Collective Efficacy,Group Threat,and Urban Change: Examining Informal Social Control Forces in Areas of Gentrification

Abstract

Since the term gentrification was first coined in the 1960s, scholars have had an interest in understanding how this process of change can impact neighborhoods. Empirical research focusing on the relationship between gentrification and crime has yielded varying results, with little examination of the contextual mechanisms that may influence the relationship. This research combines data from the Project on Human Development in Chicago Neighborhoods with several other sources, and employs multilevel modeling techniques to examine two such contextual mechanisms – collective efficacy and group threat, conceptualized as perceived neighborhood change. The results offer preliminary support for the moderating roles of collective efficacy and perceived neighborhood change mechanisms on the relationship between gentrification and crime. While there is an overall negative association between gentrification and crime, this effect is strengthened with collective efficacy, but reversed with rising levels of perceived neighborhood change.

Keywords

collective efficacy crime gentrification group threat informal social control

Introduction

Gentrification is a process that has the potential to impact the appearance, population characteristics, commercial business, and culture of a neighborhood. It involves capital investment, local politics, and is often accompanied by citizens voicing both their support and protest of the changes. This research aims to investigate whether gentrification processes are associated with levels of crime in neighborhoods, examining this phenomenon through the lens of neighborhood social control and perceived neighborhood change.

Concern over neighborhood conditions and their impact on crime is not a new phenomenon (Bursik and Grasmick, 1993; Morenoff et al., 2001; Sampson, 2012; Sampson and Groves, 1989; Shaw and McKay, 1942). However, research on neighborhood-level explanations of crime typically operates under the assumption that neighborhoods with high crime do not see much change in their demographic or economic conditions over time (Sampson, 2012). The causal mechanisms in models of collective efficacy and neighborhood control are not the structural characteristics of neighborhoods, but all neighborhood-level models specify that it is impossible to separate mechanisms such as collective efficacy, private, parochial, or public social control from the neighborhood conditions that influence their development (Sampson, 2012). Therefore, while scholars are developing a growing understanding of how mechanisms such as collective efficacy operate in historically disadvantaged, heterogeneous communities, it is unclear how these mechanisms will operate in such communities as they are undergoing change. Gentrification, a phenomenon that has impacted many impoverished urban areas in the past several decades, will be the specific type of neighborhood change examined in this research project.

Prior research on gentrification and crime has examined possible associations between these phenomena, and at times both positive and negative associations have been found (Papachristos et al., 2011; Van Wilsem et al., 2006). The current research extends this research by addressing the possibility that the effect of gentrification on crime may vary by neighborhood context. Specifically, this research will examine the following:

RQ: How do informal social control and perceived neighborhood change mechanisms interact with the process of gentrification?

(a) Is the relationship between gentrification and crime moderated by social control mechanisms such as collective efficacy?

(b) Does the relationship between gentrification and crime vary by the degree to which the changes are perceived as threatening?

The following sections will examine the concepts of gentrification, collective efficacy, and group threat, and discuss ways in which these phenomena interact and have an effect on neighborhood levels of crime over time.

Gentrification

The genesis of the term ‘gentrification’ is generally acknowledged to come from urban geographer Ruth Glass. Glass used the term gentrification to refer to young, mostly single middle- and upper-class residents purchasing property in the historically impoverished area of London’s East End (Glass, 1964). From its beginning, the term had both socially and politically charged connotations; Glass described gentrification as a dangerous process because it was eliminating housing and driving poor residents out of their neighborhoods. By the 1970s, Glass’s tone describing the phenomenon grew more somber, describing the changes in London neighborhoods such as Hampstead and Chelsea as tragedies. While many scholars had taken to interpreting the phenomenon as a benign, even positive process of neighborhood revitalization, Glass described the process as more of an upper-middle-class invasion and succession (Glass, 1973).

Modern interpretations of the term focus on the changes in social demographics and social class experienced in a gentrifying neighborhood, and interpret the process as one involving young singles or young couples, and importantly, as one involving the migration of people within the city, and not moving into the city from suburban or rural areas (Butler, 1997; Hamnett, 2003; Smith and Williams, 1986). Prior research has long paid attention to the importance of the structural characteristics of neighborhoods and their relationship to crime (Sampson et al., 1997; Shaw and McKay, 1942). Once the process of gentrification started to get scrutinized in the late-20th century, scholars began to examine how such rapid and dramatic changes to these structural factors may impact crime. Over time, research has varied in its suggestion of sometimes positive (Covington and Taylor, 1989; Lee, 2010; Taylor and Covington, 1988; Van Wilsem et al., 2006) or negative (O’Sullivan, 2005; Papachristos et al., 2011) associations between gentrification and crime. Others have explored the possibility of a nonlinear relationship, where gentrification is associated with initial rises in crime, but this effect diminishes as time passes (Kreager et al., 2011). This prior research collectively has suggested that gentrification’s disruption of social control forces may be the primary explanation for its impact on crime. While prior research has laid the foundation upon which gentrification and crime can be examined, this study hopes to build off this prior research and extend it by addressing some problems and unexamined areas. Specifically, this research will examine the sociological concepts of social control, collective efficacy, and the impact of group threat dynamics on the gentrification process.

Social control and collective efficacy

Two phenomena that will be closely examined for their relationship to gentrification and crime are informal social control and group threat. For almost a century, scholars have made important contributions to research demonstrating the importance of neighborhood informal social control mechanisms on crime (Shaw and McKay, 1942; Thrasher, 1927). In this work, scholars theorized that it is the neighborhoods, and not the people within them, which can be criminogenic.

Modern adaptations of these ideas have included the development of systemic models to demonstrate that informal social control is multidimensional and can have a multilevel impact on neighborhood crime (Bursik, 1999; Sampson, 1988; Sampson and Groves, 1989). Along with these developments, scholars introduced the concept of neighborhood collective efficacy and explored its association with crime and delinquency.

