Neighborhood Crime Control in a Changing China

Abstract

Objectives:

This research assesses the impacts of three distinctive crime control activities organized and directed by the neighborhood committees with the assistance of local police in contemporary urban China—Tiao-jie, Bang-jiao, and neighborhood watches. Tiao-jie deals with disputes and minor criminal cases. Bang-jiao provides guidance to residents who have committed minor offenses or have been released from correctional institutions to facilitate reintegration. Neighborhood watches engage local residents in crime prevention under the direction of neighborhood committees.

Method:

Using survey data recently collected in Tianjin, we examine the effects of indicators of the implementation of these neighborhood-based crime control strategies on residents’ reports of household property victimizations that occurred within the neighborhoods with multilevel logistic regressions.

Results:

Net of a range of individual-level and neighborhood-level control variables, the indicators of the level of activity of Tiao-jie, Bang-jiao, and neighborhood watches exhibit negative effects on reported household property victimization.

Conclusions:

Our findings provide suggestive evidence that the traditional strategies of neighborhood-level crime control continue to be relevant in the China of today and that the role of collective efficacy appears to differ from that observed in Western cities.

Keywords

Tiao-jie Bang-jiao crime urban China

One of the core problematics in sociological criminology over the course of the past century or so has been trying to explain the ecological patterning of levels of crime in urban settings. The basic question that motivates such inquiry can be stated quite simply, “What makes some neighborhoods safer from crime than others?” (Warner and Glubb 2013:333).

The intellectual foundations for these efforts can be traced to the pioneering work on social disorganization associated with the classical Chicago School (Kornhauser 1978; Shaw and McKay 1942). The Chicago School researchers documented appreciable differences in levels of crime across neighborhoods in the city that proved to be remarkably durable over time. Moreover, they recognized that in highly industrialized and urbanized societies, the official organizations for law enforcement have at best a limited capacity to control crime. Social disorganization theory accordingly directs attention to the scope and quality of the relationships among the private residents of different neighborhoods and to their capacity to mobilize and work together to realize collective goals—including the goal of securing a safe and secure public environment (Bursick and Grasmick 1993). Despite the theoretical and methodological criticisms that have been leveled at the Chicago School over the years (see, e.g., Bursik 1988), its distinctive insight has endured. A large and growing body of research on “neighborhood effects” indicates that the local community plays an important role in determining how safe or unsafe an urban neighborhood is likely to be (Messner and Zimmerman 2012; Sampson, Morenoff, and Gannon-Rowley 2002).

Community dynamics also play an important role in crime control in urban China, but the organizational context for such community processes differs profoundly from that in the West. In China, the state has created formal, officially recognized organizations of urban residents—the neighborhood committees (jü-wei-hui), sometimes translated as “residents’ committees.” These neighborhood committees encompass territories with explicitly defined boundaries, and they sponsor and direct a wide variety of activities.

The most important of these activities for present purposes are the crime control programs of Tiao-jie, Bang-jiao, and neighborhood watches (described below). These activities are organized and directed by the neighborhood committees with the assistance of local police agencies, and they have been cited as major factors accounting for low crime rates in urban China during the reign of Mao (Tang and Parish 2000; Whyte and Parish 1984; Zhang et al. 1996; Zhang, Messner, and Liu 2007). As Whyte and Parish (1984) and Read (2012) observed, these neighborhood programs were part of total social control under Mao, and their primary purpose was not crime prevention, but rather political control. Effective crime control became a by-product of the pervasive political control.

Following the economic reforms since the early 1980s, the ecology of urban neighborhoods in China has changed dramatically. The severe restrictions on population mobility have been relaxed allowing large numbers of rural migrants to settle in the cities (Liang, Li, and Ma 2014). An economic elite has emerged along with a poverty stratum (Bian 2002), and private ownership of housing has greatly expanded, along with the private sphere of life more generally (Zhong 2009). Crime rates, and property crimes in particular, have increased dramatically (Liu, Zhang, and Messner 2001; Zhang, Messner, and Liu 2008). Most of these crimes have been attributed by Chinese scholars to the influx of the large “floating” populations of migrants and income inequalities (Shi and Wu 2010; D. Wang et al. 2007; G. Wang and Liu 2006; Z. Wang 2006). Neighborhood committees have increasingly been “in the process of adaptation” struggling to remain relevant in the face of the general decline in communist ideology and state control over individual behavior, growing anonymity among urban dwellers, and commercial orientation in social interactions (Zhong 2009:219-21).¹ Compared to the Mao era, when neighborhood committees often represented the frontline of political campaigns and government control of urban Chinese residents, their influence is much subdued and concerned to a large extent with providing social services.

The purpose of the present study is to assess the impacts of the crime control activities implemented by the neighborhood committees in contemporary urban China. Using survey data recently collected in the city of Tianjin, we examine the effects of indicators of the implementation of Tiao-jie, Bang-jiao, and neighborhood watches on residents’ reports of property victimizations that occurred within the neighborhoods. We incorporate in our models controls for the kinds of variables that are commonly considered in the Western research on neighborhood effects (Sampson et al. 2002), that is, including measures of informal social control (collective efficacy) and public control (policing). The overarching question guiding our inquiry is as follows: To what extent do the traditional crime control strategies employed by neighborhood committees continue to play an appreciable role in contemporary China, a society undergoing profound transformation?

Neighborhood Committees and Crime Control in Urban China

A distinctive feature of social control in China continues to be its fusion of formal and informal mechanisms as reflected in “the ‘mass line’ and ‘relying on the masses’” (see Zhang et al. 1996:204; Zhong 2009:111, for discussions of the “mass line” strategy of social control in China). The neighborhood committees are the quintessential organizational manifestations of the fusion of formal and informal social control at the local level in urban China. Read (2012:3; see also Walder 2015) referred to such organizations as systems of “administrative grassroots engagement.” The leaders of these organizations serve as liaisons between the state and the residents of the city. The state, via the City–Street Office (the lowest branch of a municipal government), creates, sponsors, and directs these organizations. At the same time, members of neighborhood committees are not formal government employees, although they receive compensation. They are members of the community.

The neighborhood committees collaborate with the local police station (pai-chu-suo) to implement the community-based crime prevention programs of Tiao-jie and Bang-jiao. Tiao-jie refers literally to mediating and solving disputes and conflicts among residents, sometimes family members. A basic form of Tiao-jie is the Tiao-jie committee, which is a subcommittee of the neighborhood committee. Senior, reputable residents are often invited to participate in Tiao-jie. A Tiao-jie committee normally deals with noncriminal cases (e.g., housing problems, marital disputes, and uncollected small debts) and minor criminal cases. The committee’s goal is to resolve these cases in an informal community setting to forestall the development of more serious grievances.

Bang-jiao is intended to provide assistance and guidance to residents who have committed minor offenses or have been released from correctional institutions. Traditionally, Bang-jiao has been organized in the form of help groups that consist of the offender’s parents or relatives, a member of the neighborhood committee, and an officer from the neighborhood police station. The primary goal of Bang-jiao is to intervene in the lives of residents, especially young residents, who are at risk of becoming serious or chronic offenders and to reintegrate exoffenders back into the community.

The philosophies behind both Tiao-jie and Bang-jiao can be traced to ancient Confucianism.² Confucius and his followers cultivated a preference for informal means of conflict resolution (Zhang et al. 1996:203; see also Leng and Chiu 1985). Although impossible to determine when Tiao-jie (or mediation) began, Wall and Blum (1991:4) pointed to Confucianism as one of its roots, which perceives harmony as the most desirable state, and if disturbed, the preferred way of restoring social harmony is through compromise or the way of the middle (i.e., zhong-yong). This philosophical tenet highly values compromise and persuasion as well as the intermediary who was able to secure them (i.e., the person who does Tiao-jie; Cohen 1966), thus giving rise to the cultural practice of Tiao-jie. Various forms of mediation (i.e., Tiao-jie) were practiced throughout Chinese dynasties to manage interpersonal conflicts and maintaining social harmony (Cohen 1966; Van Der Sprenkel 1962). Even in Mao’s era, community mediation services were well in place in the communes in rural areas and integrated into neighborhood committees in the cities (Wall and Blum 1991).

