- position was controlled, collective efficacy was negatively related to homicide (Table 4, model 2). A 2-SD elevation in collective efficacy was associated with a 39.7% reduc- tion in the expected homicide rate. More- over, when collective efficacy was con- trolled, the coefficient for concentrated dis- advantage was substantially diminished, which indicates that collective efficacy can be viewed as partially mediating the associ- ation between concentrated disadvantage and homicide (33).

Control for prior homicide. Results so far were mainly cross-sectional, which raised the question of the possible confounding

effect of prior crime. For example, residents in neighborhoods with high levels of vio- lence might be afraid to engage in acts of social control (9). We therefore reestimated all models controlling for prior homicide: the 3-year average homicide rate in 1988, 1989, and 1990. Prior homicide was nega- tively related (P < 0.01) to collective effi- cacy in 1995 (r = -0.55) and positively related (P < 0.01) to all three measures of violence in 1995, including a direct associ- ation (t = 5.64) with homicide (Table 5). However, even after prior homicide was controlled, the coefficient for collective ef- ficacy remained statistically significant and substantially negative in all three models.

Further Tests

Although the results have been consistent, there are still potential threats to the valid- ity of our analysis. One question pertains to discriminant validity: how do we know that it is collective efficacy at work rather than some other correlated social process (34)? To assess competing and analytically dis- tinct factors suggested by prior theory (4, 5), we examined the measure of collective efficacy alongside three other scales derived from the CS of the PHDCN: neighborhood services, friendship and kinship ties, and organizational participation (35). On the basis of the results in Tables 3 to 5 and also to achieve parsimony, we constructed a vi- olent crime scale at the neighborhood level that summed standardized indicators of the three major outcomes: perceived violence,

Table 4. Neighborhood correlates of perceived neighborhood violence, violent victimization, and 1995 homicide events.

Model . soiModel 2: social composition Model 1: social compositon and collective efficacy V ariable__ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

Coefficient SE t Coefficient SE t

Perceived neighborhood violence* Concentrated disadvantage 0.277 0.021 13.30 0.171 0.024 7.24 Immigrant concentration 0.041 0.017 2.44 0.018 0.016 1.12 Residential stability -0.102 0.015 -6.95 -0.056 0.016 -3.49 Collective efficacy -0.618 0.104 -5.95

Violent victimizationt Concentrated disadvantage 0.258 0.045 5.71 0.085 0.054 1.58 Immigrant concentration 0.141 0.046 3.06 0.098 0.044 2.20 Residential stability -0.143 0.050 -2.84 -0.031 0.051 -0.60 Collective efficacy -1.190 0.240 -4.96

1995 homicide eventst Concentrated disadvantage 0.727 0.049 14.91 0.491 0.064 7.65 Immigrant concentration -0.022 0.051 -0.43 -0.073 0.050 -1.45 Residential stability 0.093 0.042 2.18 0.208 0.046 4.52 Collective efficacy -1.471 0.261 -5.64

*Estimates of neighborhood-level coefficients control for gender, marital status, homeownership, ethnicity, mobility, age, years in neighborhood, and SES of those interviewed. Model 1 accounts for 70.5% of the variation between neighborhoods in perceived violence, whereas model 2 accounts for 77.8% of the variation. tNeighborhood-level coefficients are adjusted for the same person-level covariates listed in the first footnote. Model 1 accounts for 12.3% of the variation between neighborhoods in violent victimization, whereas model 2 accounts for 44.4%. tModel 1 accounts for 56.1% of the variation between neighborhoods in homicide rates, whereas model 2 accounts for 61.7% of the variation.

922 SCIENCE * VOL. 277 * 15 AUGUST 1997 * www.sciencemag.org

This content downloaded from 129.137.5.42 on Tue, 26 Jan 2016 18:42:44 UTC All use subject to JSTOR Terms and Conditions

~~ RESEARCH ARTICLES

violent victimization, and homicide rate. Consistent with expectations, collective

efficacy was significantly (p < 0.01) and positively related to friendship and kinship ties (r = 0.49), organizational participation (r = 0.45), and neighborhood services (r = 0.21). Nonetheless, when we controlled for these correlated factors in a multivariate regression, along with prior homicide, con- centrated disadvantage, immigrant concen- tration, and residential stability, by far the largest predictor of the violent crime rate was collective efficacy (standardized coeffi- cient = -0.53, t = -8.59). Collective efficacy thus retained discriminant validity when compared with theoretically relevant, competing social processes. Moreover, these results suggested that dense personal ties, organizations, and local services by them- selves are not sufficient; reductions in vio- lence appear to be more directly attribut- able to informal social control and cohesion among residents (36).

