Abstract
Place attachment is the development of a psychological and emotional bond between an individual and their environmental setting. While positive experiences in the residential neighborhood are central to ongoing develop-ment of people–place bonds, whether negative experiences erode place attachment remains unknown. In this study, we explore the relationship between crime victimization, social ties, neighboring behaviors, and place attachment in Brisbane, Australia. Using multilevel linear modeling, we examine whether negative experiences, specifically crime victimization, in the residential neighborhood affect residents’ attachment to place. We also explore whether this relationship is moderated by neighborhood social ties and/or interactions with neighbors. Results indicate that the negative impact of victimization in the residential neighborhood on place attachment is attenuated through frequent social and/or functional interactions with neighbors.
Place attachment refers to the development of a psychological and emotional bond between an individual and their environmental setting (Altman & Low, 1992; Lewicka, 2011; Scannell & Gifford, 2010; Ujang & Zakariya, 2015). Developing a sense of attachment to the residential area in which one lives is beneficial for individuals and communities. Individuals’ localized social ties and neighboring experiences are integral for the development of place attachment (Lewicka, 2005, 2010, 2011). Neighboring experiences in the residential environment may be particularly important for residents of disadvantaged and disorderly neighborhoods where the physical and economic characteristics alone are less likely to facilitate positive place attachment (Cattell & Evans, 1999; Forrest & Kearns, 1999; Letki, 2008; Manzo, 2005; Scannell & Gifford, 2010; Wood & Vamplew, 1999).
Place-based research demonstrates the influence of local stressors, at the neighborhood level, such as crime, disorder, and disadvantage on the development of social ties and relationships with neighbors (Swaroop & Morenoff, 2006; Woldoff, 2002). While scholars have shown that positive experiences in the neighborhood are central to the ongoing development of the people–place bonds that form the basis of place attachment (Scannell & Gifford, 2016), whether or not negative experiences erode place attachment remains unclear. The aim of this study is to explore the relationship between crime victimization in the residential neighborhood and place attachment in a sample of neighborhoods in Brisbane, Australia. It will extend the scope of existing research by examining how negative experiences in the residential neighborhood, measured using self-reported experiences of personal crime victimization, influence residents’ attachment to place and whether this relationship is moderated by neighborhood social ties and/or social and functional neighboring behaviors.
The importance of place attachment for both individuals and communities cannot be overstated. Individuals who report higher levels of place attachment tend to experience better physical and mental health outcomes (Anton & Lawrence, 2014; Tartaglia, 2013). The sense of familiarity and rootedness that coincide with place attachment also results in greater feelings of comfort, confidence, and safety (Lewicka, 2011). Studies have consistently found that place attachment can facilitate a sense of belonging generated through shared experiences and social ties between neighbors (Jack, 2008; Lewicka, 2005, 2010, 2011). When residents have social ties with their neighbors, they express higher levels of place attachment. At the community level, benefits from high levels of social capital and place attachment include greater collective efficacy and capacity for informal social control (Brown, Perkins, & Brown, 2003; Li & McKercher, 2016; Manzo & Perkins, 2006). Place attachment can aid and enhance pre-disaster preparedness and post-disaster resilience by encouraging residents to return to the affected neighborhood (Cheshire, 2015; Scannell, Cox, Fletcher, & Heykoop, 2016). Furthermore, greater place attachment at the community level can impede crime by encouraging residents to intervene in neighborhood incivilities (Brown & Raymond, 2007; Forrest & Kearns, 2001; Górny & Toruńczyk-Ruiz, 2014). In recent decades, contemporary urban neighborhoods have been transformed by changes in transportation, lifestyle, and patterns of everyday living. In particular, there has been notable increases in residential mobility, population diversity, and density (Australian Bureau of Statistics, 2015). These factors can threaten the development of the social processes that underpin place attachment. Given recent changes in urban environments, it is vitally important to understand how to effectively foster individual place attachment which, in turn, can facilitate individual and community safety and well-being.
Conceptualizing Place Attachment
Place attachment was first conceptualized by Kasarda and Janowitz (1974), who described place attachment as sentiments toward the community driven by the desire to avoid or overcome problematic features in the neighborhood. Place attachment is facilitated and governed by a temporal process through which cognitive and emotional elements of attachments to place can be evoked and nourished (Morgan, 2010; Riger & Lavrakas, 1981). Since its original inception in the 1970s, the concept of place attachment has experienced a number of reformulations; various terms have been introduced to capture place attachment, its components, and similar social phenomena including community attachment (Kasarda & Janowitz, 1974), sense of community (Sarason, 1974), place identity (Proshansky, 1978), and place dependence (Stokols & Shumaker, 1981).