Collective efficacy is a concept that encompasses two general qualities of an area: social cohesion (the ‘collectivity’) and shared expectations for social control (the ‘efficacy’). The requirement for dense social ties to produce social control is deemphasized and it is a working trust, and not friendship, which can be enough to produce this collective efficacy in a community (Sampson, 2012). Empirically, the importance of collective efficacy in explaining crime within neighborhoods has been demonstrated both in the US and internationally (Hipp and Wickes, 2017; Maimon and Browning, 2010; Pratt and Cullen, 2005; Sampson, 2012). However, the ability of a neighborhood experiencing gentrification to exercise collective efficacy has yet to be examined. Therefore, prior research has demonstrated that collective efficacy is an asset for neighborhoods, but the assumption has been that the structural qualities of these neighborhoods would remain stable. This study will extend this prior research and examine how this neighborhood effect shapes our understanding of the association between gentrification and crime.

Group threat

There has been a recent trend in neighborhood-level research on crime and delinquency to revive a discussion of cultural mechanisms (Berg et al., 2012; Kubrin, 2017). In this study, the impact of minority and group threat are examined for their interaction with the gentrification process in affecting crime rates over time.

Group-threat theory suggests that when the majority group in an area perceives threats from a minority group, prejudice and hostility toward this minority group will have a variety of negative outcomes (King and Wheelock, 2007). This process is thought to become amplified when the minority group presence increases or they mobilize and attempt to seize additional community resources and privileges (Blalock, 1967; Bobo, 1988; Bonilla-Silva, 1997, 2006). The majority of the literature on group threat examines this dynamic looking at a White majority reacting to the invasion of a Black minority group presence. These studies have confirmed that while White residents claim to support integrated communities, they still prefer to be in the majority and thus keep receiving the majority of resources and privileges (Bobo, 1988; Clark, 1991; Frey, 1979; Levine and Campbell, 1972).

In this study, the concept of perceived neighborhood change is introduced as a possible moderating force associated with the effect of gentrification on crime. If residents feel threatened, and perceive that changes in the environment and people in their area are not for the better, this implies a sense of unrest and discomfort. It is possible, therefore, that gentrification in which there is a push back from the community may increase the likelihood of violent crime as a result of altercations between old and new residents. Similarly, this unrest and discomfort may result in an increase in property crimes, both from an increase in property theft and destruction. However, in some neighborhoods the changes may not feel as threatening, and may in fact even be embraced by longtime residents. The gentrification of Harlem in the 1990s presents a good example of this. Many of the in-movers were young, middle-class African American families, and in several news reports it was described as the ‘New Harlem Rennaissance’, because those moving in embraced the long-standing culture and mystique of the area (Freeman, 2006; Williams, 2008). Therefore, it is suggested in this study that perceived threat may significantly alter the way in which gentrification interacts with crime. Survey measures in this project will shed light on whether or not this sense of threat and discomfort has an association with crime over time.

This study will merge the theoretical concepts of collective efficacy and perceived neighborhood change with the process of gentrification. As shown in Figure 1, it is suggested that examining gentrification for its overall impact on crime will result in a negative association, but this overall relationship is masking the changing impact of gentrification depending on contextual factors, which will be revealed once contextual factors are included. When gentrification takes places and the long-standing residents in a neighborhood perceive this as perceived neighborhood change, this will modify the relationship between gentrification and crime and these neighborhoods will experience an increase in crime.

Figure 1.

Theoretical relationships among gentrification, collective efficacy, perceived neighborhood change, and crime.

It is also suggested that collective efficacy will moderate the relationship between gentrification and crime by strengthening the negative association. For example, if long-standing residents of a neighborhood notice that a large turnover in their population has occurred, and this turnover results in a change to the composition of the area, it is possible that a neighborhood high in collective efficacy would take productive action (i.e. calling meetings of neighborhood associations, organizing events designed for residents to meet and get to know one another). In contrast, a neighborhood low in collective efficacy that experiences disruptive gentrification may respond negatively – where long-standing residents are dismissive and newer, more affluent residents seek to take control of the area and separate themselves from the longtime residents.

Data and analytic strategy

Data

This project uses data from the Project on Human Development in Chicago Neighborhoods (PHDCN), specifically data from the 1994–1995 Community Survey. The PHDCN is an interdisciplinary study focusing on the causal pathways that lead to both positive and negative outcomes for children and adolescents, as well as for neighborhoods. It includes data on family dynamics, schools, and neighborhoods and their effect on youth development (Earls et al., 2007). Survey questions taken for this project include items relevant to crime, perceived neighborhood change, and items tapping into the concept of collective efficacy, which captures the neighborhood’s social cohesion and willingness to intervene to maintain social control. For the purposes of this research study, the unit of analysis is the neighborhood cluster as defined in the PHDCN data. The cluster that is mostly filled with Chicago’s O’Hare Airport is excluded, leaving a total of 342 neighborhood clusters for analysis.

Data were also gathered from archival records of the Chicago Transit Authority (CTA) to ascertain the presence of ‘L’ train stations in each of the neighborhood clusters. Good access to metro services has been identified as a leading indicator of urban gentrification elsewhere (Turner and Snow, 2001), but this has yet to be included in an examination of gentrification in Chicago, so its inclusion in this study enhances the construction of the gentrification measure. To ascertain whether or not an L train station was in operation during the time frame of this project, a system map of the CTA from 1990 was examined to identify the street addresses of each of the L train stations. These stations were then located within each of the neighborhood clusters, and a binary variable was generated (1 = L station within cluster). These data are included with the census data measuring gentrification in the clusters.

Measures and analytic strategy

A description of all the variables used in the study is included in Table 1. The following sections will describe the measurement of each variable in detail.

Table 1.

Summary of dependent and independent variables, PHDCN data.

Variables	Coding scheme	Description
Dependent variables
Perceived violence	Ordinal scale	Violence in past 6 months
Individual-level independent variables
Gender	1 = female
Race-ethnicity	4 dummy variables	Black, Latino, Other, White omitted
Age	Continuous
Socioeconomic disadvantage	Index	Family below poverty line, living on public assistance, Unemployed, female-headed households, under 18
Years at address	Continuous
Neighborhood-level independent variables
Racial-Ethnic composition	3 Dummy variables	70% A.A., 70% Latino, Mixed, with 70% White omitted
Perceived Neigh. Change	Scale	3-item scale; alpha = 0.78
Gentrification	Scale	6-item scale; alpha = 0.85
Collective Efficacy	Scale

PHDCN: Project on Human Development in Chicago Neighborhoods.