A second threat stems from the associa- tion of racial composition with concentrat- ed disadvantage as shown in Table 2. Our interpretation was that African Americans, largely because of housing discrimination, are differentially exposed to neighborhood conditions of extreme poverty (15). None- theless, a counterhypothesis is that the per- centage of black residents and not disad- vantage accounts for lower levels of collec- tive efficacy and, consequently, higher vio- lence. Our second set of tests therefore replicated the key models within the 125 NCs where the population was more than 75% black (see the first row of Table 1), effectively removing race as a potential confound. Concentrated poverty and resi-

dential stability each had significant associ- ations with collective efficacy in these pre- dominantly black areas (t = -5.60 and t = 2.50, respectively). Collective efficacy con- tinued to explain variations in violence across black NCs, mediating the prior effect of concentrated disadvantage. Even when prior homicide, neighborhood services, friendship and kinship ties, and organiza- tional participation were controlled, the only significant predictor of the violent crime scale in black NCs was collective efficacy (t = -4.80). These tests suggested that concentrated disadvantage more than race per se is the driving structural force at play.

Discussion and Implications

The results imply that collective efficacy is an important construct that can be mea- sured reliably at the neighborhood level by means of survey research strategies. In the past, sample surveys have primarily consid- ered individual-level relations. However, surveys that merge a cluster sample design with questions tapping collective properties lend themselves to the additional consider- ation of neighborhood phenomena.

Together, three dimensions of neighbor- hood stratification-concentrated disad- vantage, immigration concentration, and residential stability-explained 70% of the neighborhood variation in collective effica- cy. Collective efficacy in turn mediated a substantial portion of the association of res- idential stability and disadvantage with multiple measures of violence, which is consistent with a major theme in neighbor- hood theories of social organization (1-5).

Table 5. Predictors of neighborhood level violence, victimization, and homicide in 1995, with prior homicide controlled. For violence and victimization as outcomes, the coefficients reported in this table were adjusted for 11 person-level covariates (see Table 3), but the latter coefficients are omitted for simplicity of presentation.

Homicide in 1995 as Violence as outcome Victimization as outcome outcome

Variable

Coefficient SE t Coefficient SE t Coefficient SE t

Intercept 3.772 0.379 9.95 -2.015 0.042 -49.24 3.071 0.050 62.01 Concentrated 0.157 0.025 6.38 0.073 0.060 1.22 0.175 0.072 2.42

disadvantage Immigrant 0.020 0.016 1.25 0.098 0.045 2.20 -0.034 0.044 -0.77

concentration Residential stability -0.054 0.016. -3.39 -0.029 0.052 -0.56 0.229 0.043 5.38 Collective efficacy -0.594 0.108 -5.53 -1.176 0.251 -4.69 -1.107 0.272 -4.07 Prior homicide 0.018 0.014 1.27 0.017 0.049 0.34 0.397 0.070 5.64

Variance Between-

neighborhood variance 0.030 0.091 0.207

Percent of variance explained between neighborhoods 78.0 43.8 73.0

After adjustment for measurement error, individual differences in neighborhood composition, prior violence, and other po- tentially confounding social processes, the combined measure of informal social con- trol and cohesion and trust remained a ro- bust predictor of lower rates of violence.

There are, however, several limitations of the present study. Despite the use of decennial census data and prior crime as lagged predictors, the basic analysis was cross-sectional in design; causal effects were not proven. Indicators of informal control and social cohesion were not observed di- rectly but rather inferred from informant reports. Beyond the scope of the present study, other dimensions of neighborhood efficacy (such as political ties) may be im- portant, too. Our analysis was limited also to one city and did not go beyond its official boundaries into a wider region.

Finally, the image of local residents working collectively to solve their own problems is not the whole picture. As shown, what happens within neighbor- hoods is in part shaped by socioeconomic and housing factors linked to the wider political economy. In addition to encourag- ing communities to mobilize against vio- lence through "self-help" strategies of infor- mal social control, perhaps reinforced by partnerships with agencies of formal social control (community policing), strategies to address the social and ecological changes that beset many inner-city communities need to be considered. Recognizing that collective efficacy matters does not imply that inequalities at the neighborhood level can be neglected.

REFERENCES AND NOTES

1 For a recent review of research on violence covering much of the 20th century, including a discussion of the many barriers to direct examination of the mech- anisms explaining neighborhood-level variations, see R. J. Sampson and J. Lauritsen, in. Understanding and Preventing Violence: Social Influences, vol. 3, A. J. Reiss Jr. and J. Roth, Eds. (National Academy Press, Washington, DC, 1994), pp. 1-114.