A common perspective, and the one we adopt in this study, is that place attachment represents an affective link between people and specific places. Place attachment captures “feelings, thoughts and behaviors that demonstrate commitment to one’s local area” (Woldoff, 2002, p. 87). This multidimensional perspective suggests that place attachment develops in the presence of both affective/emotional connections and active/behavioral connections to a specific place (Raymond, Brown, & Weber, 2010; Woldoff, 2002). Adding to the complexity of this concept, these connections form in settings that vary in numerous ways—size, scope, tangibility, and familiarity. Furthermore, attitudes about place that inform attachments can be developed at the individual or community level and fluctuate over time (Altman & Low, 1992). The complexity of place attachment is reflected in the diversity of both theoretical perspectives and operations applied to measure it (Kyle, Graefe, & Manning, 2005).
Some scholars refer to two distinct types of place connections using the terms place identity and place dependence (Brown & Raymond, 2007; Kyle et al., 2005; Williams & Vaske, 2003). At its core, place identity contains an environmental past that is shared by people who have developed attachment and a sense of belonging to specific places (Proshansky, 1983). Shared “experience-in-place” creates meaning, validates a stable sense of self, and builds the connections to place that underpin place attachment in the form of memories, milestones, and personal growth (Manzo, 2005, p. 74). Place dependence refers to the importance of a place for an individual based on its functional value or its capacity to facilitate goal achievement (Jorgensen & Stedman, 2001). When conceptualized in this way, place dependence can be developed through active, functional connections to people in places that support specific outcomes, for example, interactions with neighbors to borrow milk or ask childcare advice (Raymond et al., 2010). In combination, place identity and place dependence facilitate place attachment; indeed, Altman and Low (1992) state place attachment is the product of the interplay between affect, emotions, and behaviors in reference to a place (Altman and Low, 1992, p. 5; Kyle, Mowen, & Tarrant, 2004).
This articulation of place attachment resonates with the three-pole “self-other-environment” model proposed by Gustafson (2001). Gustafson (2001) asserts that place attachment is a dynamic, active process whereby social connections within the local neighborhood (other pole) act as an intermediary between individuals (self pole) and the physical environment (environment pole). Social interactions within the physical environment influence memories and emotions, which in turn facilitate place attachment and influence future neighboring experiences. When neighboring experiences are positive, this feedback loop between locally situated interactions and emotions reproduces positive feelings of place and strengthens place attachments over time (Gustafson, 2001; Raymond et al., 2010).
Place Attachment: The Residential Neighborhood Level
Although place attachment can develop at a range of spatial scales, from an individual house to a nation (Altman & Low, 1992), the majority of scholarly attention has been given to understanding place attachment in relation to the residential neighborhood (Hidalgo & Hernandez, 2001; Lewicka, 2010). At this spatial scale, individual characteristics and neighborhood factors come together to shape neighborhood connections and place attachment (Gustafson, 2001). Generally speaking, positive neighborhood characteristics such as well-maintained public spaces and the presence of community/volunteer groups facilitate the dynamic and reciprocal relationships between an individual’s affective and functional connections within the neighborhood and place attachment (Brown, Perkins, & Brown, 2004; Lewicka, 2010). Alternately, negative neighborhood conditions such as economic disadvantage, crime, and disorder tend to impede the development of an individual’s connections within the neighborhood and, subsequently, place attachment (Brown et al., 2003, 2004; Lewicka, 2010). The extent to which positive and negative neighborhood factors influence place attachment depends on an individual’s own experiences as a local resident. For example, greater sense of security, measured as objective or subjective criminal history of an area, positively predicts place attachment (Brown et al., 2003, 2004; Lewicka, 2010; Mesch & Manor, 1998). However, Dekker (2007) observed that even in neighborhoods where crime was high, when residents possessed dense social networks and frequently participated in neighborhood social processes, they tended to exhibit stronger attachment and pride in their local area and were more likely to intervene when others engaged in antisocial or criminal behavior (Pattillo, 1998). In this sense, having social ties and interacting with neighbors ameliorated the impact of neighborhood crime on people–place bonds by empowering residents to take action in the event of crime or deviance.
