Collective Efficacy and Violence in Chicago Neighborhoods: A Reproduction

Published date01 August 2018
Date01 August 2018
DOI10.1177/1043986218769988
Subject MatterArticles
/tmp/tmp-17IcmO1CKV7hNO/input 769988CCJXXX10.1177/1043986218769988Journal of Contemporary Criminal JusticeMaxwell et al.
research-article2018
Article
Journal of Contemporary Criminal Justice
2018, Vol. 34(3) 245 –265
Collective Efficacy and
© The Author(s) 2018
Reprints and permissions:
Violence in Chicago
sagepub.com/journalsPermissions.nav
https://doi.org/10.1177/1043986218769988
DOI: 10.1177/1043986218769988
journals.sagepub.com/home/ccj
Neighborhoods:
A Reproduction
Christopher D. Maxwell1, Joel H. Garner2,
and Wesley G. Skogan3
Abstract
This research tests the reproducibility of the neighborhood-level effects of social
composition and collective efficacy on community violence that Sampson, Raudenbush,
and Earls reported in a Science article entitled “Neighborhood and Violent Crime: A
Multilevel Study of Collective Efficacy.” Based upon data from a resident survey, the
U.S. Census, and official homicide reports from Chicago, Sampson et al. found that
neighborhood collective efficacy directly affects perceived neighborhood violence,
household victimization, and homicide rates. In addition, they reported that the
relationship between residential stability and concentrated disadvantage with each
measure of violence is mediated after adding their collective efficacy measure to the
regression models. This article uses Earls, Brooks-Gunn, Raudenbush, and Sampson’s
archived data collection and other archived data collections to assess the extent
to which Sampson et al.’s core substantive findings are independently reproducible.
While the reanalysis identified some differences between the archived data and the
information provided in Sampson et al., the reanalysis produced findings in the same
reported direction and statistical significance for virtually all of Sampson et al.’s core
substantive outcomes. This confirmation of their key conclusions provides added
confidence in their collective efficacy thesis and enhances the prospects for extending
it by assessing the degree to which it also affects other crime types and whether these
effects persist over time.
1Michigan State University, East Lansing, MI, USA
2Portland State University, OR, USA
3Northwestern University, Evanston, IL, USA
Corresponding Author:
Christopher D. Maxwell, School of Criminal Justice, Michigan State University, 655 Auditorium Road,
Room 438, East Lansing, MI 48854, USA.
Email: cmaxwell@msu.edu

246
Journal of Contemporary Criminal Justice 34(3)
Keywords
violence, collective efficacy, neighborhood
For many years, scientists have speculated about and tested the extent to which varia-
tions in the social composition of neighborhoods lead to different crime outcomes
(Kornhauser, 1978; Shaw & McKay, 1942; Skogan, 1990). The most current thesis in
this area of criminology is that collective efficacy is a key social process by which
cohesion among residents coupled with their willingness to intervene influences the
quantity of criminal behavior (Sampson, 2012). While a correlated concept, self-effi-
cacy, has been used for some time (Bandura, 1986), interest in a collective efficacy
process surged following the publication of Sampson, Raudenbush, and Earls (1997)‘s
“Neighborhood and Violent Crime: A Multilevel Study of Collective Efficacy” article.
This article reports that community violence is less frequent and that the impact of
neighborhood structural disadvantage in Chicago is diminished by neighbors’ willing-
ness to intervene when their shared expectations are violated or at risk. Since their
article’s publication, other empirical evidence has attested to the relevance of collec-
tive efficacy in explaining the variation in violence across Chicago’s neighborhoods
and elsewhere.
In terms of the size of the correlation across macro-level predictors of crime, Pratt
and Cullen (2005) ranked the “collective efficacy” concept fourth largest among 31
established meta-correlates. With a mean effect size of –0.30 produced by 13 studies,
Pratt and Cullen (2005) reported that its adjusted mean effect size was just smaller
than the incarceration correlation coefficient (–0.33) and just larger than the racial
heterogeneity effect (0.29). Besides explaining variations in crime rates, the degree of
collective efficacy also has implications for crime control policy. Both community
policing and crime prevention programs often operate with an implicit assumption
about how formal government programs depend upon community mobilization and
organization for the purposes of crime control (Cancino, 2005; Serewicz, 2009; Wells,
Schafer, Varano, & Bynum, 2006). Moreover, other scientists have linked collective
efficacy to other outcomes like overall well-being and educational outcomes (Cohen,
Finch, Bower, & Sastry, 2005; Kirk, 2009; Maimon & Browning, 2010; Sampson,
2003; Sampson, Morenoff, & Earls, 1999; Simons, Simons, Burt, Brody, & Cutrona,
2005; Vega, Ang, Rodriguez, & Finch, 2011), as well as to intermediate processes such
as motivation to start projects and persistence once engaged (Bandura, 2000).
Sampson et al.’s Collective Efficacy Measure
Sampson et al.’s (1997) innovative conceptualization of collective efficacy empha-
sizes links between cohesion, social trust, shared expectations, and the willingness of
neighborhood residents to act in support of these values to address a task such as
neighborhood safety. Sampson (2004) argues that the “key casual mechanism in col-
lective efficacy theory is social control enacted under conditions of social trust” (p.
108). Triplett (2007) contends that this hypothesis remains grounded in works such as

