Interdependency,Perceived Shame,and Probability of Wrongdoing Among Chinese Students: A Partial Test of the Reintegrative Shaming Theory

Abstract

Compared with other mainstream criminological theories, Reintegrative Shaming Theory (RST) draws insufficient empirical evaluations. Given RST’s unique attention to non-Western societies and the significance of shaming in Chinese society, China provides a strategic setting for RST’s empirical assessment. Utilizing a Chinese youth sample, this study explores the association between interdependency and the probability of wrongdoing, and the mediating role of perceived shame. We find that strong school attachment is directly associated with the reduced likelihood of wrongdoing, and indirectly via perceived shame. The effects of family and neighborhood attachments on perceived shame and the probability of wrongdoing are insignificant. The results lend partial support to RST and highlight the essential role of school in children’s development in the Chinese context.

Keywords

interdependency perceived shame probability of wrongdoing juvenile China

Braithwaite’s (1989) Reintegrative Shaming Theory (RST) has commanded scholarly attention for more than three decades. Focusing on the nature of societal responses to offenders’ wrongdoing which encompasses both criminal acts and deviant behavior, RST proposes that crime can be reduced if offenders are reintegrated into society through public recognition of wrongdoing and restoration of their relationships with victims and communities in which they are embedded. It has since provided important insights for developing criminal justice policies and programs (Braithwaite, 2020; Braithwaite et al., 2008; Latimer et al., 2005; Zehr, 2015). For example, the past three decades have witnessed the development and advancement of restorative justice (Umbreit et al., 2005), for which RST is a major impetus.

Two critical concepts in RST are interdependency and shaming (Braithwaite, 1989). Interdependency refers to how much individuals are embedded in networks and attached to parents, schools, neighbors, and employers. People with a high level of interdependency are more susceptible to shaming, defined as the social process of expressing disapproval intending to invoke remorse in wrongdoers. Shaming consists of two major types: reintegrative shaming and stigmatizing shaming. The nature of shaming varies across societies. Shaming in communitarian societies, for example, is more likely to be reintegrative rather than stigmatizing (Braithwaite, 1989; Lu et al., 2002). Shaming, rooted in interdependency and collectivity, plays an essential role in exercising social control and maintaining social harmony in Eastern societies, thus explaining the relationship between the external social environment and involvement in delinquency or criminal activities (Braithwaite, 1989, 2014; Chen, 2002).

Interestingly, whereas RST has achieved widespread recognition, the empirical assessment of RST’s key theoretical postulates is relatively limited, particularly in comparison with other contemporary criminological theories. For example, few studies have directly examined the role of shaming in explaining the association between interdependency and delinquency/crime involvement. In addition, this moderate body of literature uses vastly different conceptualization and measurement tools, yielding inconsistent and often conflictual findings. This is unfortunate since RST directs unique attention to cultural differences and provides theoretical insights in non-western cultural contexts. To fill the research gap in the literature, this study tests RST’s micro-level propositions in the Chinese context. Utilizing a Chinese youth sample, we explore the association between interdependency and the probability of wrongdoing, with a particular focus on the mediating role of perceived shame. In other words, we test whether a higher level of interdependency increases children’s sense of shame, which in turn reduces the likelihood of children engaging in wrongdoing. In the following sections, we first review the literature on RST, including its theoretical framework and empirical assessments, and then we proceed with a set of hypotheses to be tested.

Literature Review

Reintegrative Shaming Theory

Braithwaite’s (1989) RST innovatively integrates various theoretical ingredients from theories of social control, subcultures, labeling, and social learning, providing a comprehensive framework concerning shaming, reintegration, and deviance at both macro- and micro-levels. The central thesis of RST emphasizes the differential effects of the societal reaction to a criminal act on offenders’ further involvement in criminal activities. The key concept in RST is shaming, which refers to social disapproval of deviant, delinquent, or criminal activities. It is conceived as “a means of making citizens actively responsible, of informing them of how justifiably resentful their fellow citizens are toward criminal behavior which harms them” (Braithwaite, 1989, p. 10). More specifically, shaming in RST is defined as “all social processes of expressing disapproval which have the intention or effect of invoking remorse in the person being shamed and/or condemnation by others who become aware of the shaming” (p. 100). It ranges from a subtle frown to verbal disapproval to an official pronouncement. Braithwaite (1989) acknowledges the important distinction between shaming (conveying lower respect in family, friends, and neighbors because of the offense) and guilt-induction (warning the offender how remorseful they should feel internally for their transgression) proposed in developmental psychology, but he considers them as “inextricably part of the same social process” (p. 57) in his theoretical framework. In the refinement of RST, the distinction between shaming as a regulatory practice and shame as an emotion is made (Ahmed et al., 2001). Existing research on RST has explored shaming practices by external agencies (Lu et al., 2002; Sakiyama et al., 2011; Tittle et al., 2003), shaming reactions perceived or expected by actual or potential wrongdoers (Rebellon et al., 2015; Ttofi & Farrington, 2008; L. Zhang & Zhang, 2004), and acknowledged or internalized feelings of shame/guilt (Braithwaite & Braithwaite, 2001; Fitch & Nazaretian, 2019; Svensson et al., 2013).

RST distinguishes two types of shaming—stigmatizing shaming and reintegrative shaming (Braithwaite, 1989). Stigmatizing shaming expresses disapproval of the transgression and labels the offender as deviant but involves no attempt to reconcile the offender with the community. The rejection and stigma foster resistance, push the offender to subcultures tolerating or supporting deviance and lead to a continuation of criminal and delinquent activities. Reintegrative shaming, in contrast, embeds social disapproval in bonds of respect and love. Signifying a deviant act is followed by forgiving the offender, identifying them as essentially good, and encouraging alternative behavior patterns. As a result, such shaming is more likely to promote reintegration and less likely to create further defiance and crime.

A critical point made in Braithwaite’s (1989) theory of shaming is that communitarianism and interdependency are the essential social conditions conducive to reintegrative shaming. A communitarian society features a dense network of individual interdependencies with strong mutual obligation, trust, and group loyalty. Like concepts such as attachment, bonding, and commitment in the social control theory, interdependency denotes to what degree individuals are enmeshed in conventional networks (i.e., family, church, school, etc.) and are dependent on one another. Strong attachment and deep commitment to these networks help develop individuals’ consciences of right and wrong, provide informal social control, and prevent new or further wrongdoing. People embedded in these communitarian societies deeply care about what their significant others think of them, and social disapproval from these external referents (e.g., friends, family, and neighbors) is usually potent to evoke morality and responsibility. As argued by Braithwaite, when individuals “are enmeshed in multiple relationships of interdependency,” they are “more susceptible to shaming” (p. 14), which helps individuals regulate their behaviors and inhibit misbehaving in the first place. Furthermore, the strength of interdependency depends on the nature, intensity, and quality of these intertwined relationships. Relationships built on love, trust, and respect are more likely to produce reintegrative shaming that is effective in controlling crime and promoting desistance. In short, RST proposes that interdependency and offending are theoretically associated while simultaneously highlighting the mediating role of reintegrative shaming.