Dependent variable

The primary outcome of interest in this study is perceived neighborhood violence, which was collected through survey questions asking citizens to report on violent incidents that have happened in the neighborhood in the past 6 months. Details of the collection of this measure are available through the publicly accessible codebook of the PHDCN Community Survey data (Earls et al., 2007).

Independent variables

Gentrification

The gentrification scale was created drawing data from the 1990 US Census and the CTA for a more complete consideration of the process. In this research, all neighborhoods with an average home value below the city’s 1990 average (US$110,000) are considered low-priced areas (United States Census Bureau, 1990). In total, there were 121 neighborhood clusters that could be defined as low-priced by this definition. To identify areas of gentrification potential in 1990, the following dichotomous variables were included in a summed scale of binary variables, in which responses for each neighborhood was either yes (= 1) or no as follows: (1) (low priced area . . .) adjacent to a high-priced area, (2) containing an L train station within its boundaries, (3) with the majority (>50%) of residences having historic architecture (built before 1940), (4) with the majority (>50%) of residences having large (5+) housing units, (5) with less than 20% eligible (persons 25+) with a bachelor’s degree, and (6) with median income below the median income for the city (US$30,707).

The responses were summed and divided by the total number of items in the scale: higher scores indicate a higher level of gentrification potential. This gentrification potential scale is modified from a similar scale used to measure gentrification potential by the Urban Institute in their 2001 analysis of Washington, DC (Turner and Snow, 2001). The gentrification scale has a validity (Cronbach’s alpha) of α = 0.85, which suggests that it has internal consistency, and the items included in the gentrification scale are appropriate for measuring the gentrification potential. Figure 2 shows the distribution of neighborhood clusters across the gentrification scale. Approximately 49% of the neighborhood clusters (170 clusters) scored relatively low on the gentrification scale (0.06 to 0.30). There were 34 neighborhood clusters (9.9%) scoring relatively high on the scale (0.75 to 1.0).

Figure 2.

Level of gentrification in neighborhood clusters (N = 342).

To verify that this scale accurately identified gentrifying neighborhoods, data from all 342 neighborhood clusters were examined with the 1990 and 2000 Census data. In this way we can confirm that those neighborhoods that seemed to have a high potential for gentrification in 1990 did in fact experience this process throughout that decade. The gentrification measures are directly measured through the 1990 and 2000 Census data, but not in the social survey data from 1994 to 1995. Neighborhoods scoring extremely high (⩾75th percentile) on the scale should have significantly changed on these measures, whereas neighborhoods scoring lower on the scale should have seen less significant changes over time. Table 2 presents the results of this verification analysis. As predicted, neighborhoods in the 75th percentile or higher on the gentrification scale changed in the predicted ways, but neighborhoods in the lower percentiles did not have significant changes to most of these indicators. This increases confidence in the validity of the gentrification scale: it accurately predicted those neighborhoods that had the most dramatic change.

Table 2.

One-tailed t-test comparison of gentrification measures from US Census (1990–2000).

	1990	2000
⩾75th percentile on scale
Home value (1990 adj.	US$59,326.43	US$136,203.83	***
Large (5+) housing units (%)	67.33	29.99	***
Eligible persons with bachelor’s degree (%)	8.97	35.33	***
Median income (1990 adj.)	US$17,336.71	US$51,030.40	***
50th–74th percentile
Home value (1990 adj.)	US$61,743.21	US$65,966.77	*
Large (5+) housing units (%)	48.25	41.42	*
Eligible persons with bachelor’s degree (%)	15.85	19.34
Median income (1990 adj.)	US$21,321.34	US$23,435.29
25th–49th percentile
Home value (1990 adj.)	US$71,487.61	US$73,589.21
Large (5+) housing units (%)	42.55	40.68
Eligible persons with bachelor’s degree (%)	26.34	28.47
Median income (1990 adj.)	US$34,899.00	US$37,641.00
1st–24th percentile
Home value (1990 adj.)	US$95,374.54	US$101,255.78	*
Large (5+) housing units (%)	15.31	14.27
Eligible persons with bachelor’s degree (%)	34.77	36.98
Median income (1990 adj.)	US$45,745.36	US$49,678.45

p < 0.05.

***

p < 0.0001.

Collective efficacy

Consistent with prior research, the measure of collective efficacy in this study is identical to the scale as was originally developed by Sampson et al. (1997), and has been used in a variety of follow-up research examining the construct both in the United States and internationally (Kirk and Papachristos, 2011; Wickes et al., 2013). The scale can be broken down to the measurement of three basic elements: social control, social cohesion, and trust. Questions assessing the level of neighborhood social control asked respondents to indicate their willingness to intervene in the following scenarios: (1) children skipping school, hanging out on street corners, (2) children spray-painting graffiti on a local building, (3) children showing disrespect to an adult, (4) fighting in front of houses and someone was being beaten or threatened, and (5) budget cuts causing closest fire station to be closed down (ICPSR 2766: 10). To measure social cohesion and trust, the following items were combined: (1) people around here are willing to help their neighbors, (2) people in this neighborhood can be trusted, (3) people in this neighborhood generally get along with each other, (4) this is a close-knit neighborhood, and (5) people in this neighborhood share the same values. The scale for collective efficacy was constructed using multilevel models, accounting for responses to each question coming from individuals who are nested within neighborhoods. From this, neighborhood-specific empirical Bayes (EB) residuals are used as the scale. In this way, similar to prior research, the collective efficacy scale represents the average level of collective efficacy in each neighborhood (Kirk and Matsuda, 2011; Sampson et al., 1997).

Perceived neighborhood change

The measure of perceived neighborhood change (examining the possibility of group threat dynamics) in the current analysis is constructed as a scale, combining responses to questions in which participants were asked about changes to the neighborhood, and whether these changes make the area ‘better’, ‘worse’, or ‘the same’. The following items were considered: (1) During the past 5 years has this (neighborhood’s looks) changed for the better, stayed about the same, or gotten worse? (2) Have the people living in the neighborhood changed for the better, stayed the same, or gotten worse? (3) In the next 5 years, do you think this neighborhood will change for the better, remain the same, or get worse? The perceived neighborhood change scale has reliability (Cronbach’s alpha) of α = 0.78. The mean score is 0.39, with higher numbers indicating a higher level of perceived neighborhood change in the neighborhood. Similar to the construction of the collective efficacy measure, to generate a variable measuring the average level of perceived neighborhood change in a neighborhood multilevel modeling was used to generate neighborhood-specific EB residuals.