Although studies have reported the potential for social ties and neighboring to offset the impacts of crime and deviance at the neighborhood level (Dekker, 2007; Takagi, Ikeda, Kobayashi, Harihara, & Kawachi, 2016), little is known about the effect of personal experiences of victimization in the residential neighborhood on individual place attachment. To date, only one study that we know of has examined the impact of personal victimization, an example of a negative neighborhood experience at the individual level, on place attachment. The study was conducted in a sample of distressed neighborhoods in Nashville, Tennessee, and focused on low-income, African American and renting households. The results found that residents in neighborhoods with high levels of physical and social disorder reported lower place attachment, but there was no additional effect for residents who experienced crime directly (i.e., as the victim) (Woldoff, 2002). These results alone cannot be considered conclusive given the very specific focus of the research on racial minority individuals within distressed neighborhoods.
The Current Study
Place attachment is associated with a number of benefits for individuals and communities. Low levels of place attachment can decrease perceptions of safety, facilitate feelings of alienation and isolation, and negatively impact individual well-being (Batson & Monnat, 2015; Bond et al., 2007; Scannell & Gifford, 2016; Wang & Peck, 2013). Feelings of attachment to the residential neighborhood develop in the presence of social connections and neighboring experiences that influence place-based memories, emotions, and behaviors (Raymond et al., 2010).
In the current study, we assess the impact of personal victimization experience on place attachment across a diverse sample of neighborhood settings. We extend previous research to explore whether localized social ties with family, friends, or acquaintances (place identity) and/or neighboring behaviors (place dependence) can buffer individuals from the detrimental impacts of negative experiences such as personal victimization on place attachment (Bruinsma, Pauwels, Weerman, & Bernasco, 2013; Forrest & Kearns, 2001; Hur & Morrow-Jones, 2008; Sampson & Groves, 1989; Sampson, Raudenbush, & Earls, 1997; Takagi et al., 2016; Wickes, Hipp, Sargeant, & Mazerolle, 2017).
Specifically, the study addresses two research questions:
In line with previous research that demonstrates the association between sense of security and/or neighborhood crime level and place attachment, we hypothesize that individuals who have experienced personal crime victimization in the local area will report lower place attachment. However, we expect that having more neighborhood-based social ties (place identity) and engaging in more frequent neighboring behaviors (place dependence) will ameliorate the detrimental effects of victimization on place attachment.
Method
To investigate the relationship between place attachment, social ties, neighboring, and victimization in Brisbane neighborhoods, we draw on Wave 2 of the Australian Community Capacity Study (ACCS). The ACCS is a longitudinal body of research funded by the Australian Research Council (ARC). The overarching goal of the ACCS is to investigate and understand whether and how neighborhood contexts and social processes explain variations in crime and disorder over time across Australian neighborhoods. The ACCS uses a nested sampling design that allows for the computation of ecologically valid measures that reliably capture neighborhood-level social processes. To control for the effects of neighborhood sociodemographic characteristics, we incorporate data from the Australian Bureau of Statistics (ABS) census and Queensland Police Service (QPS) crime incident data. Please see the ACCS website for further details (https://accs.project.uq.edu.au/).
Participants
Wave 2 of the ACCS comprises a sample of 4,249 Brisbane residents living in 147 neighborhoods. The original sample size was determined by using the number of residents needed to maintain econometric reliability based on power analyses from Optimal Design Software for multilevel samples (Spybrook et al., 2011). Participants were randomly selected using a random digit dialing method. Data collection took place using computer-assisted telephone interviewing (CATI) in August 2008. The response rate was 43%. To be eligible for the study, participants needed to be 18 years or older and reside in private dwellings with telephones in the selected Brisbane neighborhoods. Descriptive statistics are provided in Table 1.
Summary Statistics of Analytic Sample.
Note. ACCS = Australian Community Capacity Study; ABS = Australian Bureau of Statistics.
Research Site
Brisbane is the captial city of Queensland, Australia. At the 2006 census, Brisbane had a population of approximately 1.9-million people living in 429 neighborhoods. In this study, we use the term neighborhood to refer to the state suburb. 1 Unlike census tracts in the U.S. context, Brisbane neighborhoods are not demarcated by population or household size. Thus, neighborhoods vary in size with populations ranging from 15 to 21,001 2 (ABS, 2006). In 2006, 28% of the Brisbane population was born overseas and nearly 13% spoke a language other than English at home. Spatial clustering in persons born overseas, non-English speakers, and economic disadvantage is evident at the neighborhood level. Rates of crime were relatively low across Brisbane neighborhoods with an annual average (2006; 2007; 2008) of 52.58 offenses per 1,000 residents. The sample of 147 neighborhoods captured in the ACCS is representative of residential neighborhoods in Brisbane.