It is worth noting RST’s direct attention to cultural differences in its theoretical formulation. The supremacy of American-European scholarship is a longstanding problem in criminology as well as in other social science disciplines (Aas, 2012). Nevertheless, the field of international and comparative criminology has begun to respond to the challenge and expanded the territory of criminological inquiry into non-Western settings. The recent few decades have witnessed burgeoning empirical application and assessment of Western-rooted criminological theories in Asian societies (Belknap, 2016; Messner, 2015; Suzuki et al., 2018). A vast majority of this body of research focuses on borrowing and adapting imported paradigms and ideas to evaluate if Western-based criminological theories can be applied and generalized to non-Western contexts. Most of these theories were formulated based on Western sociocultural assumptions, situations, and phenomena. Nevertheless, to decolonize criminology, the distinct features of non-Western contexts need to be included as vital and internal elements in theory and knowledge production (Braithwaite, 2015; J. Liu, 2016, 2021). Distinct from many other criminological theories, RST directs unique attention to different cultures and societies and provides a useful theoretical framework to compare and bridge the West and the East.

According to Braithwaite (1989, 2014), Asian societies such as Japan and China, largely enjoying higher levels of interdependency and communitarianism than their Caucasian counterparts, foster the development of reintegrative shaming. Western culture stresses individuality and privacy, and a person’s identity is primarily built upon the individual self. Individual freedom carries paramount importance, and people internalize the prevalent social norms and are expected to be responsible for their behaviors (O. Bedford & Hwang, 2003; Gert, 1988). In this sense, violation of norms and rules is conducive to guilt with a focus on individual failure to achieve one’s expected self rather than shame due to disappointment or rejection from others. By comparison, relationships profoundly define individual identity and shape morality in Confucian culture. In other words, harmony is much more valued than freedom in oriental societies. The conceptualization, establishment, and maintenance of the self inherently embrace roles in social relationships, responsibilities to others, and obligations to the common good (O. Bedford & Hwang, 2003). As such, living up to expectations from family and community is an indispensable part of keeping up one’s self-image. With a relationship-oriented identity, people in Asian societies have a strong desire to gain approval and respect from family, peers, and community members. They are prone to feeling shame when exposed to negative perceptions or actions from the external society, which significantly promotes rule compliance (O. A. Bedford, 2004; O. Bedford & Hwang, 2003; Chen, 2002). As such, shaming, born in collectivities such as family, school, company, and community, is an important and effective means of social control in these societies. As observed by Braithwaite, crime prevention and punishment practices in these societies often include components of reintegrative shaming, which emphasizes and promotes the protection, development, and restoration of social bonds and relationships with conventional institutions. (Braithwaite, 1989; Chen, 2002).

Prior Empirical Research

While the criminology field has witnessed an explosion of theoretical development and testing during the last several decades, surprisingly, Braithwaite’s RST has drawn relatively limited empirical attention. Although numerous studies address RST’s policy implications and applications in restorative justice (Braithwaite, 2002; van Ness & Strong, 2014; von Hirsch et al., 2003; Zehr, 2015), only a moderate body of empirical research has attempted to assess the major theoretical underpinnings of RST. A few of them test RST at the macro-level (Baumer et al., 2002; Schaible & Hughes, 2011), while the others examine its micro-social process. The micro-level studies centering on RST’s core argument—the theoretical association between shaming and offending—have produced, unfortunately, mixed findings. For example, L. Zhang and Zhang’s (2004) analyses of the National Youth Survey data showed that both parental forgiveness and peer disapproval were linked to less likelihood of youth involvement in predatory offenses, which partially supported RST. Murphy and Harris (2007) applied RST in the white-collar crime context. They found that reintegrative enforcement experience was associated with fewer subsequent offenses among taxpayers, which was partially mediated by shame-related emotions. By contrast, using adult survey data collected in Russia, two studies by Botchkovar and Tittle (2005, 2008) revealed that reintegrative shaming was positively associated with projected future deviance, contrary to RST’s predictions. In the same vein, the study on jail inmates by Tangney et al. (2011) found no evidence for the inhibitory function of shame-proneness. A careful examination of this body of research reveals that the problematic status of RST’s empirical assessment largely lies in the lack of data directly bearing on its theoretical hypothesis, as well as inconsistency in conceptualization and measurement across studies (Tittle et al., 2003).

A small number of studies directed their attention to the indirect path from interdependency to delinquency, testing whether this association is mediated by reintegrative shaming or stigmatization. In one of the initial attempts to directly operationalize and test RST, Hay (2001) found that parent-child interdependency was significantly associated with parents’ use of both shaming and reintegration, and the relationship was notably stronger for reintegration than for shaming. However, shaming had a durable negative impact on predicted predatory delinquency. Ahmed (2001) reported that shame management, including shame acknowledgment and shame displacement, partially mediated the relationship between predicted bullying and predictors such as family (e.g., parent-child affection, family disharmony), school (e.g., liking for school), and personality characteristics. Based on a sample of Australian adolescents, Losoncz and Tyson (2007) found that parent-child interdependency was strongly linked to parental reintegration and stigmatization, which had a significant impact on projected offending. Adopting a vignette-based methodology, the study on Cyprus adolescents (Ttofi & Farrington, 2008) revealed that mother bonding influenced children’s expectations of parental shaming and that disintegrative shaming was associated with sibling and peer bullying via shame management. However, no significant effect of reintegrative shaming was found. Father bonding had neither direct nor indirect effects. Additionally, treating interdependency as a moderator, Tittle et al. (2003) reported that interdependency had limited and contradictory effects in boosting the positive association between crime-generative shaming and predicted misbehavior. Several other studies examined the linkages among social and family bonding, shame/shaming, and crime or deviance, but their concepts and hypotheses were either framed in a general rational choice model of offending (Grasmick & Bursik, 1990; Svensson et al., 2013) or derived from multiple theoretical perspectives (Rebellon et al., 2010, 2015). None of these purports to assess RST.

Taken together, the small body of literature informed by RST has provided mixed empirical evidence concerning the associations between interdependency, shaming/shame, and offending or criminal intent. These findings lend partial support to the theoretical segments envisioned by RST. Nevertheless, RST’s theoretical scope is broader than what has been assessed. For example, the interdependency in RST includes not only one’s relationship with parents but also their bonding to school and community (Braithwaite, 1989). To further our understanding of RST’s explanation power, it is necessary to investigate the role of these social institutions in the assessment of RST’s key postulate regarding interdependency, shaming/shame, and offending.

Assessment of RST in the Chinese Context

Shaming and shame are particularly meaningful and important in China (O.A. Bedford, 2004; O. Bedford & Hwang, 2003; Braithwaite, 1989, 2014). Confucian cultures define individual identity in relations and emphasize one’s obligations to family and society. Acting immorally not only “loses the face” of oneself, but also “loses the face” of their family and community (H. C. Hu, 1994). Consequently, shaming plays a significant role in China’s formal and informal social control (Chen, 2002). Despite the expressive relevance of RST to Chinese society, the RST-informed China research is quite thin. Several studies applied RST to evaluate China’s social control and criminal justice practices (Chen, 2002; Lu, 1999; Vagg, 1998; L. Zhang et al., 1996; Y. Zhang, 2021), while only a few have attempted to assess RST’s theoretical hypotheses in the Chinese context. Based on in-depth interviews conducted in Hong Kong and Guangzhou, Wong (2001) found that the attitudes of forgiveness and respect from parents, teachers, and social authorities were strong protective factors that restrained Chinese adolescents from the onset of delinquency. The study by Lu et al. (2002) on an adult sample of Shanghai residents suggested that reintegration and stigmatization were both present in Chinese family shaming practices. Furthermore, a higher level of neighborhood interdependency was linked with stronger support for reintegrative shaming of neighbors’ deviance. These initial attempts show the value of RST in understanding crime and social control in non-Western settings. Given RST’s inclusion of both Western and non-Western realities and its proposed universal application (Braithwaite, 1989, 2015), further assessment of this theory’s propositions in the strategic context of Chinese society is much needed.