2. J. F. Short Jr., Poverty, Ethnicity, and Violent Crime (Westview, Boulder, CO, 1997).

3. For a general assessment of the difficulties facing neighborhood-level research or! social outcomes, see S. E. Mayer and C. Jencks, Science 243, 1441 (1989).

4. R. Kornhauser, Social Sources of Delinquency (Univ. of Chicago Press, Chicago, IL, 1978); R. J. Bursik Jr., Criminology 26, 519 (1988); D. Elliott et al., J. Res. Crime Delinquency 33, 389 (1996).

5. R. J. Sampson and W. B. Groves, Am. J. Sociol. 94, 774 (1989).

6. M. Janowitz, ibid. 81, 82 (1975). 7. E. Maccoby, J. Johnson, R. Church, J. Social Issues

14, 38 (1958); R. Taylor, S. Gottfredson, S. Brower, J. Res. Crine Delinquency 21, 303 (1983); J. Hacker, K. Ho, C. Ross, Social Problems 21, 328 (1974). A key finding from past research is that many delinquent gangs emerge from unsupervised spontaneous peer groups [F. Thrasher, The Gang: A Study of 1,313 Gangs in Chicago (Univ. of Chicago Press, Chicago, IL, 1963); C. Shaw and H. McKay, Juvenile Delin-

www.sciencemag.org * SCIENCE * VOL. 277 * 15 AUGUST 1997 923

This content downloaded from 129.137.5.42 on Tue, 26 Jan 2016 18:42:44 UTC All use subject to JSTOR Terms and Conditions

quen7cy and Urbani Areas (Univ. of Chicago Press, Chicago, IL. 1969), pp. 176-185; J. F. Short Jr. and F. Strodtbeck, Group Process anid Gang Deliniqueni- cy (Univ. of Chicago Press, Chicago, IL, 1965)].

8. For example, aboLIt half of all homicides OCCur among nonifamilily members with a preexistinig rela- tionship: friends, neighibors, casual acquaintances, associates in illegal activities, or members of a rival gang. Illegal markets are especially high-r isk settings for robbery, assault, and homicide victimizationi, whether by an associate or a stranger [A. J. Reiss Jr. and J. Roth, Eds. Uniderstanclinig anid Preventing Vi- olence (National Academy Press, Washingtoni, DC, 1993), pp. 18. 79; A. J. Reiss Jr., in Criininal Careers and "Career Crirnihals," A. Blumstein. J. Cohen! J. Roth, C. Visher, Eds. (National Academy Press, Washington, DC, 1986), pp. 121-160].

9. W. Skogan, Disorder and Decline: Crime and the Spiral of Decay in American Neighborlhoods (U n iv. of California Press, Berkeley, CA, 1990).

10. J. Coleman, Foundclations of Social Thieoiy (Harvard Univ. Press, Cambridge, MA, 1990); R. Putnanm, Makin;g Democracy Work (Princeton Univ. Press, Princetoni, NJ, 1993).

11. A. Bandura, Social Founclations of ThoLight and Ac- tion: A Social Cogniitive Theol-y (Prentice-Hall, Engle- wood Cliffs, NJ, 1986).

12. See, generally, J. Logan and H. Molotch, Urban For- tLlnes: The Political Economy of Place (Univ. of Cali- fornia Press, Berkeley, CA, 1987).

13. See also J. Kasarda and M. Janowitz, Am. Sociol. Rev. 39, 328 (1974); R. Sampson, ilbid. 53, 766 (1988).

14. W. J. Wilson, The TrLIly Disadvantaged (Univ. of Chi- cago Press, Chicago, IL, 1987).

15. D. Massey and N. Denton, Americani Apartheid: Segregationi alnd the Makilg of the Ulnderclass (Har- vard Univ. Press, Cambridge. MA, 1993); D. Massey, Am. J. Sociol. 96, 329 (1990).

16. J. Brooks-Gunn, G. Duncan, P. Kato, N. Sealand. Am. J. Sociol. 99, 353 (1993); F. F. Furstenberg Jr., T. D. Cook, J. Eccles, G. H. Elder, A. Sameroff, Urban Families and Adolescent SuLccess (Univ. of Chicago Press, Chicago, IL, in press), chap. 7. Re- search has shown a strong link between the concen- tration of female-headed families and rates of vio- lence [see (1)].