Neighborhood-level controls

Several structural characteristics were included in the analysis, taken from the 1990 and 2000 Census. The 1990 data were used for the primary analyses, while the 2000 data were used to assess demographic changes to the neighborhoods relevant to our examination of group threat (see Figure 3). Consistent with prior research, the racial-ethnic composition of each neighborhood was measured using dummy variables (Kirk and Matsuda, 2011). African American NC refers to clusters with a 70% or higher African American population. Similar variables were constructed for White NC and Latino NC. For the variable Mixed NC, the clusters were coded ‘1’ if they had less than 70% of one single group. In this study, the largest proportion of neighborhoods were classified as ‘Mixed’ (52%), with 22% identified as African American, 16% identified as White, and 10% identified as Latino. The concentration of poverty of the neighborhood clusters is also examined with a scale assembled from the 1990 US Census. This measure includes the percentage of families below the poverty line, percent receiving public assistance, percent unemployed, percent female-headed households, and percent below the age of 18. This scale mirrors scales of concentrated poverty that have been constructed to examine PHDCN neighborhood clusters in prior research (Kirk and Matsuda, 2011; Sampson et al., 1997).

Figure 3.

Variation in perceived neighborhood change score in gentrifying neighborhoods by change in racial composition 1990–2000 (N = 87).

Analytic strategy

In the PHDCN Community Survey, the participants are nested within each neighborhood cluster, which means that the data include individuals who are clustered within similar surroundings and environmental conditions. Analyses utilizing single-level modeling strategies are therefore inappropriate, because these assume that the variance between individuals is constant, and that although individuals are nested within similar conditions they differ at random. It is necessary, therefore, to account for this clustering to address any violation of independent error terms across neighborhoods (Raudenbush and Bryk, 2002). Hierarchical linear modeling (HLM) is appropriate for the current analyses for several reasons. HLM will allow this project to model individual-level characteristics and neighborhood-level characteristics, while taking into account within- and between-neighborhood variation. This project uses HLM 6.0 to conduct all multilevel modeling analyses. HLM is chosen because perceived violence is a continuous variable. The model follows a normal distribution of the outcome variable. The analytic strategy for this research will first involve assessing the significance of variation in variables at the neighborhood level, and then building multilevel models in a series of steps. This will allow an examination of the relationship between the key independent variables of interest (gentrification, perceived neighborhood threat, and collective efficacy) and crime, and will also allow an examination of the interactions between these variables.

Results

Table 3 displays the descriptive statistics for the sample, which provide preliminary evidence for inter-neighborhood variation. The sample is 59% female, with an average age of 42.7 years. The same is predominantly African American (40%) and the respondents had spent an average of 10.39 years at their current address at the time of the survey. Approximately 33.4% of the respondents had lived in their neighborhood for less than 5 years. This suggests that the majority of the survey’s participants were likely not ‘gentrifiers’ to the neighborhood.

Table 3.

Descriptive statistics for study variables, PHDCN community survey 1994–1995.

	N	Mean	SD	Min	Max
Dependent variable
Perceived violence	7739	2.02	0.41	1.2739	3.3104
Individual-level variables (N = 7739)
Female		0.59	0.49	0	1
African American		0.40	0.49	0	1
White		0.28	0.45	0	1
Other		0.08	0.48	0	1
Latino		0.25	0.43	0	1
Age		42.66	16.72	17	100
Socioeconomic disadvantage index		44.26	18.07	17	97
Years at address		10.39	11.97	0	81.5
Neighborhood-level variables (N = 342)
Hispanic neighborhood (70%+)		0.10		0	1
White neighborhood (70%+)		0.18		0	1
Black neighborhood (70%+)		0.55		0	1
Mixed neighborhood		0.17		0	1
Gentrification		0.35	0.29	0.06	1
Collective efficacy		3.54	0.90	1.03	4.99
Perceived neighborhood change		1.94	1.41	0.00	4.96

PHDCN: Project on Human Development in Chicago Neighborhoods; SD: standard deviation.

The variation in the neighborhood-level variables provides preliminary evidence that there is between-neighborhood variation. For example, while the average score on gentrification scale was 0.35, neighborhoods ranged from 0.06 to 1.0. Similarly, for the variables capturing collective efficacy and perceived neighborhood change, there is considerable range in the scores across neighborhoods on both of these measures.

A zero-order correlation matrix (Table 4) indicates there is a strong negative association between perceived neighborhood change and collective efficacy (r = −0.645, p < 0.001). This suggests that neighborhoods with high levels of threat and dissatisfaction with the changes to their neighborhood are less likely to engage in collective efficacy – at least as it relates to violence and crime. There is also a strong positive association between perceived neighborhood change and gentrification (r = 0.621, p < 0.01), which indicates that in many neighborhoods where gentrification is happening, there are feelings of dissatisfaction, threat, and resentment toward the changes. The results also indicate that gentrification is more likely in predominantly African American neighborhoods (r = 0.548, p < 0.001) or in neighborhoods with no racial-ethnic group comprising the majority (r = 0.489, p < 0.01), but there is no significant association between gentrification and Latino neighborhoods (r = 0.120). Neighborhoods that are predominantly Latino (r = −0.175, p < 0.05) or African American (r = −0.376, p < 0.01) are less likely to have high collective efficacy scores, but high collective efficacy scores are significantly more likely in mixed neighborhoods (0.348, p < 0.05). In addition, there is a modest but significant negative association between collective efficacy and gentrification (r = −0.189, p < 0.05).

Table 4.

Neighborhood-level zero-order correlation matrix.

Variables	1	2	3	4	5	6
1. Perceived neighborhood change	–	–	–	–	–	–
2. African American NC	0.548***	–	–	–	–	–
3. Latino NC	0.120	−0.785***	–	–	–	–
4. Mixed NC	0.489**	−0.682**	−0.473**	–	–	–
5. Collective efficacy	−0.645***	−0.376**	−0.175*	0.348*	–	–
6. Gentrification	0.621**	0.501**	−0.242*	0.627**	−0.189*	–

NC: neighborhood cluster. N = 342.

p < 0.05.

p < 0.01.

***

p < 0.001.