The Current Study

Compared with other mainstream theories (e.g., self-control theory, social bonding theory, and social learning theory), RST has received very limited empirical attention. This is surprising given that RST is one of the few, if not the only, mainstream criminological theories that indeed direct attention to the similarities and differences of diverse cultures in its theoretical formulation and development. A considerable portion of RST’s theoretical underpinnings remains unexplored. In addition, the existing literature has significant shortcomings, with studies employing inconsistent conceptualization and measurement of shaming and stigmatization and producing mixed findings. The insufficient evaluation of RST and its relatively marginalized status is unfortunate, reflecting perhaps the Western-centric state of contemporary criminology. It is thus critical for criminologists to take advantage of RST’s global engagement and contribute to the development of Southern criminology.

To this end, the current study seeks to address the shortcomings of prior research and test the main theses of RST. Utilizing a Chinese youth sample, this study explores the association between interdependency and the probability of wrongdoing and the critical role of perceived shame in mediating this association. Informed by the postulates of RST (Braithwaite, 1989), we develop a theoretical model depicted in Figure 1. Our main hypotheses include:

H1. A higher level of interdependency (e.g., attachment to conventional social institutions such as family, school, and neighborhood) promotes perceived shame.

H2. Perceived shame significantly reduces the probability of wrongdoing.

H3. Perceived shame mediates the impact of interdependency on the probability of wrongdoing. In other words, a higher level of interdependency directly decreases the probability of children’s wrongdoing through an elevated level of perceived shame.

Data and Methods

Data

The research site of the current study is a less-developed prefecture-level city in North China. The city’s population is about 4.3 million, among whom ethnic minorities account for approximately 13 percent, and slightly more than half are urban residents. A multi-stage cluster sampling technique was employed in this city-based youth survey. The study population consists of middle school and high school students in grades 7 through 12, aged 12 to 19. Among a total of 14 administrative divisions, three urban districts, one rural county, and one minority banner (a minority banner is a county-level administrative division in China, a term first used during the Qing Dynasty) were selected in the first sampling stage. Next, eight middle schools and four high schools located in these five divisions were randomly selected for study participation. The third stage of sampling involved randomly selecting forty classes within those schools. The students and their guardians were provided with a comprehensive explanation of the study’s objectives and procedures and were reassured of the strict maintenance of anonymity and the voluntary nature of participation. Students were informed that their involvement in the study was entirely optional and that they were under no obligation to participate. The self-administered survey was conducted in classrooms without the presence of any school staff or teacher during the entire survey sessions. Consequently, 1620 valid questionnaires were collected, with an extremely high response rate of 98.9%. High response rates are often seen in student surveys. For example, the student response rates of ISRD-3 largely ranged from 78 to 95% (e.g., 92% for both Serbia and Switzerland) (Marshall et al., 2015). The Chinese youth survey conducted byH. Zhang et al. (2014) achieved a response rate of 96.9%. Furthermore, the high response rates of self-reported surveys are commonly observed in out-of-school studies in China. For instance, L. Zhang et al. (2009) reported a 97% response rate in the resident survey on routine activities and criminal victimization. Research on survey responses has indicated that cultural differences can significantly impact response rates. The collectivist orientation prevalent in East Asian societies, including China, is likely a key factor contributing to the high response rates in their social surveys (Holtom et al., 2022).

Measures

Probability of Wrongdoing

Prior research has commonly used measures of predicted offending in their empirical assessment of RST (Hay, 2001; Losoncz & Tyson, 2007; L. Zhang & Zhang, 2004). Our primary dependent variable, probability of wrongdoing, is based on adolescent participants’ responses to two imaginary situations presented in the survey questionnaire. In the first situation, they owned a 2-year-old cell phone. They convinced a classmate that the old model was great but did not say that the new model was much better and less expensive. They were able to sell the old phone to a classmate for a price that allowed them to buy themselves the brand-new model. In the second scenario, the survey participants saw something (e.g., sports shoes, an expensive T-shirt, a CD, or perfume) in a big store that they had always wanted but could not afford. They took the item home without paying. For each vignette, participants are asked to rate how likely they will do it in real life. The response categories range from low to high: 1 = not at all, 2 = probably not, 3 = undecided, 4 = probably yes, and 5 = yes surely. These two self-rated likelihood items serve as manifest variables for the latent factor probability of wrongdoing.

Perceived Shame

The latent construct, perceived shame, is manifested by three composite scales—feelings of shame if caught by parents, teachers, or best friends. The questionnaire directs respondents to imagine they were caught shoplifting and asks how ashamed they would feel if their parents, teacher, and best friend, respectively, found out about it. The answer set includes 1 = no, not at all, 2 = yes, a little, and 3 = yes, very much. Similar questions are asked about the imaginary situations of being caught physically hurting another person and being arrested by the police for committing a crime. Responses for the three situations are summed up respectively to measure perceived shame if caught by parents, teachers, or best friends, with higher values on each indicating greater levels of shame. These three measures tap into the latent construct of perceived shame.

Interdependency

As noted, the three primary independent variables measuring interdependency in our hypotheses are family attachment, school attachment, and neighborhood attachment. The latent construct of family attachment combines the measures of parent-child relationship and parental supervision. The measure of the parent-child relationship is an aggregated scale based on 3 items—getting along well with the father, getting along well with the mother, and easily getting emotional support and care from parents (Cronbach’s alpha = .763). The survey asks respondents 12 questions about parental supervision. For example, how often their parents know where they are, what they are doing, and what friends they are with when they go out; how often their parents check their homework, tell them to be back home by a certain time in the evening, ask them to call when they go home late, and so on; and how often they tell parents how they spend their money and what they do in free time, and so on. We create a scale of parental supervision by summing up the scores of these 12 items. The scale has a Cronbach’s alpha of .887, and larger values represent higher levels of parental supervision.

Two manifest variables tap into the latent trait of school attachment. First, the variable of bonding to school is a composite scale based on 4 items—how much they would miss school if moving, like going to school most mornings, like their school, and think classes are interesting (Cronbach’s alpha = .816). Second, the variable of bonding with teachers is derived from the question asking how much they would miss their teacher if they had to move.

The survey asks adolescents a set of neighborhood-related questions. The results of our preliminary factor analysis show that these items load on two factors. The first factor indicating neighborhood closeness includes “many neighbors know me” and “people in my neighborhood often do things together” (Cronbach’s alpha = .701). The items aligning into the second factor, which captures neighborhood cohesion, include help, closeness, trust, and getting along among neighbors (Cronbach’s alpha = .914). These two factors serve as the manifest variables for the latent construct of neighborhood attachment.