Dependent Variables
We conceptualize place attachment as the “feelings, thoughts and behaviors that demonstrate commitment to one’s local area” (Woldoff, 2002, p. 87). Our measure of place attachment comprises a scale of four items that together capture both the affective/attitudinal and the behavioral (problem-solving) components of place attachment described in the literature (Raymond et al., 2010; Woldoff, 2002). To measure place attachment, participants in Wave 2 of the ACCS were asked to report on a 5-point scale ranging from 1 “strongly disagree” to 5 “strongly agree” their level of agreement with four statements: “I feel that I belong to this local community; “I am, proud to live in this local community”; “I would like to be living here in three years time”; and “I feel a responsibility to make a contribution to the local community I live in” (items listed in Appendix A). The first 3 items represent attitudinal attachment while the final item measures behavioral attachment, more specifcally behavioral attachment through informal problem-solving as described by Woldoff (2002). This place attachment scale was originally developed by Christakopoulou, Dawson, and Gari (2001; Mazerolle et al., 2007) for use in the Spatial Economics Research Centre (SERC) social capital study. 3 It has high internal consistency (α = .78). The test–retest relability between Wave 1 and Wave 2 (n = 1,003) of the ACCS was .50. 4
Independent Variables
Victimization
To address the central aim of this study, which is to examine how negative experiences in the residential neighborhood influence place attachment, we used a variable from the ACCS capturing self-reported crime victimization. Respondents were asked if they (and their household members) had ever been assaulted or mugged in their community, if their home had ever been broken into, and if they had experienced property and/or vehicle damage (items listed in Appendix B). Due to the low number of persons reporting victimization in each of the three categories, responses across the three items were combined to compute a dichotomous variable indicating victimization in the local neighborhood.
Neighborhood social ties
Research shows that social ties within the residential neighborhood positively predict place attachment (Lewicka, 2010). Specifically, research reports that as social ties within the setting increase so too does emotional attachment to place (Raymond et al., 2010). Studies employ various measures of social ties, including number of friends and relatives in the neighborhood (Kasarda & Janowitz, 1974; Mesch & Manor, 1998) and less intimate ties with fellow residents (Kasarda & Janowitz, 1974; Lewicka, 2005; Moser, Ratiu, & Fleury-Bahi, 2002). Drawing on this literature, we captured neighborhood social ties using two items from the ACCS measuring total number of localized family and friendship ties and the proportion of known residents, which we conceptualize as acquaintanceship ties. These are included as two separate variables in the regression models. We adopted this approach as previous studies have shown that family/friend ties and acquaintanceship ties differentially impact social processes and neighborhood experiences (Takagi et al., 2016). The items were drawn from the Project on Human Development in Chicago Neighborhoods Survey (PHDCN). Studies have demonstrated the validity of these measures for capturing social ties in the neighborhood context (Sampson, 2012; Sampson & Groves, 1989; Sampson, Raudenbush, & Earls, 1997).
Neighboring behavior
Neighboring behaviors range from relatively rare and impersonal interactions, such as incidential meet and greets in the street, to more regular, intimate interactions that may involve doing favors, asking advice, and developing reciprocated trust with neighbors over time. It is these more regular, functional interactions that are purported to be important facilitaters of the place dependence component of place attachment (Lewicka, 2011; Raymond et al., 2010). To capture neighboring behavior, we created a six-item scale from the Wave 2 ACCS. Respondents were asked to report how often they, and the people in their community, do a range of favors for each other and activities together such as the following: ask advice about job openings or child rearing, look after each other’s properties, or visit in each other’s homes (items listed in Appendix C). Responses were measured on a 4-point Likert-type scale ranging from 1 (never) to 4 (often). Adapted from the PHDCN, the validity of this scale is well established in the literature (Sampson, 2012; Sampson & Groves, 1989; Sampson, Raudenbush & Earls, 2004). It has high internal consistency (α = .820) and reflects measures of neighbor contact found to predict place attachment in previous studies (Lewicka, 2011).