17. D. Williams and C. Collins, AnluLi. Rev. Sociol. 21, 349 (1995).

18. Cluster- analyses of census data also helped to guide the construction of internally homogeneous NCs with respect to racial and ethnic mix, SES, housing density, andc family organization. Random1-effect analyses of variance produced intracluster correla- tion coefficients to assess the degree to which this goal had been achie\ved; analyses (37) revealed that the clustering was successful in producing relative homogeneity within NCs.

19. For purposes of selecting a longitudinal cohort sam- ple, SES was defined with the use of a scale from the 1990 census that included NC-level indicators of poverty, public assistance, inconme, and education (37). Race and ethnicity were also measured with the use of the 1990 census, which defined race in five broad categories: "white," "black," "American tndian, Eskimo, or Aleut," "Asian or Pacific Islander," and "other." We use the census labels of white and black to refer to persons of European American and Afri- can American background, respectively. We use the term "Latino" to denote anyone of Latin American descent as determined from the separate census category of "Hispanic origin." "Hispanic is more properly used to describe persons of Spanish de- scent (i.e., from Spain), although the terms are com- monly used interchangeably.

20. The sampling design of the CS was complex. For purposes of a longitudinal study (37), residents in 80 of the 343 NCs were oversampled. Within these 80 NCs, a simple random sample of census blocks was selected and a systematic random sample of dwell- ing units within those blocks was selected. Within each dwelling unit, all persons over 18 were listed, and a respondent was sampled at random with the aim of obtaining a sample of 50 households within each NC. In each of the remaining NCs (n = 263),

ninle census blocks were selected with probability proportionial to population size, thl-ee dwelling unlits were selected at random-n within each block, and an adult respondent was random-nly selected from a list of all adults in the dwelling unlit. The aim was to obtain a sample of 20 in these 263 NCs. Despite these differences in sampling design, the selected dwelling units constituted a representative and approximately self-weighting sample of dwelling units within every NC (17 = 343). ABT Associates (Cambridge, MA) carried out the data collection with the cooperation of research staff at PHDCN, achievinig a final re- sponse rate of 75%.

21. 'Don't know" responses were recoded to the middle category of "neither likely nor unlikely" (informal social control) or "neither agree nor disagree" (social cohe- sion). Most respondents answered all 10 items in- cluded in the combined measure; for those respon- dents, the scale score was the average of the re- sponses. However, anyone responding to at least one item provided data for the analysis; a person- specific standard error of measurement was calcu- lated on the basis of a simple linear item-response model that took into accounlt the number and diffi- culty of the items to which each resident responded. The analyses reported here were based oni the 7729 cases having sufficient data for all models estimnated.

22. Respondents were also asked whether the incident occurred during the 6 months before the interview; about 40% replied affirmatively. Because violence is a rare outcome, we use the total violent victimization measure in the main analysis. However, in additional analyses, we examnined a summary of the prevalence of personal and household victimizations (ranginig from 0 to foulr) restricted to this 6-monith window. This test yielded results very similar to those based on the binary mneasure of total violence.

23. The original data measured the address location of all homicide incidents known to the Chicago police (regardless of arrests) during the months of the com- munity survey.

24. The alpha-scoring method was chosen because we are analyzing the universe of NCs in Chicago and are interested in maximizing the reliability of measures [H. F. Kaiser and J. Caffry, Psychornetr-ika 30, 1 (1965)]. We also estimated an oblique factor rotation, allowing the extracted dimensions to covary. A prin- cipal components analysis with varirnax rotation nonetheless yielded substantively identical results.

25. For a mnethodological procedul-e and empirical result that are similar but that used all U.S. cities as units of analysis, see K. Land, P. McCall, L. Cohen, Am. J. Sociol. 95, 922 (1990).

26. S. W. Raudenbush, B. Rowvan, S. J. Kang, J. ECdLIc. Stat. 16, 295 (1991).

27. D. V. Lindley and A. F. M. Smitl, R. Stat. Soc. J. Ser. B Methodol. 34, 1 (1972).

28. Although the vast majority of respondents answered all items in the collective efficacy scale, the measure- ment model makes full use of the data provided by those whose responses were incomplete. There is one less indicator, Dpj, than the nuimber of items to identify the intercept Th'(

29. This degree of intersubjective agreement is similar to that found in a recent national survey of teachers that assessed organizational climate in U.S. high schools [B. Rowan, S. Raudenbush, S. Kang, Am. J. EdLIc. 99, 238 (1991)].