Controlled Variables

Our structural equation modeling analyses incorporate several control variables that have been conventionally included in juvenile delinquency research. The attitude-behavior association has been widely evidenced in criminological literature (Rebellon et al., 2014; L. Zhang et al., 1996). In the revision of RST, Braithwaite and Braithwaite (2001) also point out that individual commitment to moral norms, including anti-criminal values, conditions the operation of shaming. Therefore, we control delinquent attitudes in our analyses. The latent construct, delinquent attitudes, combines two manifest scales—attitudes toward misdemeanors and attitudes toward crime. The questionnaire features eight items intended to measure attitudes toward delinquency and crime. Respondents report how wrong they think it is for someone of their age to commit each of the eight types of delinquency or crime. Our preliminary factor analysis suggests these eight items fall into two dimensions. The first dimension includes 1) lying, disobeying, or talking back to adults such as parents and teachers; 2) purposely damaging or destroying property that does not belong to them; 3) illegally downloading films or music from the internet; and 4) stealing something small like a chocolate bar from a shop. We create the scale of attitudes toward misdemeanors as a sum of the adolescent responses for these 4 items (1 = very wrong, 2 = wrong, 3 = a little wrong, 4 = not wrong at all). The Cronbach’s alpha of this scale is .754. The second scale gauges the other four more serious types of crime. They are 1) knowingly insulting someone because of his/her religion, skin color, or ethnic background; 2) breaking into a building to steal something; 3) hitting someone with the idea of hurting that person; and 4) using a weapon or force to get money or things from other people. The response scores for these 4 items are aggregated to form the scale of attitudes toward crime, yielding a Cronbach’s alpha of .847. For both scales, higher values indicate greater approvals of delinquent and criminal behaviors.

Gender is a dummy variable coded as 1 for boys and 0 for girls. Grade is a continuous variable ranging from 1 to 6 for 7th to 12th grade. Socioeconomic status is derived from the adolescents’ subjective evaluation regarding how their family is well off in comparison to others. Intact family is a dummy variable coded 1 for respondents whose father or mother passed away or whose parents are separated or divorced. Campus residence is also a dichotomous variable, coded 1 for students who attend a boarding school and live on campus.

We also control several variables that reflect distinctive features of the Chinese context. As noted, the research site is a city with a significant proportion of ethnic minority population in contrast to other regions in China. Hence, minority is included as a control variable, coded 1 for minority students. Social stratification in China has been institutionally established and shaped by its hukou system (Chan et al., 2018). The mass rural-to-urban migration over the past four decades has created 300 million migrants and nearly 70 million left-behind children, which significantly impacted children’s development and well-being in Chinese society (Chen, 2021; Liang & Li, 2021; Tong et al., 2019). Accordingly, we construct two variables to reflect hukou and migration status. The variable of hukou status is dichotomous, coded 1 for rural hukou. The variable of migrant parent differentiates respondents whose mother or father is a migrant worker in other cities (coded 1) from those who are not in a situation of being left behind (coded 0). Given the significant influence of the one-child policy on Chinese adolescent delinquency reported in prior literature (Bao & Haas, 2009; R. X. Liu et al., 2010), the final control variable we include is only child. Respondents who are the only child in their family are coded 1, while those who are not serve as the reference group.

Characteristics of the Sample

Table 1 presents the characteristics of the sample. Among the 1,620 adolescents, 49.4% are girls, 88.5% come from intact families, and 60.4% live on campus during school days. Meanwhile, 26.5% of them are ethnic minorities, 77.2% hold rural hukou, and 41.3% have one or two migrant parents. As mentioned earlier, our primary dependent latent factor, probability of wrongdoing, is measured by two observed variables: probability of deceptive cell phone trading and probability of shoplifting. A majority of the surveyed adolescents reported that they would “definitely not” or “probably not” commit these behaviors.

Table 1.

Sample Characteristics (n = 1,620).

	Mean	SD	α
Gender	0.506	0.500
Grade	2.990	1.627
SES	4.033	1.105
Intact family	0.115	0.319
Campus residence	0.604	0.489
Minority	0.265	0.442
Hukou	0.772	0.420
Migrant parent	0.413	0.493
Only child	0.405	0.491
Probability of deceptive cell phone trading	1.488	0.913
Probability of shoplifting	1.337	0.821
Perceived shame
Shame if caught by parents	8.717	0.826	.674
Shame if caught by teacher	8.690	0.878	.717
Shame if caught by friend	8.477	1.084	.704
Delinquent attitudes
Attitudes toward misdemeanor	5.954	2.131	.754
Attitudes toward crime	4.693	1.511	.847
Family attachment
Parent-child relationship	13.555	2.127	.763
Parental supervision	46.142	9.753	.887
School attachment
Bonding to school	12.271	2.890	.816
Bonding to teacher	4.717	1.200
Neighborhood attachment
Neighborhood closeness	6.460	1.592	.701
Neighborhood cohesion	13.349	3.034	.914

Analytical Strategy

We began our analyses by testing a measurement model to capture the relationships between latent constructs (e.g., family attachment) and their respective observed indicators. After demonstrating the validity of the measurement model, we proceeded to conduct structural equation modeling, testing the hypothetical dependencies among latent constructs based on path analysis. Because the distributions of our observed indicators are somewhat skewed, we used maximum likelihood with robust standard errors (MLR) to adjust for the non-normality in the data (Wang & Wang, 2019). Moreover, we utilized standard fit indices such as the Chi-square score, Comparative Fit Index (CFI), Root Mean Square Error of Approximation (RMSEA), and Standardized Root Mean Square Residual (SRMR) to assess whether the SEM model fits the data well (Hooper et al., 2008; L. Hu & Bentler, 1999). All analyses were performed using the statistical program Mplus8.6 (Wang & Wang, 2019).

Results

Measurement Model

Before testing whether the full structural equation model (Figure 1) fits the data well, we tested a measurement model (Table 2) to explore the relationships between each latent construct and its observed indicators. The results show that the Chi-square score was statistically significant, χ²(51) = 134.73, p < .01, suggesting a lack of model fit. However, the Chi-square test is sensitive to sample size and is not deemed appropriate when the sample size is large. In contrast, fit indices such as CFI (0.98), RMSEA (0.03, 90% CI [0.025, 0.038]), and SRMR (0.03) all suggested that this model fit the data extremely well (Hooper et al., 2008; L. Hu & Bentler, 1999). Table 2 presents the standardized factor loadings of observed indicators and correlations among latent constructs. All the observed indicators loaded reasonably well on their corresponding latent constructs, with standardized factor loadings ranging from 0.52 to 0.93. In addition, the correlations among the six latent constructs were all statistically significant in posited directions. First, the three measures of interdependency, as hypothesized, were moderately associated with the latent construct probability of wrongdoing, with a higher probability of wrongdoing significantly associated with weakened family attachment (β = −.42, p < .01), diminished school attachment (β = −.46, p < .01), and reduced neighborhood attachment (β = −.20, p < .01). Second, perceived shame—the hypothesized mediator in the theoretical model— had significant correlations with the three measures of interdependency (β = .26, p < .01 for family attachment; β = .29, p < .01 for school attachment; and β = .11, p < .01 for neighborhood attachment, respectively). As hypothesized, children’s feeling of shame and guilt was also significantly and negatively correlated with the probability of wrongdoing (β = −.35, p < .01). Third, the three measures of interdependency were significantly correlated with each other, with correlations ranging from .33 (the correlation between school attachment and neighborhood attachment) to .72 (the correlation between family attachment and school attachment). Finally, although not a focus of the study, children’s delinquent attitude—a control variable in the model—had weak to moderate correlations with measures of interdependency, children’s feelings of shame and guilt, and probability of wrongdoing. Children’s delinquent attitude had significant and negative associations with family attachment (β = −.52, p < .01), school attachment (β = −.47, p < .01), neighborhood attachment (β = −.27, p < .01), and perceived shame (β = −.31, p < .01), and a significant and positive association with the probability of wrongdoing (β = .33, p < .01).