To examine the effect of gentrification, collective efficacy, and perceived neighborhood change on crime, a series of multilevel models are estimated. Model 1 in Table 5 is an examination of individual-level factors, and reveals that there are significant differences in perceived violence by gender, racial-ethnic group, and by socioeconomic status. Similar to prior research (Kirk and Matsuda, 2011; Morenoff et al., 2001), being disadvantaged (b = 0.345 p < 0.05), male (b = −1.879, p < 0.001), and African American (b = 1.754, p < 0.001) is associated with a heightened perception of violence. Living more years at the current address was associated with a decreased perception of violence, but this relationship did not reach statistical significance (b = −0.212). In addition, the relationship between age and perceived violence did not reach statistical significance, but it appears that higher levels of perceived violence were reported by younger respondents (b = −0.312).

Table 5.

Multilevel model of perceived violence with individual-level characteristics.

Model 1 (perceived violence)
Variables	Coefficient		SE
Intercept	1.024	***	0.085
Individual-level variables (N = 7739)
Female	−1.879	***	0.165
African American	1.754	***	0.263
Other	−1.035	*	0.294
Latino	1.075	***	0.260
Age	−0.312		0.064
Socioeconomic index	0.345	*	0.104
Years at address	−0.212		0.078

SE: standard error.

p ⩽ 0.05.

***

p ⩽ 0.001.

Table 6 introduces both the collective efficacy and perceived neighborhood change factors at the neighborhood level. In Model 2, the primary neighborhood-level variable of interest was introduced (gentrification). For every additional increase in the level of gentrification on the scale, the level of neighborhood violence decreases by an average of 0.259, and this effect is statistically significant (b = −0.259, p < 0.01). These data suggest that neighborhoods experiencing gentrification also experience a significant decline in perceived violence. This effect is significant after controlling for the racial-ethnic composition of the neighborhood and the level of concentrated poverty in the neighborhood.

Table 6.

Hierarchical linear models of perceived violence.

Variables	Model 2			Model 3			Model 4
	b		SE	b		SE	b		SE
Intercept	1.014	***	0.088	1.018	***	0.090	1.021	***	0.089
Individual-level variables (N = 7739)
Female	−1.897	***	0.142	−1.853	***	0.138	−1.867	***	0.211
African American	1.698	***	0.234	1.742	***	0.337	1.755	***	0.286
Other	−1.024	**	0.317	−0.980	*	0.301	−0.984	*	0.304
Latino	1.107	***	0.249	1.151	***	0.243	1.162	***	0.287
Age	−0.246		0.048	−0.202		0.056	−0.238		0.102
Socioeconomic index	.383	*	0.099	0.382	*	0.102	0.401	*	0.098
Years at address	−0.241		0.077	−0.230		0.075	−0.195		0.077
Neighborhood-level variables (N = 342)
Latino NC	1.249	†	0.745	1.127		0.731	1.121		0.789
Black NC	1.568	**	0.588	1.560	**	0.426	1.534	**	0.475
Mixed NC	1.372	*	0.476	1.278	*	0.385	1.272	*	0.362
Concentration of poverty	0.211	*	0.144	0.126		0.137	0.123		0.135
Gentrification	−0.259	**	0.342	−0.197	**	0.296	−0.107		0.301
Collective efficacy				0.075		0.145	0.080		0.133
Per neighborhood change				0.530	**	0.217	0.186	†	0.223
Gent × PNC							0.374	*	0.314
Gent × collective efficacy							−1.425	**	0.392
Proportion of variance explained			0.35			0.40			0.52

NC: neighborhood cluster.

SE: standard error.

†

p < 0.10.

p ⩽ 0.05.

p ⩽ 0.01.

***

p ⩽ 0.001.

In Model 3, the negative correlation between gentrification and perceived violence remains significant, but the effect size has been reduced (b = −0.197, p < 0.01). Perceived neighborhood change has a positive and significant association with perceived violence (b = 0.530, p < 0.01), suggesting that the level of violence in a neighborhood increases by 0.530 for every increase in the level of perceived neighborhood change. The relationship between a neighborhood’s average level of collective efficacy and perceived violence is not significant (b = 0.075).

Interaction terms are introduced in Model 4 of Table 6 to examine the interaction between the key variables of interest with gentrification. Perceived neighborhood change and collective efficacy were each examined for potential interaction effects with gentrification. Model 4 suggests that the interaction between gentrification and perceived neighborhood change results in a positive and significant association with violence (b = 0.374, p < 0.05). The results reveal that the relationship between gentrification and perceived violence varies depending on the level of perceived neighborhood change in neighborhoods. An interaction between gentrification and increasing perceived neighborhood change is associated with a significant increase in levels of violence. The interaction term included to examine whether there is an interaction between gentrification and collective efficacy suggests a significant negative association (b = −1.425, p < 0.01). It appears from this result that collective efficacy in gentrifying neighborhoods may strengthen the negative association with violence.

The fact that the interaction terms suggest a different association between the factors and perceived violence seems to complement data from the Table 4 correlation matrix; it appears that there is not much overlap between neighborhoods high in collective efficacy and high in perceived neighborhood change (b = −0.645, p < 0.001), so the fact that these contextual factors produce different effects with gentrification on perceived violence indicates that the way in which gentrification impacts perceived violence is highly dependent on neighborhood contextual factors. It is informative that these divergent contextual factors have varying effects on areas experiencing gentrification. Collective efficacy appears to work as a protective factor, whereas when gentrification is accompanied by high levels of perceived neighborhood change, this increases a neighborhood’s level of violence.

This research also examines whether perceived neighborhood change may increase when a gentrifying neighborhood experiences a significant shift in the racial or ethnic profile of the area. To examine this, the neighborhoods scoring in the 75th percentile or higher on the gentrification scale were included, which resulted in 87 clusters for analyses scoring from 0.58 to 1.0 on the scale. Of the 87 neighborhoods, 61% were neighborhoods that were predominantly African American in 1990 (⩾70%). A total of 10% were neighborhoods that were predominantly White and 6% were predominantly Latino neighborhoods. If perceived neighborhood change is indeed a factor that is driven by racial-ethnic characteristics, we would expect higher levels of perceived neighborhood change in neighborhoods that saw a dramatic compositional change from 1990 to 2000. Figure 3 presents data that support this claim, showing variation in perceived neighborhood change by change in the racial-ethnic profile of the population.