Relatedly, Confucianism, patriarchal in nature, places high values on moral education. Bad thoughts that underlie bad behaviors can presumably be changed by authority figures or members of seniority in the community (Zhang et al. 1996:204), hence Bang-jiao (which literally translates “to help through lecturing”). Consequently, while Tiao-jie has largely remained a cultural practice for dispute resolution, Bang-jiao was often used during the Mao era to carry out socialist thought control policies.

Formalized mediation services were severely disrupted during the tumultuous years of the Cultural Revolution and only in recent decades resumed. However, Wall and Blum (1991) found through their literature review and field interviews in 1988 that community mediation continued more or less intact through the Cultural Revolution (1966 to 1977) and into post-Mao legal reform that began in 1978, except that the mediation services were informal and provided by respected people of the community (Wall and Blum 1991:5).

The economic development after Mao’s death allowed the Chinese government to introduce measures to once again elevate the role of the formal institutions of law and criminal justice. The revival and rapid growth in Chinese bar and legal profession also increased access to legal resolution of disputes among individuals and private companies. While civil mediation services have become more established, the number of mediation committees and disputes solved through mediation have declined steadily since the 1990s, relative to those litigated in the court. Formal court proceedings are increasingly accepted by the public as the preferred venue for conflict resolution in China because many disputes nowadays involve business contracts, complex financial arrangements, foreign entities, and personal grievances against government exercise of eminent domain over land use, which are all too challenging for the traditional form of civil mediation (Hong 2011). A third community-based crime control initiative in urban China is the implementation of neighborhood watches (she-qu-xun-luo). Their routine activities are the patrolling of the neighborhood by identifying and questioning strangers, checking the security of residents’ property, and reporting suspicious activities. By name, these efforts might appear to be similar to those implemented in the West, but they have distinctive features. In the West, community-based crime prevention programs reside securely within the domain of civil society, being grounded in voluntary associations of private citizens despite varying degrees of collaboration with local law enforcement. Neighborhood watches in urban China, in contrast, are typically organized, directed, and supervised by the politically authorized neighborhood committees, which coordinate closely with the neighborhood police stations. Similar to Tiao-jie and Bang-jiao, these neighborhood watches are thus characterized by a distinctive fusion of formal and informal means of social control.

It is important to recognize that the differences between China and the West in the blending of informal and formal means of neighborhood crime control are matters of degree. See, for example, Rosenbaum and Schuck’s (2012) analysis of the growing interest in “coordinated community-wide strategies” for crime prevention in the West, which entail partnerships between law enforcement agencies and civic associations, and Carr’s (2003) arguments about the “new parochialism.” Nevertheless, these are important matters of degree. As Zhong (2009:111) observed, the fusion of formal and informal control mechanisms in China typically appears to be “remarkable in the eyes of Western observers.”

Despite the long-standing recognition of the centrality of these community-based strategies for crime control in China, there is very little quantitative evidence about their actual effectiveness. We are unaware of any quantitative analyses of neighborhood watches, and the few studies that have applied quantitative techniques to assess the impacts of Tiao-jie and Bang-jiao have yielded mixed findings. An early study conducted by the Tianjin Academy of Social Sciences in 1991 is suggestive of at least some crime control effects of these programs (Zhang et al. 1996). This study involved surveys of inmates in prisons and reform schools in Tianjin. The inmates were asked questions about the implementation of Tiao-jie and Bang-jiao in the neighborhoods, in which they lived prior to incarceration. The measure of Bang-jiao activity was a dummy variable indicating whether a Bang-jiao group had been established in the inmate’s neighborhood. The Tiao-jie measure was the respondent’s rating of the level of Tiao-jie activity in the neighborhood, classified in three ordered categories: “not at all,” “for some disputes,” and “for all disputes.” The relationships between these responses and the inmates’ reports on the general levels of deviance and crime in their neighborhoods were assessed. Information was also collected on whether the inmate was a first or a repeat offender.

The results revealed a significant negative effect of the Bang-jiao indicator on the measure of deviance—reports of deviance were lower in neighborhoods with an established Bang-jiao group. The Tiao-jie indicator exhibited significant negative effects on the measures of both deviance and crime. In addition, the Bang-jiao indicator exhibited a negative association with reports of repeat offending (the coefficient for the Tiao-jie indicator was nonsignificant for repeat offending). The statistically significant coefficients were thus all consistent with the inference that these traditional crime control strategies do indeed help prevent crime.

More recently, surveys of the general population in two Chinese cities have offered conflicting findings about the impact of Tiao-jie. A household victimization study conducted in Tianjin in 2004 examined the association between an indicator of Tiao-jie activity and household burglary (Zhang et al. 2007). Tiao-jie activity was conceptualized as a form of “semipublic control” and was measured with a dummy variable differentiating respondents who reported that the Tiao-jie subcommittee in their neighborhood was “active” versus “not active.” The dependent variable for the analyses was respondent-reported household burglary within the past five years. The results of multilevel logistic regressions revealed no significant effect of the Tiao-jie measure. The authors speculated that these null results were likely to due to the crude measurement of “semi-public control” (Zhang at al. 2007:934).

Implementing a research design similar in many respects to that in the Tianjin study, Jiang, Land, and Wang (2013) conducted a household-level criminal victimization survey in the city of Guangzhou in 2007. The primary outcome variable in this study was perceptions of property crimes occurring within the respondents’ neighborhoods within the past six months. One of the key independent variables was conceptualized as “semiformal control,” which was measured with a dummy variable intended to capture the level of activity of the Tiao-jie subcommittee. The questionnaire item asked whether this subcommittee was important in the respondents’ neighborhoods with a response set of “not important,” “uncertain,” and “important.” The categories uncertain and important were combined and were scored “1” to capture a “non-negative evaluation of importance of the neighborhood mediation committee” (Jiang et al. 2013:216). The results of the multilevel regressions diverged from those reported in the Tianjin survey. The coefficient for the Tiao-jie dummy variable was statistically significant in the expected negative direction.

Recent research by Zhang, Messner, and Zhang (2014) further complicates the picture. The primary purpose of this study was to relate different types of neighborhood social control to residents’ perceptions of different forms of public disorder—criminal activity, social disorder, physical disorder, and total disorder. The researchers included a global measure of “semipublic control,” which encompassed a wide range of activities of neighborhood committees, including but not limited to indicators of the functioning of Tiao-jie and Bang-jiao (e.g., assistance to low-income families, help with employment, organizing recreational activities, and efforts to clean up the neighborhood). The results of their multilevel regression analyses revealed unexpected positive effects for the measure of semipublic control: Greater engagement of the neighborhood committees was associated with higher perceived levels of all forms of disorder. To interpret these seemingly anomalous findings, the authors speculated that the neighborhood committees might be serving as “communication vehicles” that transmit information and enhance awareness among residents of the disorderly conditions in their neighborhood, including levels of criminal activity. Consistent with this interpretation, Zhang et al. found that their “global” measure of semipublic control exhibited a negative rather than a positive association with reports of household property victimization that occurred within the neighborhood.

Research Objectives and Hypotheses

The present study is intended to build on and extend the prior research on the role of neighborhood crime control strategies in China. The studies that focused specifically on Tiao-jie and Bang-jiao reviewed above have been limited in significant respects. Representativeness of the early exploratory study by the Tianjin Academy of Social Science is obviously problematic; the respondents were sampled from incarcerated offenders. In addition, little information was available on conditions in the inmates’ neighborhoods, other than a crude measure of their socioeconomic composition. The 2004 Tianjin study and the 2007 Guangzhou study were based on general population samples rather than incarcerated inmates, but only the Tiao-jie mediation subcommittee was considered, and the level of activity of this subcommittee was operationalized rather crudely as a dummy variable in both.

The more recent research by Zhang et al. (2014) based on the 2013 Tianjin study, in contrast, incorporated indicators of both Bang-jiao and Tiao-jie activity, and the measures were more refined (ordinal measures with four categories for each). However, given the objectives of the study, these indicators were combined in a composite measure of semipublic control along with activities of neighborhood committees that have no direct bearing on crime prevention. Moreover, the primary focus of the research was perceptions of disorderly conditions in the neighborhood, and as noted above, different results were observed for reports of actual property victimization, suggesting that in China as in the West, the interconnections between “actual” disorder and perceived disorder in urban neighborhoods are highly complex (Hipp 2010; Kubrin 2008; Sampson 2012; Sampson and Raudenbush 2004; Wallace 2011; Wallace, Louton, and Fornango 2015).