Table 2.

Measurement Model: Standardized Factor Loadings and Correlations Among Latent Constructs (n = 1,620).

Latent construct	Factor loadings	Correlations among latent constructs
Probability of wrongdoing			1	2	3	4	5	6
Deceptive trading	0.59**	Probability of wrongdoing (1)	1
Shoplifting	0.53**	Perceived shame (2)	−0.35**	1
Perceived shame		Delinquent attitude (3)	0.33**	−0.31**	1
Caught by parents	0.88*	Family attachment (4)	−0.43**	0.26**	−0.52**	1
Caught by teacher	0.91**	School attachment (5)	−0.46**	0.29**	−0.47**	0.72**	1
Caught by friend	0.83**	Neighborhood attachment (6)	−0.20**	0.11**	−0.27**	0.41**	0.33**	1
Delinquent attitude
Toward misdemeanor	0.93**
Toward crime	0.68**
Family attachment
Parent-child relationship	0.54**
Parental supervision	0.67**
School attachment
Bonding to school	0.77**
Bonding to teacher	0.61**
Neighborhood attachment
Neighborhood closeness	0.89**
Neighborhood cohesion	0.52**

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

Full Model

After confirming that the observed indicators loaded well on their respective latent constructs,we proceeded to test the hypothesized theoretical model (Figure 1).The full model controlled for a series of individual and household characteristics, including gender, grade, family socioeconomic status, ethnicity, whether the respondent was the only child in the family, campus residence, intact family, hukou status, parental migration status, and children’s delinquent attitude. Overall, the full model fits the data well. The chi-square test of the overall model fit was statistically significant, χ²(119) = 374.23, p < .01; however, fit indices such as CFI = 0.95, RMSEA = 0.04, and SRMR = 0.03 suggested that the model fit the data satisfactorily (Hooper et al., 2008; L. Hu & Bentler, 1999). For the sake of a parsimonious presentation, we only reported the standardized coefficients of key independent variables (Figure 2), but details of the full model are available upon request.

Among the three measures of interdependency, school attachment (β = .13, p < .05) was the only measure significantly associated with a higher level of perceived shame, highlighting the critical role of the school as a formal institution in internalizing conventional social norms in China. Strong bonding with parents and neighborhood attachment increased children’s perceived shame; however, these associations were not statistically significant. As hypothesized in Figure 1, perceived shame served as an inhibitor against children’s involvement in delinquent and criminal activities, with a higher level of perceived shame and guilt significantly associated with a lower level of projected wrongdoing (β = −.20, p < .01). The SEM results also reveal a direct effect of school attachment, with children’s school bonding significantly and negatively associated with the probability of wrongdoings (β = −.24, p < .05), net of control of perceived shame and other theoretical correlates. Combined, these results highlight the critical role of the school as a formal social control institution in China; not only does school attachment directly decrease children’s delinquency and criminal involvement, but it also increases children’s feelings of embarrassment, guilt, and shame, ultimately reducing children’s delinquency involvement indirectly.

In terms of the effects of demographic variables and other controls, gender was statistically associated with the probability of wrongdoing, with boys reporting a higher level of delinquency engagement (β = .15, p < .01). Other control variables were not statistically associated with the probability of wrongdoing. Their effects, however, were likely to be mediated by children’s perceived shame. For example, children living off campus (β = .09, p < .05), children with a rural hukou (β = −.07, p < .05), or those with a higher level of socioeconomic status (β = −.07, p < .05) reported lower levels of perceived shame than their counterparts. Children’s delinquent attitude was also highly associated with perceived shame (β = −.19, p < .01), with higher levels of delinquent attitude significantly reducing children’s feelings of shame and guilt (results are not available in Figure 2 but available upon request).

Indirect Effects Analyses

To formally test whether the associations between interdependency and delinquency were mediated by children’s perceived shame and guilt, we calculated the total, direct, and indirect effects of the three interdependency measures on the probability of wrongdoing. Following Preacher and Hayes’s (2008) suggestion, we investigated the significance of the direct and indirect effects using the bias-corrected bootstrapping method with 5,000 resampling (Table 3).

Table 3.

Measures of Interdependency and Probability of Wrongdoing: Total, Direct, and Indirect Effects (n = 1,620).

	Lower 2.5%	B	Upper 2.5%
Family attachment
Total effects	−0.219	−0.074	0.060
Total direct effects	−0.214	−0.064	0.066
Total indirect effects	−0.027	−0.009	0.003
School attachment
Total effects	−0.121	−0.064	−0.014
Total direct effects	−0.114	−0.058	−0.007
Total indirect effects	−0.015	−0.006	−0.001
Neighborhood attachment
Total effects	−0.076	−0.015	0.048
Total direct effects	−0.081	−0.020	0.044
Total indirect effects	−0.004	0.005	0.015

Note. Bold numbers indicate a significant effect.

The indirect effect analyses confirm that school attachment indirectly influenced children’s probability of wrongdoing through increased feelings of shame and guilt. Results reveal that the total effect of children’s bonding with school and teachers on their probability of wrongdoing was statistically significant (b = −0.064, 95% CI [−0.121, −0.014]). Specifically, strong attachment to the school directly decreased children’s projected participation in delinquent activities (b = −0.058, 95% CI [−0.114, −0.007]), and indirectly decreased the probability of wrongdoing through the elevated feelings of shame and guilt (b = −0.006, 95% CI [−0.015, −0.001]). Although researchers have not reached a consensus on the best statistic to measure the effect size of an indirect effect, Wen and Fan (2015) suggest that the ratio of an indirect to the total effect should be preferred over other measures. As shown in Table 2, the indirect effect of school attachment through perceived shame accounted for about 10% of the total effect, suggesting that the specific pathway to delinquency through perceived shame is practically meaningful in the Chinese context.

Discussion and Conclusion

Although more than three decades have passed since Braithwaite’s (1989) book Crime, Shame and Reintegration was published, the empirical assessment of RST remains limited in both quantity and scope. By far only a moderate body of studies, based on inconsistent conceptualization and measurement, have examined a few segments of RST, yielding highly mixed findings. Utilizing a sample of Chinese adolescents, this study extends the focus beyond the shaming/shame-crime linkage and tests a broader scope of RST’s posits in a sociocultural context that features interdependency and shaming. We investigate the micro-level mechanism underlying the associations between interdependency shame, and wrongdoing, with a special focus on the mediation effects of children’s perceived shame. Notable findings concerning our three hypotheses are discussed below.