For gentrifying neighborhoods that began the decade predominantly African American, those with a decline in the proportion of African Americans had significantly higher perceived neighborhood change scores than those where the neighborhood remained predominantly African American; a one-tailed t-test at α 0.05 yielded a p value of 0.032. For gentrifying neighborhoods that began the decade predominantly Latino, those that were no longer predominantly Latino by the end of the decade had higher perceived neighborhood change scores than those that remained predominantly Latino, but this difference was only marginally significant (p value 0.091). For gentrifying White neighborhoods, those remaining predominantly White were not significantly different from those whose racial profile changed in their level of perceived neighborhood change (p value 0.621).

These results offer partial support for the idea that perceived neighborhood change is higher when gentrification causes the racial-ethnic composition of a neighborhood to significantly change. For gentrifying neighborhoods that began predominantly composed of racial-ethnic minorities, changes to this composition were associated with higher levels of perceived neighborhood change. Given that the perceived neighborhood change measure specifically focuses on feelings related to population change, it is reasonable to assume that threat and changing racial-ethnic dynamics are linked. These findings complement prior research, while also presenting the novel finding that these dynamics are similar when other racial-ethnic groups are the majority in an area.

Discussion and conclusion

Gentrification continues to enter into the public discourse throughout the country. It is doubtful that gentrification will cease, and the tensions surrounding this process remain high. While many neighborhoods in Chicago experienced gentrification in the 1990s, the process continues, and elicits powerful reactions from the residents experiencing the changes. In January 2015, for example, anonymous protestors taped up signs (shown below) in front of a few coffee shops in the Chicago neighborhood of Pilsen to protest gentrification efforts that are changing the area.

Picture 1.

Photo taken outside Bow Trust Coffee Roasters, Pilsen, Chicago, IL., January, 2015.

As this process continues to be debated in the public discourse, it is incumbent upon the field of criminology to continue to examine gentrification and how it interacts with neighborhood factors and crime. As gentrification remains an extremely polarizing topic, many are continuously asking the basic question, ‘Will gentrification make neighborhoods better?’ The answer to this question may help to develop policies and strategies that can be put into place as such neighborhood transformations occur.

This study finds that when all neighborhoods experiencing gentrification are examined together, they do seem to have lower levels of violence. This result is consistent with the prior research (O’Sullivan, 2005; Papachristos et al., 2011; Velez et al., 2012), and seems to suggest that, in the aggregate, neighborhoods that experience gentrification will experience less perceived violence. However, this study extends prior research by examining ways in which this effect might vary by neighborhood context. With the finding that collective efficacy moderates the gentrification and perceived violence relationship by strengthening the negative association, it is suggested that a more cohesive, trusting neighborhood might have an even stronger crime-reduction benefit as gentrification unfolds. This study supports this idea; collective efficacy seems to strengthen the crime-reduction benefit of gentrification on perceived violence. There are several potential reasons why this may be the case. Collective efficacy has sometimes been described as a dynamic factor that is partly endogenous, or contingent upon the challenges at hand (Morenoff et al., 2001; Sampson, 2012). When posed with the challenges that accompany neighborhood turnover and change, perhaps neighborhoods high in collective efficacy do better because they take collective action to welcome new residents and new businesses into the area. They may go to meet their new neighbors and invite them to community meetings and events, instead of ignoring them and excluding them from local organizations. And earlier in the process, they may make efforts to partner with developers to become attached and invested in the outcome of the process.

Edison Park, a northwest Chicago neighborhood, provides anecdotal evidence for the power of collective efficacy in action during gentrification efforts. In this neighborhood, the residents took an active role in the process of redevelopment. Newspaper articles from the time suggest that while residents enjoyed most of the changes, they took action to address certain elements they did not support:

Recently, a community group, worried that the streets might become more clogged with parked cars and related problems if new bars and restaurants move in, asked the city to prohibit any additional liquor licenses from being issued in the ‘Town Center’ area along Northwest Highway. The proposed ban has attracted few critics and is supported by Edison Park’s Chicago City Council representative, Ald. Brian Doherty (41st). (Chase, 1998: 1)

In an article discussing changes to the Bucktown area, Alderman Terry Gabinsky (32^nd District), a 46-year-old resident of the neighborhood had a similar positive experience with the gentrification process because of residents being proactive to participate in the changes:

We’re meeting our goal, which, from my point of view, is a change from the absentee landlords we suffered through in the late ’60s and ’70s. Owner-occupied property is a majority now, though I can’t give you a number, and brings back stability. People who own their property are more aggressive about what they want in terms of city services, whereas absentee landlords only care about the rent. (Lauerman, 1992: 2)

A neighborhood without the ability to take action and intervene to tackle such issues in this manner might grow upset over time if the changes to their area were upsetting their ability to use public space. This study supports the idea that gentrification, coupled with collective efficacy, can help to improve neighborhood perceived violence. By taking action and addressing problems, perhaps gentrifying neighborhoods with collective efficacy enjoy lower violence and victimization, because they stop problems from developing into criminogenic situations. This result is also complementary to prior work on collective efficacy. Scholars have often noted that this neighborhood quality can play a powerful role in solving problems and reducing crime (Mazerolle et al., 2010; Morenoff et al., 2001; Sampson et al., 1997; Wickes, 2010). In this project, the results complement prior research by demonstrating that collective efficacy can specifically help reduce perceived violence in gentrifying neighborhoods.