The present study takes the prior research by Zhang et al. (2014) as its point of departure, but given our explicit interest in crime, we specifically examine indicators of the prominent crime control programs that are initiated by neighborhood committees—Tiao-jie, Bang-jiao, and neighborhood watches. Each of these neighborhood crime control programs is hypothesized to contribute to lower levels of victimization within the neighborhood. Tiao-jie activity is intended to reduce levels of crime by resolving disputes and conflicts among neighbors that might otherwise escalate and promote serious criminal offending. Bang-jiao activity is geared toward reintegrating offenders and thereby minimizing risks of further offending, some of which would be directed at convenient targets in the neighborhood. Neighborhood watches have the principal purpose of mobilizing residents to enhance guardianship and surveillance within the neighborhood. In our statistical analyses, we consider the effects on property victimization of each of these neighborhood crime control programs independently as well as their joint impact as captured in a composite index.

Our statistical models also include indicators of the two forms of neighborhood control that have been examined extensively in Western research—collective efficacy and public control. As explained above, a distinctive feature of China has been the fusion of informal and formal social controls. This raises questions about the salience of collective efficacy—a form of distinctively informal control—in the Chinese context. The household surveys conducted by Zhang et al. (2007) in Tianjin and by Jiang et al. (2013) in Guangzhou reported evidence for a negative effect of their indicators of collective efficacy on victimization, but the measures of the semipublic controls associated with the neighborhood committees were rather crude. With respect to public control, the findings from these two prior studies are somewhat inconsistent. Zhang et al. (2007) found a negative effect of their measure of public control on burglary victimization, whereas the indicator of police activity in the study by Jiang et al. (2013) was not significantly related to perceptions of property crime in the neighborhood. It is thus instructive to assess the effects of these “Western” forms of neighborhood social control along with those of the semipublic forms of control that are employed in urban China.

Data and Method

Research Design and Analytic Framework

We follow the strategy employed in the previous Chinese studies that have examined neighborhood committees and crime control (Jiang et al. 2013; Zhang et al. 2007, 2014). These studies implemented a correlational design guided by the Western neighborhood effects analytic framework (see especially Sampson et al. 2002). This general analytic framework models variation in levels of crime across neighborhoods with reference to measures of structural characteristics of neighborhoods along with indicators of social control processes that are theorized to serve as intervening mechanisms (for reviews of the Western literature, see Kubrin and Weitzer 2003; Messner and Zimmerman 2012; Skogan 2012; Warner and Clubb 2013). As elaborated below, our model specification is informed by this general analytic framework.

Data Collection

The data for our analyses are based on a survey that was conducted in collaboration with the Department of Sociology at Nankai University. The sampling entailed a multistage cluster design with a target of 2,500 households distributed across 50 neighborhoods in Tianjin. The selection of neighborhoods for research in urban China entails a different approach than is commonly used in Western research. In the West, researchers generally rely on territorial units that were developed for purposes of collecting official statistics (e.g., census tracts). The situation in China is different. As noted above, neighborhoods are politically designated units of social organization that are managed by the officially recognized neighborhood committees. Neighborhood committees in the past encompassed relatively small populations (100–500 households), but these units have increased in size along with the growth of cities. At present, neighborhood committees in Tianjin typically have jurisdiction over approximately 2,000–3,000 households.

Due to logistical and political constraints, it was not possible to implement strict probability sampling, and thus a combination of probability and purposive sampling techniques were implemented. The starting point was the probability sample of 16 neighborhoods that are used in the Chinese General Social Survey (CGSS). Twelve of the neighborhoods are located in the six traditional districts of the city—Heping, Nankai, Hongxiao, Hexi, Hebei, and Hedong—and four are in the district of Binhai, which is outside the central area of the city. The CGSS is a national survey of about 10,000 households in urban and rural areas of China, which was initiated in 2003. The Tianjin sample is part of the national survey (see Bian and Li 2012, for a description of the CGSS). Using the CGSS sample as a starting point, the Chinese collaborates in the research selected the remaining neighborhoods based on their extensive survey experience in the six traditional districts of the city. They assessed socioeconomic conditions of candidate neighborhoods and their geographic locations to enhance the representative of the overall sample.³

With assistance from neighborhood committees, the research team conducted systematic sampling to select approximately 53–56 households from each of the selected neighborhoods using household addresses. The research team contacted the respondents to schedule in-home interviews. In households with more than one person aged 18 years or older, the resident with a birthday closet to a criterion date was selected to be the respondent. Face-to-face interviews were conducted in the selected households. Consistent with standard institutional review board (IRB) protocols, respondents were assured of the voluntary nature of their participation, their right to refuse to answer questions, and the confidentiality of their responses. After an interview was completed, the interviewer placed the questionnaire in an envelope that was sealed and all the completed questionnaires were transmitted directly to the chief researcher, who secured them in a safe location after a round of interviews was completed in a selected neighborhood. A total of 2,497 valid questionnaires were obtained after some replacement, a response rate that is remarkably high by Western standards—about 99 percent. The high response rate is likely due to the survey method. The neighborhood committees facilitated the initial contacts between the researchers and the sampled households, which might have enhanced willingness to participate. This procedure might raise questions about subtle coercion, although as noted respondents were explicitly provided with standard IRB protocols, and some respondents declined to answer certain questions. The statistical analyses reported below are based on a sample with list-wise deletion of cases (N = 2,479).

Measures

Dependent variable

The principal dependent variable for the analyses is respondent-reported household property victimization. Respondents were asked whether they or any member of their households had been the victim of a crime that involved the loss of property (without the use of force or threat of force) during the past 12 months.⁴ If respondents answered “yes,” they were asked to identify how many of the incidents occurred and how many of the incidents occurred within their own neighborhoods. These responses were used to create a dummy variable scored 1 for a report of any household property victimization within their neighborhood during the time frame and scored 0 for no victimization within their neighborhood. One tenth (10 percent) of respondents reported that a member of their household had experienced a property victimization that had occurred within the neighborhood during the past year, a figure comparable to that reported in the United States at 11 percent (Truman and Morgan 2016).⁵ Descriptive statistics for variables entered into the regression equations are reported in Appendix Table A1.

Independent variables

The primary independent variables are indicators of the major crime prevention activities sponsored by the neighborhood committees—Tiao-jie, Bang-jiao, and neighborhood watches. Respondents were asked how often they had known of, or had heard about, any Tiao-jie activities provided by their neighborhood committee during the past year. The response categories and the corresponding numerical scores were 1 = never, 2 = once, 3 = a few times, and 4 = many times. Respondents were asked an analogous question about Bang-jiao activity, which was scored in the same manner. The survey responses on each of these items was averaged to represent the level of activity at the neighborhood level.

Respondents were also asked if they were aware of any neighborhood watch activities being organized by their neighborhood committee in the last 12 months, and if so, whether these activities operated on a regular basis. We combined the information from the two items to create a measure scored 0 to represent no report of neighborhood watch activity, 1 to represent some neighborhood watch activity but not on a regular basis, and 2 to represent regular neighborhood watch activity. To capture the combined influence of the three forms of neighborhood crime control, we created a scale based on factor scores, given the high degree of shared variance (the factor loadings range from 0.7 to 0.86), which was aggregated to the neighborhood level.