Individual-level controls
As place attachment is a manifestation of the affective connection that people form toward place over time via repeated interactions with people and place (Gustafson, 2001; Manzo, 2005; Scannell & Gifford, 2010), individual demographic characteristics that influence time spent in the neighborhood and opportunities to interact with neighbors would be expected to also affect place attachment (B. Brown et al., 2003; Dallago et al., 2009). To account for the influence of individual characteristics, we included the following demographic variables in our analytic models: gender, relationship status, presence of children within the household, employment, home ownership, duration of residence, and country of birth. Further details of all variables and summary statistics are provided in Table 1.
Neighborhood-level variables
To control for the influence of neighborhood context (Bailey, Kearns, & Livingston, 2012; Lancee & Dronkers, 2011), we included measures of neighborhood disadvantage, residential mobility, ethnic composition, population density, and crime in our analytic models. These measures were drawn from ABS census and QPS crime incident data. All measures are computed at the level of the residential state suburb.
Crime
Drawing from the QPS data, we calculated the average, annual neighborhood total crime rate between 2006 and 2008. Crime rate is calculated by dividing total number of reported incidents by the neighborhood population and multiplying by 1,000 to achieve a neighborhood crime rate per 1,000 people.
Neighborhood disadvantage
Neighborhood disadvantage is a factor score computed from percentage low-income households, percentage of individuals with less than high school–level education, and percentage lone-parent households (α = .70).
Language diversity
To capture ethnic diversity at the neighborhood level, we computed a Blau index of language diversity (Blau, 1977). Language is strongly tied to ethnicity (Anderson, 2006) and nationhood (Calhoun, 1993). As an observable cue of social distance (Wickes, Hipp, Zahnow, & Mazerolle, 2013), speaking a language other than the official national language may hinder neighborhood connections and communication (Leigh, 2006). The Blau index, also referred to as the Herfindhal index (Harrison & Klein, 2007), is scaled between 0 and 1 where 0 denotes a completely homogeneous neighborhood and 1 indicates high ethnic diversity. Our index uses nine language regions represented in ABS census data including Northern, Southern, and Eastern European languages; South West Central, Southern Asian, South East and Eastern Asian languages; Australian Indigenous; other languages; and speaks English only.
Residential mobility
We include a measure of the proportion of residents living at a different address 5 years prior the census to control for residential mobility at the neighborhood level.
Population density
Living in populated areas may be associated with a greater number of ties and contact with neighbors due to increased opportunities for interacting and bonding (Hipp, Corcoran, Wickes, & Li, 2014). Also from the ABS, this variable captures number of persons per hectare.
Analytic Strategy
We estimate five multilevel linear models that account for the nested structure of the data. Below is the generalized expression of a multilevel linear regression model. Yij is the place attachment score for individuals i within Brisbane neighborhoods (j) in the following expression:
where µ represents the grand mean for the population, being average place attachment scores from individuals clustered within 146 5 Brisbane neighborhoods; Uj is the random effect of neighborhoods j; and finally, Rij denotes the residual effect or error for individuals i within neighborhood j. Model 1, the base model, provides a preliminary exploration of the relationship between social ties, contact with neighbors, demographic control variables at the individual and neighborhood levels and place attachment. Model 2 includes our key variable of interest, crime victimization experience in the residential neighborhood, and addresses Research Question 1 by assessing the relationship between crime victimization experience and individual place attachment after controlling for demographic and ecological variables. Models 3 to 5 address Research Question 2. Here we incorporate variables capturing interactions between victimization and (a) family and friendship ties, (b) acquaintanceship ties, and (c) neighboring behavior, to assess whether or not the impact of victimization on place attachment varies dependent on an individual’s social ties and social and functional interactions in the residential neighborhood.
Results
The null model revealed 7.6% of the variation in individual place attachment is at the neighborhood level. Model 1 examines the relationship between social ties, neighboring behavior, demographic factors and place attachment. The results demonstrate positive relationships between neighborhood family/friend ties (β = .035, p < .001), neighborhood acquaintanceship ties (ref: know no neighbors; know a few neighbors, β = .367, p < .001; know many neighbors, β = .498, p < .001; know most neighbors, β = .566, p < .001), neighboring behavior (β = .261, p < .001), and place attachment. A number of individual and contextual variables are also significant predictors of place attachment. While age is positively associated with place attachment (β = .005, p < .001), male residents (β = –.074, p < .001), single residents (β = –.058, p < .01), and individuals either on a pension (β = –.129, p < .001) or home duties (β = –.077, p < .01) report lower levels of place attachment than females, married residents, and employed persons respectively. Neighborhood disadvantage (β = –.103, p < .001) is also negatively associated with individual place attachment.