With respect to our first hypothesis, we find that adolescents’ attachment to school fosters the feeling of shame if their parents, teachers, and friends find out about their misconduct. However, we fail to find any evidence of the association between family attachment and shame reported in earlier studies (Losoncz & Tyson, 2007; Ttofi & Farrington, 2008). The effect of neighborhood attachment on perceived shame is not significant either. These results lend partial support to RST’s postulate about interdependency and shame/shaming and highlight the role of school in children’s development.

These statistically significant and null findings are likely shaped by the importance children attribute to the social networks within which they are embedded. Note that the study’s sample consisted of 7th to 12th graders. Research across various disciplines has documented the crucial role of schools in shaping adolescents’ identities and self-perceptions during the second decade of life (Kinney, 1993). This is particularly pertinent for Chinese adolescents for two reasons. First, education is of paramount importance for children and parents in China, with children’s daily activities revolving around school, homework, and peer students. Indeed, in many areas, children spend more than 10 hours each day in school. Second, in China’s exam-based education system, teachers are endowed with high levels of respect and authority and serve as important disciplinary agents. Any detected wrongdoing, even if it is not study-related, will likely lead to disciplinary action imposed by school officials and teachers. Everyday social interactions with teachers and peers in the educational setting reinforce adolescents’ definitions of moral values, social reality, and personal identity (Schwartz, 1987). By contrast, the impact of family on adolescents might be more distant compared with schools. For example, Song et al. (2009) reported that the quality of school peer attachment, instead of parental attachment, was preeminently associated with Chinese high school students’ self-esteem, self-liking, and self-competence.

In the same vein, children of this age group may not place great importance on the verbal or behavioral reactions of their neighbors to their deviant behavior, as much as they do on the feedback from their teachers and peers at school. Numerous studies have found that although the effect of neighborhood environment is significant, it is considerably smaller than the effects of other social control agents such as family, school, and peers (e.g., Le Blanc, 2006). Meanwhile, the dramatic social transformation in post-reform China has weakened the role of the neighborhood as an informal social control institution (Y. Liu et al., 2017; Whyte & Parish, 1984; Zhuo, 2012). The traditional neighborhoods, usually centered on work units or kin villages and featuring close ties, frequent contacts, and watchful eyes, have declined significantly in contemporary China. As Braithwaite (1989) points out (pp. 171–173), the locus of interdependency in modern societies may shift from a spatially localized neighborhood to a community of interest. Compared with residential neighborhoods, schools are undoubtedly more of a community of interest for adolescents. Our results indicate that the school community is more important than the neighborhood for Chinese adolescents nowadays.

These findings carry important implications for policymakers and program planners in the fields of adolescent development and juvenile justice. It is evident that Chinese children attach immense importance to how they are perceived by their teachers and peers within the school environment. A strong attachment to school contributes to heightened feelings of shame regarding wrongful actions. Policymakers and educators should acknowledge that schools serve as powerful agents of socialization, shaping the attitudes, behaviors, and prospects of young individuals. Interventions aimed at enhancing the connections between students and educational institutions can promote reintegrative shaming and offer promising avenues for crime prevention and desistance (H. Zhang, 2023). Furthermore, policymakers should consider measures that bridge the gap between school and home, reinforcing the collaboration between families, educators, and communities.

Providing empirical support for the second hypothesis, our results indicate that perceived shame significantly reduces the probability of wrongdoing. In other words, the higher the perceived shame, the lower the probability of wrongdoing. This finding is consistent with RST’s prediction and echoes relevant empirical studies (Braithwaite, 1989; Rebellon et al., 2015; Tibbetts & Herz, 1996; L. Zhang & Zhang, 2004). For example, Rebellon et al. (2015) found that anticipated shaming is a significant predictor of crime intent, especially among girls. As discussed earlier, shaming/shame is an essential part of social control in Asian societies including China. With relationship-shaped self-identity, people are concerned about losing honor among relational others if they have done wrong. The imagined disapproval of others and feelings of shame, occurring in anticipation of perpetration, act as a general deterrent that refrains people from doing wrong (O. Bedford & Hwang, 2003; Braithwaite, 1989; Harris, 2001).

Our third hypothesis focuses on an often-neglected hypothesis in Braithwaite’s (1989) RST theory, the mediating role of shame. Our indirect effect analysis reveals that the effect of school attachment on the probability of wrongdoing is partially mediated by children’s perceived shame. This finding lends support to RST’s postulates that shame mediates the effect of interdependency on crime (Braithwaite, 1989; Losoncz & Tyson, 2007). Strong attachment to school conduces to shame, which improves rule compliance. Meanwhile, closeness to school still directly predicts the probability of wrongdoing. Nevertheless, we fail to find any significant direct or indirect effect of family attachment and neighborhood attachment. Taken together, these results further highlight the prominent role of the school in Chinese children’s lives, both as an educational and as a social control institution. Schools may be even more salient than other conventional institutions for adolescents, who start establishing social bonds beyond family relations and developing personal identity.

Cautions should be exercised in the interpretation of the results. We acknowledge that the survey was conducted in schools, which could potentially bias students’ response towards emphasizing the importance of school. Furthermore, we rely on cross-sectional data to assess our hypotheses. The directional paths between factors are not certain, and no causal conclusions can be drawn. For example, we find delinquent attitudes are significantly associated with perceived shame. To make the analysis focused and parsimonious, delinquent attitudes are treated as a controlled variable in our model. RST addresses the linkage between shaming and moral conscience building (Braithwaite, 1989). Earlier research also provides empirical evidence regarding the association between reintegrative shaming in a school context and adolescents’ antisocial attitudes (Lee & Kavanaugh, 2015). Longitudinal studies could further investigate the process among interdependency, shaming, moral values and conduct norms, and misbehavior over time.

It should be noted that this study adopts a scenario approach, in which the key variables are defined via perceived shame and self-evaluated probability of wrongdoing. Although intentions and actions are usually highly related, expected feelings and behaviors do not always match situations in the real world (Exum & Bouffard, 2010). Furthermore, the conceptualization of shame in our study is from an internal perspective. Earlier studies focus on either external shaming practice (Lu et al., 2002; Tittle et al., 2003) or internal shame feelings (Fitch & Nazaretian, 2019). Future research may be directed toward examining the relationship between the reintegrative/stigmatizing shaming practice and the internal sensation of shame (Harris, 2001).

In addition, although not the focus of the current study, our results reveal significant differences in both perceived shame and probability of wrongdoing between boys and girls. Existing studies have reported highly mixed findings regarding gender effects in shaming/shame and offending (Fitch & Nazaretian, 2019; Rebellon et al., 2015; Walters, 2022). Further research is needed for a better understanding of gender in RST’s theoretical nexus, particularly in Asian settings. Lastly, like most antecedent empirical assessments of RST, this study is conducted at the micro-level with individual adolescents as the unit of analysis. More research is needed to test the macro-level portions of RST. A more interesting avenue might be to explore the interplay between contextual-level communitarianism and individual-level shaming/shame by taking advantage of multilevel analysis (Braithwaite, 1989).

The theoretical framework of RST provides important insight as regards associations between interdependency, shaming/shame, and offending. The current study attempts to empirically assess this mechanism in China, a social setting featuring interdependency and communitarianism. Consistent with RST’s predictions, this study reveals a significant direct deterrent effect of shame on wrongdoing. Meanwhile, attachment to school diminishes the probability of wrongdoing both directly and indirectly via shame. Adolescents situated in greater attachment and commitment to school are more susceptible to shame, which refrains them from doing wrong. These findings add to our understanding of RST’s explanation power. Additionally, they highlight the critical role of school in adolescent development. By closing, we call for further empirical assessment of RST in various social contexts as well as more attention to schools in crime prevention.