Control variables

Consistent with previous analyses in China (Zhang et al. 2007, 2014) and the Western neighborhood effects analytic framework, we estimated multilevel regression equations that include indicators of relevant respondent and household-level characteristics, along with measures of neighborhood structural characteristics and measures of social control processes other than the crime prevention strategies under investigation. Information was collected on two structural characteristics of neighborhoods: poverty and residential stability. Poverty was measured as the percentage of families with monthly incomes below 1,999 yuan (Chinese dollars). This is a reasonable poverty level according to the 2013 official poverty line in the city of Tianjin. The poverty line was 600 Chinese dollars per family member and the average family size was 2.75 in the city in 2013 (Tianjin Statistical Yearbook 2014). Residential stability was measured as the mean number of years the respondents has lived at their current neighborhoods.⁶

Measures of two forms of neighborhood social control other than Tiao-jie, Bang-jiao, and neighborhood watches were included in the regression models, serving mainly as statistical control variables for present purposes.⁷ One is generally conceptualized as a form of “public control.” Public control refers to the capacity of neighbors to secure resources from outside the neighborhood to maintain order, especially law enforcement resources (Bursik and Gramsick 1993:17). Research in the West has affirmed the utility of incorporating public control in explanations of neighborhood variation in crime and disorder (Velez 2001; see also Carr 2003), and evidence also suggests that public control (policing) affects levels of crime across Chinese neighborhoods, although the results have been inconsistent (cf. Zhang et al. 2007 with Jiang et al. 2013). Our measure of public control captures residents’ evaluations of police activity in their neighborhood. We created a composite index based on survey items dealing with the frequency of visits from the local police officer, police officer’s level of concern with safety/security of the neighborhood, the extent to which police respond in a timely and effective manner, and the competence of police.

A second form of neighborhood social control that needs to be controlled to isolate any distinctive effect of crime control activities of the neighborhood committees is “collective efficacy.” Collective efficacy is a higher order theoretical construct that combines shared expectations among neighbors for informal control with key elements of social cohesion—trust and mutual support (Sampson 2006:152; see also Sampson 2012). Neighborhoods characterized by a high degree of collective efficacy are those in which residents trust one another, support one another, and are confident that fellow neighbors will act collectively to solve common problems. Empirical research in the West has offered considerable support for the claims about the role of collective efficacy in explaining variation in levels of crime across urban neighborhoods (Morenoff, Sampson, and Raudenbush 2001; Sampson 2006; Sampson, Raudenbush, and Earls 1997). Similarly, the multilevel study by Zhang et al. (2007) detected a significant effect (with a one-tailed test) of their indicator of collective efficacy on neighborhood levels of burglary.

We computed a measure of collective efficacy to conform to procedures widely used for this concept in Western research (Sampson et al. 1997). The first step was to create an index of social cohesion in the neighborhood. This index was computed by summing responses to items asking about getting along with neighbors, exchanges of assistance, how close-knit the neighborhood is, shared concerns, and levels of trust. We also created an additive index of willingness of neighbors to exercise informal control based on items asking about the likelihood of intervention of neighbors to deal with problems in the neighborhood. These indexes were aggregated to the neighborhood level and summed to yield the score on collective efficacy. The correlation between the two components of collective efficacy—the aggregated scores on social cohesion and the willingness of neighbors to intervene—was .82, highly similar to the correlation (.80) reported for the research based on neighborhood clusters in Chicago (Sampson et al. 1997:920).⁸

Given the multilevel design of the survey, it is possible to control for respondent and household characteristics that might affect the reports of household victimization irrespective of conditions in the neighborhood. We created measures of three household-level factors that on theoretical grounds might reflect differential risks of victimization within any given neighborhood. Our selection of these respondent-level and household-level control variables draws upon the logic of routine activities theory (for an illustration in previous research in China, see Zhang et al. 2007). Households with higher family incomes might be particularly attractive targets for property crimes, especially burglaries. In contrast, residents who have lived in a neighborhood a long time may have well-established social relationships that provide guardianship, thereby reducing victimization risk. We accordingly included the following measures: family income (monthly)—nine response categories ranging from 1 = below 1,000 yuan to 9 = 8,000 yuan and length of residence—number of years the respondent reports having lived at the current residence. A third household-level variable potentially relevant to risk of victimization is the presence of youths at home—the number of residents living in the household under the age of 18. The predictions for this variable are somewhat ambiguous. On the one hand, youths tend to be relatively attractive targets for minor property crimes in the immediate neighborhood (e.g., bicycle thefts and larcenies). On the other hand, the presence of young people is likely to increase guardianship of the household itself.

We also created measures for demographic variables to be able to assess their possible effects on respondents’ victimization reports. Recall that these reports refer not only to the respondents’ personal experiences but to victimizations that have been perpetrated against any member of the household. Awareness of incidents affecting other household members might conceivably vary by demographic characteristics, which would constitute measurement bias for present purposes. These variables include age, gender, education, marital status, and employment status. They were measured as follows: age—in years; gender—a dummy variable coded in the direction of males; education—seven response categories ranging from 1 = no formal schooling to 7 = graduate school or above; marital status—1 = married and 0 = unmarried; and employment—1 = having employment and 0 = no employment.

Analytical Strategy and Statistical Procedures

Our statistical models for the hypothesis testing are based on multilevel regressions. Given the dichotomous nature of the main dependent variable (household property victimization), we estimated the following logistic regressions. Assume that for the ith individual in the jth neighborhood, we observe a dichotomous response for victimization in his or her household:

Y_ij = 1 for sample individual who responds “yes.”

Y_ij = 0 for sample individual who responds “no.”

We assume that the Y_ij responses are distributed independently as a Bernoulli random variable, and we denote by p_ij the probability that the ith individual observation will respond “yes.” Then we define the logit of victimization for individual i in neighborhood j as logit (Y_ij ) = log[p_ij /(1 − p_ij )]. Thus, for the individual i residing in neighborhood j, our level 1 model is:

Logit (Y_{i j}) = β_{0 j} + β_{1 j} X_{2 i j} + β_{2 j} X_{2 i j} + \dots + β_{P j} X_{P i j},

where β _pj (p = 0, 1,…, Q) are coefficients and X_pij is individuals’ pth household-level variables or demographic characteristics for case i in neighborhood j. At the neighborhood level, we model β_0j as a dependent variable:

\begin{array}{l} β_{0 j} = γ_{00} + γ_{01} W_{1 j} + γ_{02} W_{2 j} + \dots + γ_{0 S 0} W_{S 0 j} + u_{0 j} \\ = γ_{00} + \sum_{s = 1}^{S_{0}} γ_{0 s} W_{s j} + u_{0 j}, \end{array}

where γ_0s (s = 0, 1,…, S_q ) are coefficients for neighborhood-level variables, W_sj is a neighborhood-level explanatory variable, and u_0j is a neighborhood-level random effect.

Following conventional procedures, we begin the regression analyses by estimating an intercept-only model. The results indicate whether or not household property victimization varies significantly across neighborhoods. The next model is a baseline model, which includes the level 1 variables (indicators of household risk for victimization and the demographic controls for differential awareness of victimizations within the household) along with the neighborhood-level variables that for our purposes serve as control variables. The measures of Tiao-jie, Bang-jiao, and neighborhood watches are then introduced separately, followed by a model with the composite Index of Neighborhood Controls. Given that the signs of the relationships for the indicators of the crime control activities of the neighborhood committees are predicted, we apply one-tailed tests of statistical significance for their coefficients. We apply two-tailed tests for other variables that serve as controls for our purposes, even though prior theory and research might justify one-tailed tests in some instances. The regression tables report both coefficients and standard errors for the predictor variables. In further exploratory analyses, we treat the Index of Neighborhood Controls as a dependent variable and estimate hierarchical linear models.

Results

The results of our multilevel regression analyses are presented in Table 1. The first model is the intercept-only model. The results indicate that reported property victimization varies significantly across the 50 neighborhoods as reflected in the significant variance component (χ² = 45.91**). Model 2 introduces the level 1 variables (household and demographic characteristics) and the neighborhood-level controls. None of the variables has a significant effect on property victimization. The null effect for the measure of poverty is particularly surprising and contrary to those findings in the West, although it is consistent with the earlier study of neighborhood burglary victimization in Tianjin by Zhang et al. (2007) and Zhang et al.’s (2014) more recent research on perceived neighborhood criminal disorder. Zhang et al. (2007) interpreted the lack of any significant effect of poverty on crime by noting that neighborhood control in urban China is not as dependent on local financial resources as in the West, given the role of the city government in sponsoring neighborhood organization through the neighborhood committees.

Table 1.

Multilevel Logistic Regressions of Household Property Victimizations.