Maxwell et al.
247
Shaw and McKay (1942)’s community sources of delinquency theories because it con-
nects factors such as residential stability and ethnic diversity with a neighborhood’s
capability to address criminal behavior. Thus, Sampson et al.’s (1997) collective effi-
cacy concept remains connected with other contemporary theories about how com-
munity structures and organizations are linked to crime rates (e.g., Bursik & Grasmick,
1993; Elliott et al., 1996; Skogan, 1990).
Sampson et al. (1997) also articulated an innovative approach to measuring and
then testing the direct and indirect effects of collective efficacy on criminal behavior.
Based on their 1995 survey of Chicago residents, Sampson et al. (1997, p. 920) used
individual-level responses to ten questions to construct two scales that they labeled
“informal social control” and “social cohesion.” They then combined these two scales
at the neighborhood level using an item response model to produce a single measure
that they labeled “collective efficacy.” This measure captured what they argued was
the degree of “linkage” between a neighborhood’s “mutual trust” and its “willingness
to intervene for the common good.” In more concrete language, Sampson et al. (1997)
asserted that the “collective efficacy of residents is a critical means by which urban
neighborhoods inhibit the occurrence of interpersonal violence” (p. 919).
From the same community survey, Sampson et al. (1997, p. 921) also created two
measures of community violence. The first measure of violence was a scale based on
five questions about the respondent’s perceptions of violence in their neighborhood in
the past 6 months. The second measure was based on one question about whether the
respondent or anyone in their family had experienced violence in their current neigh-
borhood. Sampson et al. (1997) also used a third measure of violence derived from
Chicago Police Department’s homicide records for parts of 1995. In addition, Sampson
et al. (1997, p. 920) used 1990 U.S. Census data to construct measures of concentrated
disadvantage, immigrant concentration, and residential stability to capture the relevant
social compositions of Chicago’s neighborhoods.
To test their two main hypotheses, Sampson et al. (1997) formulated three statisti-
cal models. To model the dependent measure of perceived violence, they specified a
three-level hierarchical linear model (HLM) with the individual questions at the first
level, the 11 individual respondent demographics characteristics at the second level,
and the neighborhood measures of collective efficacy, concentrated disadvantage,
immigrant concentration, and residential stability at the third level. Their modeling
approach for the measure of experienced violence was to specify a two-level HLM
with the one victimization question and the individual demographic characteristics at
the first level and the four neighborhood factors specified at the second level. Their
approach to analyzing the homicide data was to specify in HLM a Poisson regression
model with overdispersion and with the logged homicide rate and the four neighbor-
hood factors at the same level (p. 922).
Sampson et al.’s Key Collective Efficacy Results
Sampson et al. (1997) first reported the impact of their three neighborhood-level social
composition factors on each of their three measures of violence. Based upon their

248
Journal of Contemporary Criminal Justice 34(3)
initial analysis, they report that neighborhoods with more concentrated disadvantage
have significantly higher levels of all three forms of violence. Similarly, they report
that higher levels of immigrant concentration are associated with statistically signifi-
cant increases in both perceived and experienced violence; however, they did not find
that immigrant concentration influenced homicide rate. They also reported that neigh-
borhoods with higher...

To continue reading

Request your trial

VLEX uses login cookies to provide you with a better browsing experience. If you click on 'Accept' or continue browsing this site we consider that you accept our cookie policy. ACCEPT