30. The analysis of collective efficacy and violence as Out- comes uses a three-level model in which the level 1 model describes the sources of measurement error for each of these outcomes. The level 2 and level 3 models together describe the joint distribution of the "true scores" within and between neighborhoods. Given the joint distribution of these ouItcomes, it is then possible to describe the conditional distribution of violence given "true' collective efficacy and all other predictors, thus automatically adjusting for any errors of measurement of collective efficacy. See S. Raucien- bush and R. J. Sampson (paper presented at the conference "Alternative Models for Educational Data," National Institute of Statistical Sciences Research Tri- angte Park, NC, 16 October 1996) for the necessary derivations. This work is an extension of that of C. Ctogg, E. Petk<ova, and A. Haritou [Am. J. Sociof. 100,

1261 (1995)] and P. Allison (ibid., p. 1294). Note that census blocks were not included as a "level" in the analysis. Thus, person-level and block-level variance are confounided. However, this confounding has no effect on standard errors reported in this manuscript. If explanatory variables had been measured at the level of the census block, it would have been impor- tant to represent blocks as an additional level in the model.

31. The resulting model is a logistic regression model with random effects of neighborhoods. This model was estimated first with penalized quasi-likelihood as described by N. E. Breslow and D. G. Clayton [J. Am. Stat. Assoc. 88, 9 (1993)]. The doubly iterative algo- rithm used is described by S. W. Raudenbush ["Pos- terior modal estimation for hierarchical generalized linear models with applications to dichotomous and couint data" (Longitudinal and Multilevel Methods Project, Michigan State Univ., East Lansing, Ml, 1993)]. Then, using those results to model the mar- ginal covariation of the errors, we estimated a pop- uilation-average model with robust standard errors [S. Zeger, K. Liang, P. Albert, Biometrics 44, 1049 (1988)]. Results were similar. The results based on the population-average model with robust standard errors are reported here.

32. The analysis paralleled that of criminal victimization, except that a Poisson sampling model and logarith- milic link were used in this case. Again, the reported results are based on a population-average model with robust standard errors.

33. Although the zero-order correlation of residential sta- bility with homicide was insiginificant, the partial co- efficient in Table 4 is significantly positive. Recall from Table 3 that stability is positively linked to collective efficacy. But higher stability without the expected greater collective efficacy is not a positive neighbor- hood quality according to the homicide data. See (14).

34. T. Cook, S. Shagle, S. Degirmencioglu, in Neighbor- hood Poverty: Context and Conisequenices for Chil- dren, vol. 2, J. Brooks-Gunn, G. Duncan, J. L. Aber, Eds. (Russell Sage Foundation, New York, in press).

35. "Neighborhood seivices" is a nine-item scale of local activities and programs (for example, the presence of a block group, a tenant association, a crime pre- vention program, and a family health seivice) com- bined with a six-item inventory of seivices for youth (a neighborlhood youth center, recreational programs, after-school programs, mentoring aidcounselingser- vices, mental health seivices, and a crisis interven- tion program). "Friendship and kinship ties" is a scale that measures the numliber of friends and relatives that respondents report are living in the neighbor- hood. "Organizational participation" measures actu- al involvement by residents in (i) local religioLis orga- nizations; (ii) neighborhood watch programs; (iii) block groLup, tenant association, or community council; (iv) business or civic groups; (v) ethnic or nationality clubs; and (vi) local political organizations.

36. Similar results were obtained when we controlled for a measure of social interaction (the extent to which neighbors had parties together, watched each oth- er's homes, visited in each others' homes, ex- changed favors, and asked advice about personal matters) that was positively associated with collec- tive efficacy. Again the direct effect of collective effi- cacy remained, suggesting that social interaction, like friendship and kinship ties, is linked to reduced violence through its association with increased levels of collective efficacy.

37. R. J. Sampson, S. W. Raudenbush, F. Earls, data not shown.

38. Major funding for this project came from the John D. and Catherine T. MacArthur Foundation and the Na- tional Institute of Justice. We thank L. Eisenberg and anonymous reviewers for helpful comments; S. Buka and A. J. Reiss Jr. for important contributions to the research design; and R. Block, C. Coldren, and J. Morenoff for their assistance in obtaining, cleaning, geo-coding, and aggregating homicide incident data to the NC level. M. Yosef and D. Jeglum-Bartusch assisted in the analysis.

16 January 1997; accepted 20 June 1997

924 SCIENCE ? VOL. 277 ? 15 AUGUST 1997 o www.sciencemag.org

This content downloaded from 129.137.5.42 on Tue, 26 Jan 2016 18:42:44 UTC All use subject to JSTOR Terms and Conditions