Variables	1	2	3	4	5	6
Constant	−2.410** (0.132)	−4.260 (2.242)	−2.629 (2.219)	−0.767 (2.290)	−5.562* (2.374)	−4.753* (2.114)
Level 2
Poverty level		0.012 (0.024)	0.004 (0.023)	0.007 (0.022)	0.002 (0.024)	−0.000 (0.023)
Residential stability		0.000 (0.025)	0.020 (0.025)	0.023 (0.024)	0.009 (0.026)	0.023 (0.025)
Public control		−0.047 (0.049)	−0.046 (0.046)	−0.056 (0.044)	−0.011 (0.053)	−0.027 (0.045)
Collective efficacy		0.140 (0.075)	0.139 (0.071)	0.070 (0.071)	0.148* (0.074)	0.117 (0.070)
Tiao-jie			−0.772** (0.295)
Bang-jiao				−0.844** (0.257)
Neighborhood watch					−0.507^† (0.344)
Index of Neighborhood Crime Controls						−0.828** (0.279)
Level 1
Family income		0.007 (0.037)	0.001 (0.037)	−0.000 (0.037)	0.004 (0.037)	−0.001 (0.037)
Length of residence		0.010 (0.007)	0.010 (0.007)	0.010 (0.007)	0.010 (0.007)	0.010 (0.007)
Children at home		0.249 (0.134)	0.237 (0.134)	0.233 (0.134)	0.251 (0.134)	0.239 (0.134)
Variance (constant)	.516	.464	.385	.340	.438	.361
χ²	45.91**	38.05**	31.27**	26.75**	35.55**	28.78**

Note. N = 2,479 at the individual level and 50 at the neighborhood level. Level 1 controls of education, age, gender, marital status, and employment status are included in models 2–6 but are not shown. See text for explanation. Based on our predictions, we apply one-tailed tests of statistical significance for indicators of the crime control activities of neighborhood committees. Standard errors are given in parentheses.

^† p = .07. *p < .05. **p < 0.01.

Zhang et al. (2007) also reported a significantly positive coefficient for residential stability, in contrast with our finding of a nonsignificant coefficient. They interpreted this unexpected positive effect with reference to features of the housing market at the time—newly constructed housing projects had many attractive features, which reduced victimization risks despite the short tenure of the residents. It seems plausible that the discrepancy in the results for residential stability across studies reflects changes in the housing market in Tianjin over the course of the intervening decade. This interpretation is consistent with the nonsignificant coefficient for residential stability in Zhang et al.’s (2014) recent study of perceived criminal disorder in Tianjin.

Models 3–6 allow for assessments of the influence of the community crime control programs of Tiao-jie, Bang-jiao, and neighborhood watches on the risks of household property victimization. As expected, the indicators of all three activities directed by the neighborhood committees are negatively related to household property victimization. The coefficients for Bang-jiao, Tiao-jie, and the Index of Neighborhood Crime Controls reach statistical significance, while the coefficient for neighborhood watch is borderline significant (p = .07). In other words, neighborhoods in which the traditional Chinese crime control programs are reported to be actively implemented do indeed appear to be characterized by comparatively low levels of property crime as reflected in reported household victimization.

Although we conceptualize the implementation of Tiao-jie, Bang-jiao, and neighborhood watches as properties of neighborhoods, our measurement is based on the individual-level survey responses aggregated to the neighborhood level. Respondents essentially served as informants of the conditions within their neighborhoods. We have no theoretical reason to expect that individual perceptions of these crime control programs affect victimization risk above and beyond the effects of their actual implementation. Nevertheless, to assess the robustness of our findings, we reestimated the regression models including the individual-level analogues for the aggregated measures of Tiao-jie, Bang-jiao, and neighborhood watches, group-mean centered, as level 1 variables to partition individual-level and contextual effects (Feaster et al. 2011). The results for these models are substantively the same as those in Table 1. Similar results are obtained when the regression models are expanded to include individual-level analogues as level 1 predictors for the other perceptual indicators of neighborhood conditions (public social control and collective efficacy).

It is particularly noteworthy that the indicators of Tiao-jie, Bang-jiao, and neighborhood watches are the only neighborhood-level measures that yield significant coefficients in the regression models predicting property victimization in the neighborhood, with the exception of the one coefficient for collective efficacy, which has an unexpected positive sign. The null effect for public control (policing) is consistent with the results reported by Jiang et al. (2013) in the 2007 Guangzhou survey but is inconsistent with finding reported by Zhang et al. (2007) based on data on burglary from the 2004 Tianjin survey. The discrepancy between our findings and those for the earlier Tianjin study could reflect methodological differences or perhaps changes over time, although it is not clear why the effect of public control would have weakened in the ensuing decade. The null effect of the measure of public control could also reflect reciprocal causal processes. Insofar as a high level of crime in the neighborhood results in more intensive police activity, and this activity gets translated into greater awareness and positive evaluations of policing, a positive association would be created between measures of public control and neighborhood criminal activity, which would be opposite to any negative effect reflecting processes of control. The possibility of reciprocal causal processes with counterbalancing implications has long been recognized in the macrolevel research on policing and crime rates (see, e.g., Kubrin et al. 2010; Levitt 1997; Marvell and Moody 1996).

The results for the measure of collective efficacy are also surprising, given the prominence of this concept in Western research. Note that the models in Table 1 are directed toward assessing any independent effect of collective efficacy, net of the other neighborhood semipublic controls (and other covariates). It is possible, however, that collective efficacy, or at least aspects of it, affects victimization indirectly through the semipublic controls.⁹ To explore this possibility, we estimated multilevel linear regression models with the Index of Neighborhood Crime Controls serving as the dependent variable as a household-level variable. In these models, we examined the effects of the original measure of collective efficacy, its two components separately (cohesion and willingness of residents to intervene), and an elaborated indicator of social cohesion (explained below).

The results are reported in Table 2. All level 1 covariates are included in these models, but the coefficients are suppressed to facilitate the readability of the table. The only level 1 covariate with a significant effect is age. Older respondents report lower levels of crime control activity directed by the neighborhood committees than do younger respondents.

Table 2.

Regressions of Index of Neighborhood Crime Controls.

Variables	1	2	3	4	5
Constant	−0.256 (1.086)	−0.221 (1.158)	−0.499 (0.992)	−0.428 (1.193)	−3.254** (0.999)
Level 2
Poverty level	−0.014 (0.012)	−0.013 (0.012)	−0.013 (0.012)	−0.014 (0.012)	−0.005 (0.010)
Residential stability	0.021 (0.013)	0.022 (0.013)	0.019 (0.013)	0.019 (0.014)	−0.005 (0.012)
Public control	0.033 (0.025)	0.029 (0.025)	0.034 (0.025)	0.034 (0.025)	0.049* (0.021)
Collective efficacy	−0.043 (0.039)
Cohesion		−0.062 (0.066)		−0.011 (0.100)
Intervention			−0.094 (0.081)	−0.084 (0.124)	−0.240** (0.073)
Cohesion (expanded)					0.173** (0.036)
Variance (constant)	.764	.764	.764	.764	.764
χ²	407.34**	409.43**	407.09**	406.84**	257.91**

Note. N = 2,479 at the individual level and 50 at the neighborhood level. Level 1 controls include family income, length of residence, the number of children at home, education, age, gender, marital status, and employment status. All but age and length of residence are nonsignificant. See text for explanation. Standard errors are given in parentheses.

*p < .05. **p < .01.

Models 1–4 indicate the effects of collective efficacy as commonly measured in Western research on the Index of Neighborhood Crime Controls. As shown in model 1, the coefficient for collective efficacy is nonsignificant. The same is true for its constituent elements—the measures of social cohesion and willingness of neighbors to intervene—when entered in place of the composite measure either separately (models 2 and 3) or together (model 4). In other words, indicators of the Western conceptualization of collective efficacy and its subdimensions, as conventionally operationalized, are unrelated to the crime prevention efforts of the neighborhood committees in Tianjin.

We also explored the effects of an expanded measure of social cohesion. As noted above, our measurement of collective efficacy was guided by the approach adopted in Western research. We accordingly selected items that were similar to those previously used to capture neighborhood social cohesion in the West. However, the survey instrument also includes additional indicators of “neighboring” that might be conceptualized as reflecting social cohesion, especially as manifested in a collectivistic society, such as China. Specifically, respondents were asked how often they invited their neighbors to their home or were invited to their neighbors’ home, how often they grocery shopped for their neighbors, how often they helped care for elderly neighbors or watched over their neighbors’ children, and how easy it was for respondents to pick out outsiders in the neighborhood.