Footnotes

Declaration of Conflicting Interests

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

Funding

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

ORCID iDs

Yue Zhuo

Xiaojin Chen

Xin Jiang

Author Biographies

Yue Zhuo is an associate professor in the Department of Sociology and Anthropology at St. John’s University. Her research interests include crime and victimization, juvenile delinquency, drug abuse and mental health.

Xiaojin Chen is an associate professor in the Department of Sociology at Tulane University. His research interests include migration and children’s development in China, life course criminology, victimization, research methods, and social statistics.

Jianhong Liu is a distinguished professor in the Faculty of Law at the University of Macau. His research interests include international and comparative criminal law and justice, crime control and prevention, research methods, and social statistics.

Xin Jiang is an assistant professor in the Department of Sociology at Tulane University. Her research interests include juvenile delinquency, criminological theory, crime control, and immigrant assimilation.

References

Aas

K. F.

(2012). ‘The earth is one, but the world is not’: Criminological theory and its geopolitical divisions. Theoretical Criminology, 16, 5–20.

Ahmed

(2001). Shame management: Regulating bullying. In Ahmed

Harris

Braithwaite

(Eds.), Shame management and regulation (pp. 211–311). Cambridge University Press.

Ahmed

Harris

Braithwaite

(Eds.). (2001). Shame management and regulation. Cambridge University Press.

Bao

Haas

(2009). Social change, life strain, and delinquency among Chinese urban adolescents, Sociological Focus, 42(3), 285–305.

Baumer

E. P.

Wright

Kristinsdottir

Gunnalaugsson

(2002). Crime, shame, and recidivism: The case of Iceland. British Journal of Criminology, 41, 40–59.

Bedford

Hwang

K.-K.

(2003). Guilt and shame in Chinese culture: A cross-cultural framework from the perspective of morality and identity. Journal for the Theory of Social Behaviour, 33(2), 127–144.

Bedford

O. A.

(2004). The individual experience of guilt and shame in Chinese culture. Culture & Psychology, 10(1), 29–52.

Belknap

(2016). Asian criminology’s expansion and advancement of research and crime control practices. Asian Journal of Criminology, 11, 249–264.

Botchkovar

E. V.

Tittle

C. R.

(2005). Crime, shame, and reintegration in Russia. Theoretical Criminology, 9(4), 401–442.

10.

Botchkovar

E. V.

Tittle

C. R.

(2008). Delineating the scope of reintegrative shaming theory: An explanation of contingencies using Russian data. Social Science Research, 37, 703–720.

11.

Braithwaite

(1989). Crime, shame, and reintegration. Cambridge University Press.

12.

Braithwaite

(2002). Setting standards for restorative justice. The British Journal of Criminology, 42(3), 563–577.

13.

Braithwaite

(2014). Crime in Asia: Toward a better future. Asian Journal of Criminology, 9, 65–75.

14.

Braithwaite

(2015). Rethinking criminology through radical diversity in Asian reconciliation. Asian Journal of Criminology, 10, 181–183.

15.

Braithwaite

(2020). Restorative justice and reintegrative shaming. In Chouhy

Cochran

J. C.

Jonson

C. L.

(Eds.), Criminal justice theory: Explanation and effects (Vol. 26, pp. 281–308). Routledge.

16.

Braithwaite

Ahmed

Braithwaite

(2008). Shame, restorative justice, and crime. In Cullen

Wright

Belvins

(Eds.), Taking stock: The status of criminological theory, advances in criminological theory (Vol. 15, pp. 397–417). Transaction Publishers.

17.

Braithwaite

(2001). Shame and shame management. In Ahmed

Harris

Braithwaite

(Eds.), Shame management and regulation (pp. 3–69). Cambridge University Press.

18.

Chan

K. W.

Cai

Wan

Wang

(2018). Urbanization with Chinese characteristics: The hukou system and migration. Routledge.

19.

Chen

(2002). Social control in China: Applications of the labeling theory and the reintegrative shaming theory. International Journal of Offender Therapy and Comparative Criminology, 46(1), 45–63.

20.

Chen

(2021). Parental migration and children’s problem behaviours in rural China: Testing an integrative theoretical model. British Journal of Criminology, 61(6), 1592–1611.

21.

Exum

M. L.

Bouffard

J. A.

(2010). Testing theories of criminal decision making: Some empirical questions about hypothetical scenarios. In Piquero

A. R.

Weisburd

(Eds.), Handbook of Quantitative Criminology (pp. 581–594). Springer.

22.

Fitch

C. H.

Nazaretian

(2019). Examining gender differences in reintegrative shaming theory: The role of shame acknowledgment. Crime, Law, and Social Change, 72, 527–546.

23.

Gert

(1988). Morality: Its nature and justification. Oxford University Press.

24.

Grasmick

H. G.

Bursik

R. J.

(1990). Conscience, significant others, and rational choice: Extending the deterrence model. Law & Society Review, 24(3), 837–861.

25.

Harris

(2001). The relationship between shame and shaming. In Ahmed

Harris

Braithwaite

(Eds.), Shame management and regulation (pp. 157–176). Cambridge University Press.

26.

Hay

(2001). An exploratory test of Braithwaite’s reintegrative shaming theory. Journal of Research in Crime and Delinquency, 38(2), 132–153.

27.

Holtom

Baruch

Aguinis

A Ballinger

(2022). Survey response rates: Trends and a validity assessment framework. Human Relations, 75(8), 1560–1584.

28.

Hooper

Coughlan

Mullen

(2008). Structural equation modelling: Guidelines for determining model fit. The Electronic Journal of Business Research Methods, 6(1), 9.

29.

H. C.

(1994). The Chinese concepts of “face.” American Anthropologist, 46(1), 45–64.

30.

Bentler

P. M.

(1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling: A Multidisciplinary Journal, 6(1), 1–55.

31.

Kinney

D. A.

(1993). From nerds to normals: The recovery of identity among adolescents from middle school to high school. Sociology of Education, 66(1), 21–40.

32.

Latimer

Dowden

Muise

(2005). The effectiveness of restorative justice practices: A meta- analysis. The Prison Journal, 85(2), 127–144.

33.

Le Blanc

. (2006). Self-control and social control of deviant behavior in context: Development and interactions along the life course. In Wikström

P. O. H.

Sampson

R. J.

(Eds.), The explanation of crime context, mechanisms, and development (pp. 124–151). Cambridge University Press.

34.

Lee

Kavanaugh

P. R.

(2015). Discipline, shaming and antisocial attitude in Philadelphia middle schools. Sociological Spectrum, 35(6), 504–517.

35.

Liang

(2021). Migration and children in China: A review and future research agenda. Chinese Sociological Review, 53, 312–331.

36.

Liu

(2016). Asian paradigm theory and access to justice. Journal of Contemporary Criminal Justice, 32(3), 205–224.

37.

Liu

(2021). Asian criminology and non-western criminology: Challenges, strategies, and directions. International Annals of Criminology, 59(2), 103–118.

38.

Liu

R. X.

Lin

Chen

Z.-y.