Another conclusion to be drawn from this study is that neighborhoods perceiving the changes to their community as threatening may experience rises in perceived violence. Prior research supports these conclusions, although most prior research assumes that the group being threatened is middle-class White residents (Eitle and Taylor, 2008; Hipp, 2007; King and Wheelock, 2007). In part, the results in this study may be due to violent clashes between new and longtime residents, and it may manifest itself through property crimes designed to target newly redeveloped spaces and residences. It could also be the case that new residents instigate confrontations regarding the use of public space, appearance of residences, and power over resources. The results in this study found similar positive associations between gentrification and perceived neighborhood change for violence. The sense of threat in these changing Chicago neighborhoods is perhaps best personified with this quote from Erskine Sankey, a longtime resident of the northeast neighborhood Uptown along Chicago’s border with Lake Michigan. Sankey was interviewed about the rising tensions in the area:

My main reason for [protesting] is the woman who has two children and makes $300 a month on public aid isn’t welcome here, but the guy who makes $100,000 a year is. (Martin, 1996: 1)

The supplementary analyses examining variation in perceived neighborhood change by change in the racial-ethnic composition of gentrifying neighborhoods also supports the importance of group threat dynamics. The results suggest that gentrifying neighborhoods that began the process predominantly composed of one racial-ethnic minority scored much higher on the perceived neighborhood change measure if they lost predominance in the neighborhood. In this way, the study supports the idea that racial and ethnic dynamics remains an important way in which neighborhoods establish their identity; the threatening of this racial-ethnic identity seems to be a driving force behind perceptions of perceived neighborhood change. Prior research in group threat dynamics offers confirmation that race-ethnicity often plays an important role in establishing group dynamics (Kane, 2003; King and Wheelock, 2007), and this study reinforces these findings.

In sum, this study both complements and extends prior research examining the relationship between gentrification and perceived violence within neighborhoods. Several of the most recent studies have found a negative relationship between gentrification and subsequent crime (Kreager et al., 2011; Papachristos et al., 2011; Velez et al., 2012) and in the examination of the overall effect in Chicago, this study comes to a similar conclusion. However, this study also extends prior research by examining potential mechanisms that either strengthen or weaken this relationship, and the results suggest that the context within which gentrification emerges plays a large role in its effect.

There are a few important areas of uncertainty and limitations to this study, which warrant attention and suggestions for future study. First, as the dependent and moderating variables were gathered through the same survey, the cross-sectional nature of these data makes it impossible to fully examine the directions of causal relationships over time. For example, perceived violence in a neighborhood over time could possibly have its own impact on collective efficacy. Rose and Clear’s (1998) research on the cyclical nature of violence on neighborhood social control certainly suggests this as a possibility. While the data used in this study make examination of this impossible, it will be important for future research to examine such variables over multiple points in time. Another important element of uncertainty in the results of this study is the potential for measurement error in capturing gentrification by examining neighborhoods with data gathered 10 years apart. This study aimed to identify neighborhoods that seemed to have the highest potential for gentrification, and then verify that it took place with subsequent data, but it would be preferable in the future to gather data on key indicators of gentrification as it is happening. It will also be important in future research to examine variations in the rapidity of gentrification across and within neighborhoods. Future research would also benefit from examining gentrification processes in other urban areas. The city of Chicago has provided criminologists with a rich dataset to examine neighborhood context for several decades, but the examination of gentrification and crime should be expanded to other cities. Several scholars have already begun to do this, examining gentrification and crime patterns in Baltimore, Boston, San Francisco, Seattle, and New York City, to name a few (Atkinson, 2002; Kreager et al., 2011; McDonald, 1986; Taylor and Covington, 1988). However, these studies have not yet examined the role of neighborhood contextual factors. The survey administered through the PHDCN has been administered in a few other urban areas (Wickes et al., 2011); future research in gentrification and crime should take advantage of these replications and gather the appropriate data on gentrification to determine if neighborhood contextual factors operate in similar manners as they have in this study.

Footnotes

Acknowledgements

Raymond Paternoster is acknowledged for his past advice and guidance in the research and writing of this manuscript.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Author biography

Kathryn Kozey is an Assistant Professor of Criminal Justice at Western New England University. She received her PhD in Criminology and Criminal Justice from the University of Maryland in 2015, and her research focuses on the integration of sociological and criminological theories to examine crime within neighborhood and community contexts. She is also currently pursuing a qualitative research project examining the use of social media in disclosing accounts of past sexual assaults.

References

Atkinson

(2002) Does gentrification help or harm urban neighborhoods? An assessment of the evidence-base in the context of the new urban agenda. CNR paper 5. Available at: http://neighbourhoodchange.ca/wp-content/uploads/2011/06/Atkinson-2002-Gentrification-Review.pdf

Berg

Stewart

Schreck

, et al (2012) The victim-offender overlap in context: Examining the role of neighborhood street culture. Criminology 50(2): 1–31.

Blalock

(1967) Toward a Theory of Minority-group Relations. New York: John Wiley & Sons.

Bobo

(1988) Group conflict, prejudice, and the paradox of contemporary racial attitudes. In: Katz

Taylor

(eds) Eliminating Racism. New York: Springer, pp. 85–114.

Bonilla-Silva

(1997) Rethinking racism: Toward a structural interpretation. American Sociological Review 62(3): 465–480.

Bonilla-Silva

(2006) Racism without Racists: Color-blind Racism and the Persistence of Racial Inequality in the United States. Lanham, MD: Rowman & Littlefield.

Bursik

(1999) The informal control of crime through neighborhood networks. Sociological Focus 32(1): 85–97.

Bursik

Grasmick

(1993) Neighborhoods and Crime: The Dimensions of Effective Community Control. New York: Lexington.

Butler

(1997) Gentrification and the Middle Classes. London: Ashgate.

10.

Chase

(1998) Lincoln Park Northwest? Gentrification hits Edison Park. Chicago Tribune, 20 December. Available at: https://www.chicagotribune.com/news/ct-xpm-1998-12-08-9812080144-story.html

11.

Clark

WAV

(1991) Residential preferences and neighborhood racial segregation: A test of the Schelling segregation model. Demography 28(1): 1–19.

12.

Covington

Taylor

(1989) Robbery and larceny changes in appreciating Baltimore neighborhoods during the 1970s. Urban Affairs Quarterly 25(1): 142–172.

13.

Earls

Brooks-Gunn

Raudenbush

, et al. (2007) Project on Human Development in Chicago Neighborhoods: Community Survey, 1994-1995. Ann Arbor, MI: Inter-university Consortium for Political and Social Research.

14.

Eitle

Taylor

(2008) Are Hispanics the new ‘Threat’? Minority group threat and fear of crime in Miami-Dade County. Social Science Research 37(4): 1102–1115.

15.

Freeman

(2006) There Goes the ‘Hood: Views of Gentrification from the Ground Up. Philadelphia, PA: Temple University Press.

16.

Frey

(1979) Central city white flight: Racial and nonracial causes. American Sociological Review 44(3): 425–448.

17.

Glass

(1964) London: Aspects of Change. London: MacGibbon & Kee.

18.