Model 2 includes our indicator of crime victimization experience to examine its effect on place attachment when controlling for neighborhood social ties, neighboring behavior, and other individual and contextual factors. The results reveal a significant, negative association between victimization and place attachment (β = –.090, p < .001). Specifically, individuals who have experienced victimization report levels of place attachment that are approximately 9% lower than those who have not been victiimized. This demonstrates that even after controlling for neighborhood social ties, neighboring behavior, demographic and contextual factors, the negative experience of crime victimization in the residential neighborhood is detrimental for individual place attachment. Consistent with Model 1, neighborhood family/friend ties (β = .035, p < .001), neighborhood acquaintanceship ties (ref: know no neighbors; know a few neighbors, β = .366, p < .001; know many neighbors, β = .496, p < .001; know most neighbors, β = .559, p < .001), and neighboring behavior postively predict place attachment (β = .259, p < .001). The effects of individual and contextual variables on place attachment are also consistent with Model 1 (see Table 2).
Predictors of Place Attachment.
Note. AU = Australia; ICC = intraclass correlation coefficient; BIC = Bayesian information criteria; AIC = Akaike’s information criteria.
p < .05. **p < .01. ***p < .001, unstandardized coefficients.
Models 3 to 5 examine the moderating effects of neighborhood family/friend social ties, neighborhood acquaintanceship ties, and neighboring behavior on the relationship between victimization and place attachment through the inclusion of variables capturing the interaction between each of the three measures and our indicator of victimization experience. Models 3 and 4 test whether the effect of victimization on place attachment varies dependent on the number of social ties people have in the neighborhood with family/friends and acquaintances respectively. Neither of these interactions were significant at the .05 level. Model 5 revealed a significant moderating effect of neighboring behavior on the relationship between victimization and place attachment. Specifically, engaging with neighbors socially and/or for functional purposes, such as to ask advice about child rearing or to keep watch of each other’s homes, appears to buffer individuals from the negative effects of personal crime victimization on place attachment (β = .103, p < .001) (see Table 3 and Figure 1).
Moderating Effect of Neighborhood Family/Friend Ties, Acquaintanceship Ties, and Neighboring.
Note. AU = Australia; ICC = intraclass correlation coefficient; BIC = Bayesian information criteria; AIC = Akaike’s information criteria.
p < .05. **p < .01. ***p < .001, unstandardized coefficients.
Moderating effect of neighboring on relationship between victimization and place attachment.
Discussion
Place attachment is dynamic, multifaceted, and develops from “a potpourri of memories, conceptions, interpretations, ideas and related feelings” (Proshansky, Fabian, & Kaminoff, 1983, p. 60). Feelings of attachment emerge through localized ties and neighboring experiences, both positive and negative, that shape an individual’s place-based memories, emotions, and future behavioral intentions. The aim of this study was to extend current understandings of place attachment by examining the impact of personal victimization experience in the residential neighborhood on place attachment and exploring whether neighborhood social ties and/or neighboring behaviors could ameliorate the detrimental impacts of the negative experience of victimization on place attachment. Scholarship demonstrates that crime and disorder at the neighborhood level can undermine the development of positive connections to place (Dekker, 2007; Lewicka, 2010; Pattillo, 1998). However, the impact of directly experiencing crime, as a victim, on place attachment remains relatively unexplored. We applied a multilevel modeling strategy incorporating survey, census, and crime data, to examine how person–place bonds are influenced by negative experiences (i.e., individual crime victimization) in the residential neighborhood and whether this relationship can be moderated by positive experiences (i.e., neighbor contact and social ties).
Our study indicates several important findings about place attachment: First, as hypothesized, individuals who experience crime victimization in the residential neighborhood report lower levels of place attachment. We found that personal crime victimization impacts place attachment above and beyond neighborhood-level crime rate and other neighborhood contextual factors associated with social disorganization and disorder such as economic disadvantage, residential instability, and population density. Our results also demonstrated that residents were buffered from the negative effects of crime victimization on place attachment by their engagement in social and functional neighboring behaviors. There was no evidence that the mere presence of neighborhood social ties engendered the same moderating influence; neither neighborhood family/friend ties nor neighborhood acquaintancship ties served to offset the detrimental impact of victimization experience on place attachment.