We conducted exploratory factor analyses on the original items included in the social cohesion measure along with these additional items. The results indicated that they all loaded quite well on a single factor with one exception. The question about how often respondents got along well with their neighbors had a factor loading of only 0.34. Without that item, the other items loaded on the factor at 0.5 or higher. We accordingly constructed an expanded measure of social cohesion without the question on “getting along well.”¹⁰

We reestimated the regression equation reported in model 4 of Table 2, predicting the Index of Neighborhood Crime Controls with the expanded measure of social cohesion substituted for the original measure of social cohesion. The results are quite intriguing (model 5). The coefficient for social cohesion is now significantly positive (0.173**). This finding suggests that in the Chinese context, social cohesion is indeed relevant to crime control, but it plays a somewhat different role than it does in the West. High levels of social cohesion apparently help to stimulate and sustain the traditional Chinese crime control strategies.

The coefficient for the willingness of neighbors to intervene directly is now significantly negative (−0.240**). Residents in neighborhoods where fellow neighbors are willing to intervene directly for purposes of social control report lower levels of crime control activities directed by the neighborhood committees. We suspect that this is likely to reflect reverse causation or reciprocal causation. Insofar as the community crime control programs of Tiao-jie, Bang-jiao, and neighborhood watches are implemented actively, there is less need for residents to intervene directly to deal with collective problems. Causal order for collective efficacy and the crime control operations of neighborhoods committees is also uncertain and possibly reciprocal. Social cohesion is likely to stimulate greater engagement of the neighborhood committees, which reinforces and amplifies social cohesion. Accordingly, the regression results reported in Table 2 must be regarded as exploratory and at most suggestive, given the cross-sectional nature of the data. Disentangling the causal processes referenced above is a task for future research.

Summary and Conclusions

Scholars have long recognized that social control in contemporary China involves a blending of formal and informal mechanisms. The Chinese government has established, and continues to nourish, an organizational infrastructure in urban areas designed to mobilize the general population for purposes of governance—the neighborhood committees. As Read (2012:17) has observed, these organizations effectively “graft part of the apparatus of government into the fibers of community networks.” Among the key governance functions of these neighborhood committees is to promote a safe and secure local environment—to control crime. Tiao-jie, Bang-jiao, and authoritatively directed neighborhood watches are the most widely employed operational strategies at the neighborhood level. To date, very little quantitative evidence has been available to evaluate the extent to which these traditional community crime control programs actually influence criminal victimization in urban neighborhoods. We analyzed data from a recent survey in Tianjin in an effort to fill this conspicuous gap in the criminological literature.

Our multilevel analyses find that Tiao-jie, Bang-jiao, and neighborhood watch programs do in fact appear to contribute to crime control for property victimization. Net of indicators of structural conditions in the neighborhood, indicators of other forms of neighborhood control (policing, collective efficacy), and measures of household-level and individual-level characteristics, the indicators of the level of activity of Tiao-jie, Bang-jiao, and neighborhood watches exhibit negative effects on reported household property victimization. Similarly, the coefficient for a composite index of these activities is significantly negative. The analyses thus indicate that neighborhoods in Tianjin in which these traditional crime control programs are actively implemented exhibit lower levels of property crime victimization than do those in which such programs are implemented relatively infrequently. These findings for the community crime control programs stand in marked contrast to those for a measure of collective efficacy—a concept that plays a prominent role in Western research on crime in urban neighborhoods. Our measure of collective efficacy does not yield the expected negative effect on household property victimization.

We further explored the possibility of a facilitating effect of collective efficacy on crime control by regressing the composite Index of Neighborhood Crime Controls on collective efficacy and its constituent elements along with the other covariates. The results differ depending on measurement decisions. When social cohesion is operationalized using items analogous to those used in Western research, there are no significant associations between collective efficacy or its components and the Index of Neighborhood Crime Controls. In contrast, when social cohesion is operationalized with items reflecting broader forms of “neighboring,” the coefficients for social cohesion and the measure of the willingness of neighbors to intervene are both statistically significant. The effect of social cohesion is positive: The more cohesive social neighborhoods tend to be characterized by higher levels of Tiao-jie, Bang-jiao, and neighborhood watch activity. The effect of the measure of the willingness of neighbors to intervene, in contrast, is negative. Neighborhoods in which residents tend to refrain from direct intervention are those characterized by comparatively high levels of community crime prevention activity directed by the neighborhood committees. Causal order in this instance is not entirely clear, and reciprocal causation seems plausible.

While our findings suggest some important theoretical implications on social order maintenance in urban China with its organic social control apparatus called neighborhood committees, our study is not without limitations. First, due to logistical necessities, we employed a hybrid of probability and purposive sampling to boost our territorial coverage and sample size. The representativeness of the sample may thus be problematic, as is the application of formal tests of statistical significance. The implementation of the survey was facilitated by members of the neighborhood committees, who provided access to the residents. It is hard to imagine how a survey of this type could be done without such assistance, given the sociopolitical realities in contemporary urban China. Nevertheless, the extent to which such “authorization” and “endorsement” from the neighborhood committee affected responses to the survey is unknown. Finally, cities in China are obviously quite diverse in size and demographic composition. The generalizability of our findings to other large Chinese cities must await future research.

It is also important to acknowledge that our analyses are limited to explaining the risks of property crime victimization across neighborhoods. The data do not reveal the identity of offenders, and thus it is not possible to probe any “developmental effects” of the neighborhood crime control strategies—the effects on individual offending. As Sampson (2006:157) observed, there is a logical separation between explaining “…variation of crime event rates across neighborhoods regardless of who commits the acts (residents or otherwise)….” and explaining “…how neighborhoods influence the individual behavior of residents no matter where they are” (see also Wikström and Sampson 2003). The crime control strategy of Bang-jiao in particular is explicitly oriented toward developmental effects, that is, reintegrating offenders into the community and thus reducing the likelihood of their criminal activity. Although the developmental impacts of such a program should be manifested to some extent in victimizations within the neighborhoods wherein those participating in the program reside, a fuller assessment of the efficacy of Bang-jiao would require a different research design and data sources (e.g., self-reports of offending that can be linked with the residence of participants).

With these limitations in mind, our analyses have some important implications for comparative criminology more generally. Much attention has been devoted to the identification of large-scale “homogenizing” forces. For example, classical modernization theory directed attention to the various ways that “becoming modern” tends to weaken controls and stimulate criminal motivations (Shelley 1981; see also Neapolitan 1997; Shelley 1986). Early theories of globalization similarly highlighted criminogenic pressures that are common to nations as they become more interconnected to a globalized world, many of these pressures associated with the adoption of a market economy (Currie 1991; Teeple 1995). Evidence indicates that some of these forces are applicable to the Chinese case. Rising crime rates have been documented and have been interpreted with reference to processes of modernization, global marketization, and accompanying anomie, although with some qualifications (Liu 2006; Liu et al. 2001; Zhao 2008).

However, more recent scholarship has emphasized the fact that whatever “master forces” might accompany the developmental process, nations follow distinctive “pathways” that reflect their cultural traditions and their unique institutional arrangements (Messner, Liu, and Karstedt 2008:275). Our analyses are quite consistent with approaches that reject a sharp divide between tradition and modernity in favor of the view that contemporary practices build on, and are legitimated by, long-standing traditions. Despite the rapid transition to modernity and the profound social transformations associated with the economic reform in China, the traditional crime prevention mechanisms of organized and directed by the neighborhood committees evidently remain quite relevant in the newly emerging urban landscape. Perhaps this should not be too surprising, given that these mechanisms embody the core value orientations of the society—a strong emphasis on collectivism and communitarianism.

Our results also have implications for the theorizing about neighborhood effects in criminology (Sampson et al. 2002). Without question, the formulation and the application of the concept of collective efficacy have contributed greatly to an understanding of the mechanisms that link the structural conditions in urban neighborhoods in Western cities with levels of criminal activity (see Sampson 2006, for a review of the empirical literature). We have tried to measure the construct in a way that closely resembles the procedures that are commonly used in the West, and it appears that we have been reasonably successful in doing so. Measurement properties are similar. Yet the measure of collective efficacy does not perform very well in our regression models. It does not emerge as a very useful predictor of variation in neighborhood levels of property crime victimization with the survey data for Tianjin, either directly or indirectly via an impact on the crime control activities of the neighborhood committees.