(2010). The Effect of parental responsiveness on differences in psychological distress and delinquency between singleton and non-singleton Chinese adolescents. Journal of Child and Family Studies, 19, 547–558.

39.

Liu

(2017). Changing neighbourhood cohesion under the impact of urban redevelopment: A case study of Guangzhou, China. Urban Geography, 38(2), 266–290.

40.

Losoncz

Tyson

(2007). Parental shaming and adolescent delinquency: A partial test of reintegrative shaming theory. The Australian and New Zealand Journal of Criminology, 40(2), 161–178.

41.

(1999). Bang Jiao and reintegrative shaming in China’s urban neighborhoods. International Journal of Comparative and Applied Justice, 23, 115–125.

42.

Zhang

Miethe

T. D.

(2002). Interdependency, communitarianism and reintegrative shaming in China. The Social Science Journal, 39, 189–201.

43.

Marshall

I. H.

Enzmann

Hough

Killias

Kivivuori

Steketee

(2015). Youth victimization and reporting to police: First results from the third round of the International Self-Report Delinquency Study (ISRD3). https://web.northeastern.edu/isrd/wp-content/uploads/2016/01/ISRD3_TechRep_03.pdf

44.

Messner

(2015). When west meets east: Generalizing theory and expanding the conceptual toolkit of criminology. Journal of Asian Criminology, 10(2), 117–129.

45.

Murphy

Harris

(2007). Shaming, shame and recidivism: A test of reintegrative shaming theory in the white-collar crime context. British Journal of Criminology, 47(6), 900–917.

46.

Preacher

K. J.

Hayes

A. F.

(2008). Asymptotic and resampling strategies for assessing and comparing indirect effects in multiple mediator models. Behavior Research Methods, 40(3), 879–891.

47.

Rebellon

C. J.

Manasse

M. E.

Van Gundy

K. T.

Cohn

E. S.

(2014). Rationalizing delinquency: A longitudinal test of the reciprocal relationship between delinquent attitudes and behavior. Social Psychology Quarterly, 77(4), 361–386.

48.

Rebellon

C. J.

Piquero

N. L.

Piquero

A. R.

Tibbetts

S. G.

(2010). Anticipated shaming and criminal offending. Journal of Criminal Justice, 38(5), 988–997.

49.

Rebellon

C. J.

Wiesen-Martin

Piquero

N. L.

Piquero

A. R.

Tibbetts

S. G.

(2015). Gender differences in criminal intent: Examining the mediating influence of anticipated shaming. Deviant Behavior, 36, 17–41.

50.

Sakiyama

Liang

(2011). Reintegrative shaming and juvenile delinquency in Japan. Asian Journal of Criminology, 6, 161–175.

51.

Schiable

L. M.

Hughes

L. A.

(2011). Crime, shame, reintegration, and cross-national homicide: A partial test of reintegrative shaming theory. The Sociological Quarterly, 52(1), 104–131.

52.

Schwartz

(1987). Beyond conformity or rebellion: Youth and authority in America. University of Chicago Press.

53.

Song

Thompson

R. A.

Ferrer

(2009). Attachment and self-evaluation in Chinese adolescents: Age and gender differences. Journal of Adolescence, 32(5), 1267–1286.

54.

Suzuki

Pai

C.-F.

Islam

Md. J

. (2018). Systematic quantitative literature review on criminological theories in Asia. Asian Journal of Criminology, 13(2), 129–151.

55.

Svensson

Weerman

F. M.

Pauwels

L. J. R.

Bruinsma

G. J. N

Bernasco

(2013). Moral emotions and offending: Do feelings of anticipated shame and guilt mediate the effect of socialization on offending? European Journal of Criminology, 10(1), 22–39.

56.

Tangney

J. P.

Stuewig

Mashek

Hastings

(2011). Assessing jail inmates’ proneness to shame and guilt: Feeling bad about the behavior of the self? Criminal Justice and Behavior, 38, 710–734.

57.

Tibbetts

S. G.

Herz

D. C.

(1996). Gender differences in factors of social control and rational choice. Deviant Behavior, 17(2), 183–208.

58.

Tittle

C. R.

Bratton

Gertz

M. G.

(2003). A test of a micro-level application of shaming theory. Social Problems, 50(4), 592–617.

59.

Tong

Yan

Kawachi

(2019). The factors associated with being left-behind children in China: Multilevel analysis with nationally representative data. PLoS One, 14(11), e0224205.

60.

Ttofi

M. M.

Farrington

D. P.

(2008). Reintegrative shaming theory, moral emotions and bullying. Aggressive Behavior, 34, 352–368.

61.

Umbreit

M. S.

Vos

Coates

R. B.

Lightfoot

(2005). Restorative justice in the twenty-first century: Social movement full of opportunities and pitfalls. Marquette Law Review, 89(2), 251–304.

62.

Vagg

(1998). Delinquency and shame: Data from Hong Kong. The British Journal of Criminology, 38(2), 247–264.

63.

van Ness

D. W.

Strong

K. H

. (2014). Restoring justice: An introduction to restorative justice (5th ed.). Routledge.

64.

von Hirsch

Roberts

J. V.

Bottoms

A. E.

Roach

Schiff

. (Eds.). (2003). Restorative justice and Criminal Justice: Competing or reconcilable paradigms. Bloomsbury Publishing.

65.

Walters

G. D.

(2022). Parents and best friends perceived reaction to participant deviance and self-reported delinquency: Moderation by sex and mediation by youth attitudes toward deviance in mid- and late adolescents. Criminal Justice Review, 47(3), 302–317.

66.

Wang

(2019). Structural equation modeling: Applications using Mplus (2nd ed.). Wiley.

67.

Wen

Fan

(2015). Monotonicity of effect sizes: Questioning kappa-squared as mediation effect size measure. Psychological Methods, 20(2), 193.

68.

Whyte

M. K.

Parish

W. L.

(1984). Urban life in contemporary China. University of Chicago Press.

69.

Wong

D. S. W.

(2001). Pathways to delinquency in Hong Kong and Guangzhou (South China). International Journal of Adolescence and Youth, 10, 91–115.

70.

Zehr

(2015). The little book of restorative justice: Revised and updated. Simon and Schuster.

71.

Zhang

(2023). Restorative practice in school: Reflections from an East Asian perspective. The International Journal of Restorative Justice, 6(2), 179–190.

72.

Zhang

Zhao

Ren

(2014). Social attachment and juvenile attitudes toward the police in China: Bridging Eastern and Western wisdom. Journal of Research in Crime and Delinquency, 51(6), 703–734.

73.

Zhang

Messner

Liu

Zhuo

(2009). Guanxi and fear of crime in contemporary urban China. British Journal of Criminology, 49, 472–490.

74.

Zhang

(2004). Reintegrative shaming and predatory delinquency. Journal of Research in Crime and Delinquency, 41(4), 433–453.

75.

Zhang

Zhou

Messner

S. D.

Liska

A. E.

Krohn

M. D.

Liu

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.

76.

Zhang

(2021). Police discretion and restorative justice in China: Stories from the street-level police. International Journal of Offender Therapy and Comparative Criminology, 65(4), 498–520.

77.

Zhuo

(2012). Social capital and satisfaction with crime control in urban China. Asian Journal of Criminology, 7(2), 121–136.