Although the results only partially support our hypotheses, they align with two key bodies of research that highlight the nuanced difference between the presence of neighborhood social ties (i.e., knowing people) and experiencing the process of neighboring by activating social ties (Sampson & Raudenbush, 1999; Scannell & Gifford, 2010). A number of place attachment scholars highlight the importance of experiences in place for the development of memories and emotions that facilitate affective place connections. For example, Scannell and Gifford (2010) articulate the people–process–place components in their attachment framework. This model of place attachment emphasizes the vital, intermediary role of processes in linking people and place through place-based interactions. Multidimensional models that consider place attachment to be a product of the interplay between place identity and place dependence iterate this sentiment (Altman & Low, 1992; Raymond et al., 2010). Our finding—that engagement in neighboring behaviors offsets the detrimental impacts of victimization experience but neighborhood social ties per se do not—resonates with this perspective. A direct negative experience in the residential neighborhood, operationalized here as crime victimization, is embedded in a person’s place-based memory and subsequently informs feelings and emotions toward place. Social and functional neighboring behaviors, such as spending time in each other’s homes and asking each other for advice, also constitute place-based experiences, albeit of a positive nature. These experiences generate positive memories and emotions connected to place. Alternately, knowing many neighbors or having neighborhood social ties—the mere existance of these things—does not necessarily generate active experiences in the neighborhood that will become embedded in one’s place-based memories (Proshansky et al., 1983).
The delineation between the presence of social ties and the active experience of connecting through social ties was first made explicit by Sampson and Raudenbush (1999) who asserted the need to clearly differentiate “the process of activating/converting social ties to achieve the desired outcomes from the ties themselves” (p. 635). Collective efficacy theory builds on the systematic model of neigborhood social disorganization (Bursik & Grasmick, 1993; Shaw & McKay, 1942) to explain variations in neighborhood crime and disorder as a function of the capacity for residents to generate shared expectations for acceptable public behavior and activate social ties to engage in informal social control (Sampson & Raudenbush, 1999; Sampson et al., 1997). Empirical evidence to support collective efficacy theory finds that neighborhoods with higher collective efficacy have lower levels of crime and residents of such neighborhoods feel safer (Gibson, Zhao, Lovrich, & Gaffney, 2002; Hinkle, 2013; Morenoff, Sampson, & Raudenbush, 2001; Yuan & McNeeley, 2017; Zhang, Messner, & Liu, 2007). This is because the activation of neighborhood social ties, through contact with fellow residents, empowers individuals to intervene in the event of a crime by facilitating shared expectations that all neighborhood residents will do something when a problem arises (Collins & Guidry, 2018; Sampson, 2012). Our results reflect this line of argument, albeit at the individual level. We suggest frequent contact with neighbors may serve to enhance feelings of security in the neighborhood by reinforcing positive place-based experiences and shared expectations for safe social behavior in the neighborhood.
Conclusion
The aim of this study was to assess the effect of individual-level crime victimization on place attachment. Findings highlight that people who have been victimized display lower feelings of attachment. This is consistent with both place and criminological research that demonstrates people’s perceptions, attitudes, and behaviors can be tarnished by victimization experiences. Although our results suggest crime victimization is associated with lower place attachment, engaging in neighboring behaviors counteracted these experiences, demonstrating the interconnected and complex evolution of place attachment and the development of person–process–place dimensions (Scannell & Gifford, 2010).
First acknowledged at the turn of the 20th century (Shaw & McKay, 1942), the importance of community participation and localized social ties for community regulation and well-being has recently returned to the research and policy agenda of scholars and government bodies (B. Brown et al., 2003; Swaroop & Morenoff, 2006; Wickes et al., 2017). While increasingly used as a discursive tool by local government, place attachment tends to be the focus of short-term sporadic community projects that do little for residents’ long-term empowerment. Often in disadvantaged neighborhoods it is assumed that poor neighborhood conditions undermine the capacity for place attachment and that residents would leave the area if it were financially viable; for this reason, these types of communities are rarely the target of community projects that aim to engender positive place identity, place dependence, and attachment. Here we demonstrate that in addition to objective neighborhood conditions, experiences in the residential neighborhood, whether positive neighboring behaviors or negative victimization experiences, are important for predicting place attachment. The findings suggest that the benefits of place attachment can be accrued in any type of neighborhood so long as individuals interact with other residents in the local area (Lancee & Dronkers, 2011; Skogan, 1990) and that engaging in positive neighboring experiences can help to neutralize the effect of negative experiences on place attachment. Encouraging social interactions through community engagement projects and providing open public spaces to support incidential interactions between neighbors is integral to establishing authentic place bonds across neighborhood contexts.