We would not dismiss, however, the utility of the general notion of collective efficacy. Rather, it is likely to have widespread applicability if it is reconceptualized at a slightly higher level of abstraction. The current conceptualization joins neighborhood social cohesion with the willingness of residents to engage in what can be characterized as “voluntary self-governance” (Read 2012:3). This voluntary self-governance is based on the collective action of largely independent actors who choose to join together on their own to deal with some perceived local need. Such a mechanism of informal social control makes a good deal of sense in societies with predominantly individualistic value orientations and strong civic institutions. In the Chinese context, there is a different organizational infrastructure available to the residents of urban neighborhoods that is more compatible with the collectivistic values of the society—the neighborhood committees. These committees sponsor and facilitate the social control efforts of residents.

Finally, although our focus in this study has been on crime control, it is important to recognize that the control exercised by neighborhood committees can also be experienced as being highly intrusive and repressive. Read (2012) provided a compressive discussion of the complexities surrounding the operations of neighborhood committees in contemporary urban China and of residents’ attitudes toward them. Nonetheless, rather than relying on motivated individuals to mobilize community members to realize collective goals, these neighborhood committees evidently become salient features of the sociocultural landscape in urban China, thus challenging us to rethink how collective efficacy may manifest itself in different sociocultural context.

Our observations about reconceptualization are quite compatible with Sampson’s (2012:159) conclusion in his recent review of the evidence concerning the “reach of collective efficacy.” Sampson (2012:168) acknowledged cross-cultural variation in the manifestations of collective efficacy, and he observed that “what is striking is the fact that large community-level variations can be detected and econometrically assessed in such divergent country settings.” Our analyses of Tiao-jie, Bang-jiao, and neighborhood watches in contemporary urban China offer an additional case in point. Community-level variations in traditional forms of crime control have measurable effects on risks of property crime victimization. A key challenge for future comparative criminological research on neighborhood effects is to cast the net widely to identify additional forms of collective efficacy across divergent settings and to adapt and modify theories as need be to accommodate them.

Footnotes

Appendix

Table A1.

Descriptive Statistics of Variables.

Variables	Mean	Standard Deviation
Dependent variable
Property victimization	0.10	0.30
Level 1 independent variables
Male	0.47	0.50
Employed	0.46	0.50
Married	0.80	0.40
Age	49.61	15.13
Education (scale of 1–7)	4.24	1.26
Length of residence	15.94	11.34
Children at home	0.30	0.51
Family income	5.64	2.11
Level 2 independent variables
Poverty level	6.90	5.84
Residential stability	15.91	6.06
Public control	45.80	3.33
Collective efficacy	28.29	2.08
Cohesion	17.81	1.23
Intervention	10.47	0.94
Cohesion (expanded)	22.57	1.89
Tiao-jie	2.36	0.45
Bang-jiao	1.63	0.55
Watch	1.24	0.47
Index of Neighborhood Controls^a	0.00	0.50

Note. N = 2,479 at the household level and 50 at the neighborhood level.

^aThe Index of Neighborhood Controls is measured at the household level when serving as a dependent variable in the multilevel regressions and as a neighborhood-level variable when serving as an independent variable. See text for explanation.

Authors’ Note

Any opinions, findings, and conclusions or recommendations expressed herein are those of the authors and do not necessarily reflect the views of the National Science Foundation.

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: This article is based on research supported by the National Science Foundation under grant no. 1126175.

Notes

References

Bian

Yanjie

. 2002. “Chinese Social Stratification and Social Mobility.” Annual Review of Sociology 28:91–116.

Bian

Yanjie

Lulu

. 2012. “The Chinese General Social Survey (2003–2008): Sample Design and Evaluation.” Chinese Sociological Review 45:70–97.

Bursik

Robert J.

Jr . 1988. “Social Disorganization and Theories of Crime and Delinquency: Problems and Prospects.” Criminology 26:519–51.

Bursik

Robert J.

Jr Grasmick

Harold G.

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

Carr

Patrick J.

2003. “The New Parochialism: The Implications of the Beltway Case for Arguments Concerning Informal Social Control.” American Journal of Sociology 108:1249–91.

Cohen

Jerome A.

1966. “Chinese Mediation on the Eve of Modernization.” California Law Review 54: 1201–26.

Currie

Elliott

. 1991. “Crime in the Market Society: From Bad to Worse in the Nineties.” Dissent 38:254–59.

Feaster

Daniel

Brincks

Ahnalee

Robbins

Michael

Szapocznik

Jose

. 2011. “Multilevel Models to Identify Contextual Effects on Individual Group Member Outcomes: A Family Example.” Family Process 50:167–83.

Hipp

John R.

2010. “Resident Perceptions of Crime and Disorder: How Much is ‘Bias,’ and How Much is Social Environment Differences?” Criminology 48:475–508.

10.

Hong

Dongying

. 2011. Research on Changes in Mediation in Contemporary China (in Chinese). Shanghai, China: Shanghai People’s Press.

11.

Jiang

Shane

Land

Kenneth C.

Wang

Jin

. 2013. “Social Ties, Collective Efficacy and Perceived Neighborhood Property Crime in Guangzhou, China.” Asian Criminology 8:207–23.

12.

Kornhauser

Ruth Rosner

. 1978. Social Sources of Delinquency: An Appraisal of Analytic Models. Chicago: University of Chicago Press.

13.

Kubrin

Charis E.

2008. “Making Order of Disorder: A Call for Conceptual Clarity.” Criminology & Public Policy 7:203–14.

14.

Kubrin

Charis E.

Messner

Steven F.

Deane

Glenn

McGeever

Kelly

Stucky

Thomas D.

. 2010. “Proactive Policing and Robbery Rates Across U.S. Cities.” Criminology 48:57–97.

15.

Kubrin

Charis E.

Weitzer

Ronald

. 2003. “New Directions in Social Disorganization Theory.” Journal of Research in Crime and Delinquency 40:374–402.

16.

Leng

Shaou-Chuan

Chiu

Hungdua

. 1985. Criminal Justice in Post-Mao China: Analysis and Documents. Albany, NY: SUNY Press.

17.

Levitt

Steven D.

1997. “Using Electoral Cycles in Police Hiring to Estimate the Effect of Police on Crime.” American Economic Review 87:270–90.

18.

Liang

Zai

Zhen

Zhongdong

. 2014. “Changing Patterns of the Floating Population in China, 2000-2010.” Population and Development Review 40:695–716.

19.

Liu

Jianhong

. 2006. “Modernization and Crime Patterns in China.” Journal of Criminal Justice 34:119–30.

20.

Liu

Jianhong

Zhang

Lening

Messner

Seven F.

. 2001. Crime and Social Control in a Changing China. Westport, CT: Greenwood Press.

21.

Hong

Miethe

Terance D.

. 2001. “Community Integration and the Effectiveness of Social Control.” Pp. 105–21 in Crime and Social Control in a Changing China, edited by Liu

Jianhong

Zhang

Lening

Messner

Steven F.

. Westport, CT: Greenwood Press.

22.

Lubman

Stanley

. 2000. Bird in a Cage: Legal Reform in China after Mao. Stanford, CA: Stanford University Press.

23.

Guon

. 2001. “Population Migration and Crime in Beijing, China.” Pp. 65–71 in Crime and Social Control in a Changing China, edited by Liu

Jianhong

Zhang

Lening

Messner

Steven F.

. Westport, CT: Greenwood Press.

24.

Marvell

Thomas B.

Moody

Carlisle E.

. 1996. “Specification Problems, Police Levels, and Crime Rates.” Criminology 34:609–46.

25.

Messner

Steven F.

Liu

Jianhong

Karstedt

Susanne

. 2008. “Economic Reform and Crime in Contemporary Urban China: Paradoxes of a Planned Transition.” Pp. 271–93 in Urban China in Transition, edited by

Logan

John R.

Malden, MA: Blackwell.

26.

Messner

Steven F.

Zimmerman

Gregory M.

. 2012. “Community-level Influences on Crime and Offending.” Pp. 155–72 in The Oxford Handbook on Crime Prevention, edited by Welsh

Brandon C.