Glass

(1973) The mood of London. In: Donnison

Eversely

(eds) London: Urban Patterns, Problems and Policies. London: Heinemann, pp. 159–183.

19.

Hamnett

(2003) Gentrification and the middle-class remaking of inner London. Urban Studies 40(12): 252–261.

20.

Hipp

(2007) Block, tract, and levels of aggregation: Neighborhood structure and crime and disorder as a case in point. American Sociological Review 72(5): 659–680.

21.

Hipp

Wickes

(2017) Violence in urban neighborhoods: A longitudinal study of collective efficacy and violent crime. Journal of Quantitative Criminology 33(4): 783–808.

22.

Kane

(2003) Social control in the metropolis: A community-level examination of the minority group-threat hypothesis. Justice Quarterly 20(2): 265–295.

23.

King

Wheelock

(2007) Group threat and social control: Race, perceptions of minorities and the desire to punish. Social Forces 85(3): 1255–1280.

24.

Kirk

Matsuda

(2011) Legal cynicism, collective efficacy, and the ecology of arrest. Criminology 49(2): 443–472.

25.

Kirk

Papachristos

(2011) Cultural mechanisms and the persistence of neighborhood violence. American Journal of Sociology 116(4): 1190–1233.

26.

Kreager

Lyons

Hays

(2011) Urban revitalization and Seattle crime, 1982–2000. Social Problems 58(4): 615–639.

27.

Kubrin

(2017) Cultural disorganization and crime. In: Cullen

(ed.) Challenging Criminological Theory: The legacy of Ruth Rosner Kornhauser. New York: Routledge.

28.

Lauerman

(1992) Growing Pains: New money and new residents are changing the character of Bucktown and Wicker Park. Not everyone thinks the changes are for the better. Chicago Tribune, 18 October. Available at: https://www.chicagotribune.com/news/ct-xpm-1992-10-18-9204040535-story.html

29.

Lee

(2010) Gentrification and crime: Identification using the 1994 Northridge earthquake in Los Angeles. Journal of Urban Affairs 32(5): 549–577.

30.

Levine

Campbell

(1972) Ethnocentrism: Theories of Conflict, Ethnic Attitudes, and Group Behavior. Oxford: John Wiley & Sons.

31.

McDonald

(1986) Does gentrification affect crime rates? Crime and Justice 8: 163–201.

32.

Maimon

Browning

(2010) Unstructured socializing, collective efficacy, and violent behavior among urban youth. Criminology 48(2): 443–474.

33.

Martin

(1996) Uptown gets hot under blue, white collars: Community policing office divides 2 sides. Chicago Tribune, 13 March. Available at: https://chicagotribune.newspapers.com/browse/Chicago%20Tribune_4351/1996/03/13/Main%20Edition

34.

Mazerolle

Wickes

McBroom

(2010) Community variations in violence: The role of social ties and collective efficacy in comparative context. Journal of Research in Crime and Delinquency 47(1): 3–30.

35.

Morenoff

Sampson

Raudenbush

(2001) Neighborhood inequality, collective efficacy, and the spatial dynamics of urban violence. Criminology 39(3): 517–560.

36.

O’Sullivan

(2005) Gentrification and crime. Journal of Urban Economics 57(1): 73–85.

37.

Papachristos

Smith

Scherer

, et al. (2011) More coffee, less crime? The relationship between gentrification and neighborhood crime rates in Chicago, 1991 to 2005. City and Community 10(3): 215–240.

38.

Pratt

Cullen

(2005) Assessing macro-level predictors and theories of crime: A meta- analysis. Crime and Justice 32: 373–450.

39.

Raudenbush

Bryk

(2002) Hierarchical Linear Models: Applications and Data Analysis Methods. 2nd ed. Thousand Oaks, CA: SAGE.

40.

Rose

Clear

(1998) Incarceration, social capital, and crime: Implications for social disorganization theory. Criminology 36(3): 441–480.

41.

Sampson

(1988) Local friendship ties and community attachment in mass society: A multilevel systemic model. American Sociological Review 53(5): 766–779.

42.

Sampson

(2012) Great American City: Chicago and the Enduring Neighborhood Effect. Chicago: University of Chicago Press.

43.

Sampson

Groves

(1989) Community structure and crime: Testing social- disorganization theory. American Journal of Sociology 94(4): 774–802.

44.

Sampson

Raudenbush

Earls

(1997) Neighborhoods and violent crime: A multilevel study of collective efficacy. Science 277(5328): 918–924.

45.

Shaw

McKay

(1942) Juvenile Delinquency in Urban Areas. Chicago: University of Chicago Press.

46.

Smith

Williams

(1986) The Gentrification of the City. Boston: Allen & Unwin.

47.

Taylor

Covington

(1988) Neighborhood changes in ecology and violence. Criminology 26(4): 553–589.

48.

Thrasher

(1927) The Gang: A Study of 1,313 Gangs in Chicago. Chicago: University of Chicago Press.

49.

Turner

Snow

(2001) Leading indicators of gentrification in D.C. neighborhoods: D.C. Policy Forum. Available at: http://www.urban.org/publications/900461.html

50.

United States Census Bureau (1990) Selected Housing Characteristics. Available at: https://www.census.gov/main/www/cen1990.html

51.

Van Wilsem

Wittebrood

De Graaf

(2006) Socioeconomic dynamics of neighborhoods and the risk of crime victimization: A multilevel study of improving, declining, and stable areas in the Netherlands. Social Problems 53(2): 226–247.

52.

Velez

Lyons

Boursaw

(2012) Neighborhood housing investments and violent crime in Seattle: 1981–2007. Criminology 50(4): 1025–1056.

53.

Wickes

(2010) Generating action and responding to local issues: Collective efficacy in context. Australian and New Zealand Journal of Criminology 43(3): 423–443.

54.

Wickes

Zahnow

White

, et al. (2011) Sensing cohesion or diversity? Examining the impact of ethnic diversity on cohesion and neighbor networks in urban communities. In: International RC21 conference, Amsterdam, 7–9 July 2011.

55.

Wickes

Zahnow

White

, et al. (2013) Ethnic diversity and its impact on community social cohesion and neighbourly exchange. Journal of Urban Affairs 36(1): 51–78.

56.

Williams

(2008) Harlem Renaissance. Bloomington, IN: AuthorHouse.