Although this study adds to our understanding of place attachment as a multilayered, multilevel construct by drawing on a large data set of Brisbane residents living in 147 neighborhoods, it is not without limitations. First, we acknowledge that Brisbane is a relatively low-crime city when compared with its international counterparts in the United States and the United Kingdom. Therefore, more research is required before the study findings can be generalized to high-crime and population-dense cities in other national contexts. Also a recent study by Hipp and Wickes (2018) found evidence to suggest that people update their perceptions of the neighborhood over time. Given the cross-sectional nature of this study, we are unable to examine the temporality of the development of people–process–place bonds (Altman & Low, 1992; Hidalgo & Hernandez, 2001; Riger & Lavrakas, 1981; Scannell & Gifford, 2010). Furthermore, we focus here on neighborhood place attachment yet are unable to account for residents’ attachment to their “home” environment; previous research demonstrates that respondents’ level of attachment may be entangled with feelings toward the home rather than solely the neighborhood (Lewicka, 2011; Manzo, 2005). There is also evidence that attachment to the home may be stronger than neighborhood place attachment when the two are considered in isolation (Manzo, 2005). Given this, future research that differentiates between attachment to the home and attachment to the neighborhood and that considers the relative importance of predictors of each is necessary for a holistic understanding of place attachment. Relatedly, we acknowledge the modifiable areal unit problem (MAUP), ubiquitous in neighborhood research, and caution against generalizing the findings of this study to other spatial units such as micro-neighborhoods or street blocks. Finally, we acknowledge the age of the data (collected in 2008) may be seen as potentially limiting given the large societal changes occuring over the last decade. However, longitudinal analyses using this (ACCS Wave 2) and additional waves of the ACCS collected in 2010 and 2012 demonstrate patterns of social processes, neighboring, demographics, and crime have remained remarkably stable over time in Brisbane neighborhoods (Wickes, Britt, & Broidy, 2017). Despite these limitations, our results contribute to an under-explored area of place attachment by interrogating whether negative experiences, specifically direct crime victimization in the residential neighborhood, can erode people–place bonds and how engaging in neighboring behaviors can serve as a buffer to minimize this influence.
Footnotes
Appendix A
Place Attachment Survey Items.
| Strongly disagree | Disagree | Agree | Strongly agree | Don’t know | Refuse | |
|---|---|---|---|---|---|---|
| 1. I feel that I belong to this local community. | 1 | 2 | 3 | 4 | 5 | 6 |
| 2. I would like to be living in this local community in 3 years time. | 1 | 2 | 3 | 4 | 5 | 6 |
| 3. I am proud to live in this local community. | 1 | 2 | 3 | 4 | 5 | 6 |
| 4. I feel a responsibility to make a contribution to the local community I live in. | 1 | 2 | 3 | 4 | 5 | 6 |
Appendix B
Question prelude: The next section asks about victimization that may have happened in your community, to yourself or to members of your household.
Appendix C
Neighboring Behaviors Survey Items.
| Never | Rarely | Sometimes | Often | Don’t Know | Refused | |
|---|---|---|---|---|---|---|
| 1. About how often do you and people in your community do favors for each other? | 1 | 2 | 3 | 4 | 5 | 6 |
| 2. When a neighbor is not at home, how often do you and other neighbors watch over their property? | 1 | 2 | 3 | 4 | 5 | 6 |
| 3. About how often do you and people in your community ask each other advice about things such as child rearing or job openings? | 1 | 2 | 3 | 4 | 5 | 6 |
| 4. About how often do you and people in your community visit in each other’s homes or on the street? | 1 | 2 | 3 | 4 | 5 | 6 |
| 5. About how often do you and people in your community have parties or other get togethers? | 1 | 2 | 3 | 4 | 5 | 6 |
| 6. About how often do you and people in your community spend leisure time together going out for dinner, to the movies, to a sporting event, etc.? | 1 | 2 | 3 | 4 | 5 | 6 |
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) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Wave 2 of the Australian Community Capacity Study was funded by the Australian Research Council (DP0771785).