Farrington

David P.

. New York: Oxford University Press.

27.

Morenoff

Jeffrey D.

Sampson

Robert J.

Raudenbush

Stephen W.

. 2001. “Neighborhood Inequality, Collective Efficacy, and the Spatial Dynamics of Urban Violence.” Criminology 39:517–59.

28.

Neapolitan

Jerry

. 1997. Cross-national Crime: A Research Review and Sourcebook. Westport, CT: Greenwood Press.

29.

Read

Benjamin L.

2012. Roots of the State: Neighborhood Organization and Social Networks in Beijing and Taipei. Stanford, CA: Stanford University Press.

30.

Read

Benjamin L.

Chen

Chun-Ming

. 2008. “The State’s Evolving Relationship with Urban Society: China’s Neighborhood Organizations in Comparative Perspective.” Pp. 315–35 in Urban China in Transition, edited by

Logan

John R.

Malden, MA: Blackwell.

31.

Rosenbaum

Dennis P.

Schuck

Amie M.

. 2012. “Comprehensive Community Partnerships for Preventing Crime.” Pp. 226–46 in The Oxford Handbook on Crime Prevention, edited by Welsh

Brandon C.

Farrington

David P.

. New York: Oxford University Press.

32.

Sampson

Robert J.

2006. “Collective Efficacy Theory: Lessons Learned and Directions for Future Inquiry.” Pp. 149–67 in Taking Stock: The Status of Criminological Theory, (Advances in Criminological Theory, Vol. 15), edited by Cullen

Francis T.

Wright

John Paul

Blevins

Kristie R.

. New Brunswick, NJ: Transaction Publishers.

33.

Sampson

Robert J.

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

34.

Sampson

Robert J.

Morenoff

Jeffrey D.

Gannon-Rowley

Thomas

. 2002. “Assessing ‘Neighborhood Effects’: Social Processes and New Directions in Research.” Annual Review of Sociology 28:443–78.

35.

Sampson

Robert J.

Raudenbush

Stephen W.

. 2004. “Seeing Disorder: Neighborhood Stigma and the Social Construction of ‘Broken Windows.’” Social Psychology Quarterly 67:319–42.

36.

Sampson

Robert J.

Raudenbush

Stephen W.

Earls

Felton

. 1997. “Neighborhoods and Violent Crime: A Multilevel Study of Collective Efficacy.” Science 15:918–24.

37.

Shaw

Clifford R.

McKay

David D.

. 1942. Juvenile Delinquency and Urban Areas. Chicago: University of Chicago Press.

38.

Shelley

Louise I.

1981. Crime and Modernization: The Impact of Industrialization and Urbanization on Crime. Carbondale, IL: Southern Illinois Press.

39.

Shelley

Louise I.

1986. “Crime and Modernization Reexamined.” Annales Internationales de Criminologie 24:7–21.

40.

Shi

Jin-chuan

Xing-jie

. 2010. “Floating Population, Income Inequality, and Crime” (in Chinese). Shandong University Journal 2:1–15.

41.

Skogan

Wesley G.

2012“Disorder and Crime.” Pp. 173–88 in The Oxford Handbook of Crime Prevention edited by Welsch

Brandon C.

David

. Farrington. New York: Oxford University Press.

42.

Tang

Wang

Parish

William L.

. 2000. Chinese Urban Life under Reform: The Changing Social Contract. New York: Cambridge University Press.

43.

Teeple

Gary

. 1995. Globalization and the Decline of Reform. Atlantic Highlands, NJ: Humanities Press.

44.

Tianjin Statistical Yearbook. 2014. Tianjin Municipal Bureau of Statistics and Survey Office (in Chinese). Beijing: China Statistics Press.

45.

Truman

Jennifer L.

Morgan

Rachel E.

. 2016. Criminal Victimization, 2015. Washington, DC: Bureau of Justice Statistics, U.S. Department of Justice. Retrieved (https://www.bjs.gov/content/pub/pdf/cv15.pdf) (accessed on November 28, 2016).

46.

Van Der Sprenkel

Sybille

. 1962. Legal Institutions in Manchu China: A Sociological Analysis (London School of Economics Monographs on Social Anthropology No. 24). London: Athlone Press.

47.

Velez

Maria B.

2001. “The Role of Public Social Control in Urban Neighborhoods: A Multilevel Analysis of Victimization Risk.” Criminology 39:837–63.

48.

Walder

Andrew G.

2015. China Under Mao: A Revolution Derailed. Cambridge, MA: Harvard University Press.

49.

Wall

James A.

Jr Blum

Michael

. 1991. “Community Mediation in the People’s Republic of China: Roots of….” The Journal of Conflict Resolution 35:3–20.

50.

Wallace

Danielle

. 2011. “A Test of Routine Activities and Neighborhood Attachment Explanations for Bias in Disorder Perceptions.” Crime & Delinquency 20:1–23.

51.

Wallace

Danielle

Louton

Brooks

Fornango

Robert

. 2015. “Do You See What I See? Perceptual Variation in Reporting the Presence of Disorder Cues.” Social Science Research 51:247–61.

52.

Wang

Da-zhong

Chai

Yan-ru

Zhang

Xiao-dong

Guo

Bin

. 2007. “A Survey of Crimes Committed by the Floating Population in Beijing” (in Chinese). Journal of Chinese People’s Public Security University 126:9–15.

53.

Wang

Gui-xin

Liu

Yi-yun

. 2006. “An Analysis of the Characteristics and Causes of Crimes Committed by the Floating Population in Shanghai” (in Chinese). Population Journal 157:44–48.

54.

Wang

Zhi-qiang.

2006. “An Empirical Analysis of Crimes Committed by Floating Populations in Recent Years” (in Chinese). Journal of Chinese People’s Public Security University 120:44–51.

55.

Warner

Barbara D.

Clubb

Audrey C.

. 2013. “Neighborhood Ties, Control, and Crime.” Pp. 333–51 in The Oxford Handbook of Criminological Theory, edited by Cullen

Francis T.

Wilcox

Pamela

. New York: Oxford University Press.

56.

Whyte

Martin K.

Parish

William L.

. 1984. Urban Life in Contemporary China. Chicago, IL: University of Chicago Press.

57.

Wikström

Per-Olof

Sampson

Robert J.

. 2003. “Social Mechanisms of Community Influences on Crime and Pathways in Criminality.” Pp. 118–48 in Causes of Conduct Disorder and Serious Juvenile Delinquency, edited by Lahey

Ben

Moffitt

Terrie

Caspi

Avshalom

. New York: Guilford Press.

58.

Jianhua

. 2013. “Urbanization and Inevitable Migration: Crime and Migrant Workers.” Pp. 209–23 in Routledge Handbook of Chinese Criminology, edited by Cao

Liqun

Sun

Ivan Y.

Hebenton

Bill

. New York: Routledge.

59.

Zhang

Lening

Messner

Steven F.

Liu

Jianhong

. 2007. “A Multilevel Analysis of the Risk of Household Burglary in the City of Tianjin, China.” British Journal of Criminology 47:918–37.

60.

Zhang

Lening

Messner

Steven F.

Liu

Jianhong

. 2008. “A Critical Review of Recent Literature on Crime and Criminal Justice in China: Research Findings, Challenges, and Prospects (Introduction).” Crime, Law, and Social Change 50:125–30.

61.

Zhang

Lening

Messner

Steven F.

Zhang

Sheldon

. 2014. “Neighborhood Social Control and Disorder in Contemporary Urban China: Preliminary Findings from a Large Survey in the City of Tianjin.” Presented at the Sixth Annual Conference of the Asian Criminological Society, Osaka, Japan, June 27–29.

62.

Zhang

Lening

Zhou

Dengke

Messner

Steven F.

Liska

Allen E.

Krohn

Marvin D.

Liu

Jianhong

Zhou

. 1996. “Crime Prevention in a Communitarian Society: Bang-jiao and Tiao-jie in the People’s Republic of China.” Justice Quarterly 13:199–222.

63.

Zhao

Linda Shuo

. 2008. “Anomie Theory and Crime in a Transitional China (1978–).” International Criminal Justice Review 18:137–57.

64.

Zhong

Lena Y.

2009. Communities, Crime and Social Capital in Contemporary China. Devon, UK: Willan.