Explaining Chinese Delinquency: Self-Control,Morality,and Criminogenic Exposure

Abstract

This study extends the testing of situational action theory (SAT) to a Chinese population, and sheds new light on the directions of the moderation relationships between self-control and morality, and between crime propensity and criminogenic exposure on delinquency. Relying on a large, representative sample of middle school students from two areas of Guizhou, China (N = 2,498), we find that both self-control and morality have significant inhibiting effects on delinquency. Moreover, self-control has a more profound curbing effect on delinquency among adolescents with higher levels of morality. Meanwhile, the promoting effect of crime propensity on delinquency decreases when levels of risky exposure increase. When adolescents have more unsupervised activities and delinquent peers, their crime propensity affects delinquency to a lesser extent. Our study confirms that individual and situational factors interlock in determining delinquency, and reiterates the value of empirical testing across cultures to validate and possibly improve general theories of crime.

Keywords

situational action theory self-control morality China delinquent peers unsupervised socializing

Generations of criminologists have pinpointed the centrality of an individual–environment interplay in shaping crime and delinquency (Bernard et al., 2010). Arguably, one of the most tested and supported individual explanations of crime is Gottfredson and Hirschi’s (1990) General Theory of Crime, positing that the ability to exercise self-control is a critical trait of a person’s crime propensity. A recent stream of research based on situational action theory (SAT, Wikström, 2004) has meaningfully expanded the understanding of crime propensity by including the dimension of morality, and brought the interaction between crime inclination and environmental risks back to a full discussion. As a promising general theory, SAT has gained research momentum in the United Kingdom, and across North American and European countries, generating insights into how self-control, morality, and criminogenic exposure interact in shaping crime (Pauwels et al., 2018). Missing from this literature, however, is East Asia, an area that has deep-seated cultural and social differences from Western countries. With strong emphases on individual capacity to self-discipline, morality, and interdependence between individual and context, China represents a critical geographic area for testing the generalizability of SAT.

Drawing upon a large, probability sample of middle school students from two areas of Guizhou, China, this study adds to the SAT literature in important ways. Unlike previous SAT studies, we distinguish between violent and property delinquencies, as the two differ in nature and may be associated with varying explanatory factors and mechanisms. We also take the initiative to simultaneously examine the interaction effects between self-control and morality, and between crime propensity and risky exposure in a single study. Although we did not test the third proposition of SAT regarding the interactional effect between deterrence and crime propensity in shaping crime, our study has advantageously captured two key propositions of SAT.

This study also contributes to the literature on the delinquency of Chinese adolescents. Whereas a number of studies have assessed the role of self-control on delinquency in the Chinese context (N. W. Cheung & Cheung, 2008; Chui & Chan, 2016; Jiang et al., 2020; Li et al., 2014; Lu et al., 2013; Pyrooz & Decker, 2013), little attention has been paid to the matter of morality or the interplay between self-control and morality in determining delinquency. Also missing is an evaluation of the potential interaction effects between individual traits and environmental factors. External factors such as peer delinquency and risky lifestyles may not only influence juvenile delinquency independently from crime propensity traits such as low self-control (e.g., N. W. Cheung & Cheung, 2008; Jiang et al., 2020), but also, importantly, condition the effects of crime propensity on delinquency. To fill these gaps, this study incorporates morality as an indispensable dimension of crime propensity, and examines the nexus between propensity toward crime, criminogenic exposure, and delinquency by uncovering potential moderating relationships.

Crime Propensity, Risky Exposure, and Crime

SAT

People commit crime often out of a synergy of how well they can control themselves, how strong their prosocial values are, and how many criminogenic situations they are subject to (Wikström, 2004). Self-control has been a popular concept in criminological theorizing and empirical testing (Pratt & Cullen, 2000; Vazsonyi et al., 2017), with low self-control including such characteristics as being impulsive, insensitive, physical, risk-taking, shortsighted, and nonverbal (Gottfredson & Hirschi, 1990). Morality, different from self-control, refers to normative values and rules that are socially desirable and culturally transmissible (Hofmann et al., 2018). Morality regards what is right or wrong, and serves as another major constraint on delinquent impulsivity. Relatively understudied in criminology, morality is highlighted in a recent line of research that explores its interaction with self-control in shaping criminality (e.g., Antonaccio & Tittle, 2008; Wikström & Svensson, 2010). Dominating this nascent body of research is SAT, positing that morality not only independently affects delinquency but also moderates the influence of self-control on delinquency (Wikström, 2004). Specifically, self-control matters more for individuals who have lower levels of morality, as they count on their ability of self-control to escape from the temptation of crime. For those with higher levels of morality, however, desires or urges to follow conventional rules can effectively restrain them from delinquent acts, regardless of their levels of self-control (Wikström & Svensson, 2010).

SAT further postulates that self-control and morality interlock and constitute crime propensity, the tendency to view and choose crime as a behavior option (Wikström & Svensson, 2010). Crime propensity, however, is only part of the crime equation, with the other part being criminogenic exposure that highlights the role that environment plays in crime involvement. On one hand, when people are in situations with increased risks of delinquency, such as associating with delinquent peers and engaging in unsupervised leisure activities, their ability to exercise self-control and their will to act in line with conventional morals can help them make right choices (Schepers & Reinecke, 2018). On the other hand, notwithstanding levels of propensity, physical and social settings that often connect to risky lifestyles and activities are conducive to crime (Wikström & Svensson, 2010). Moreover, growing evidence demonstrates an interaction effect between crime propensity and criminogenic exposure, with the influence of exposure amplifying when propensity toward crime intensifies (Pauwels et al., 2018). In other words, people who have stronger crime propensity are subject to a greater influence of risky situations on their engagement in delinquent behaviors.

Empirical Evidence on Self-Control × Morality

A large amount of evidence has supported the significance of self-control, and to a lesser extent, morality in shaping crime (Stams et al., 2006; Vazsonyi et al., 2017). Recently, scholars have examined the interactive effects of self-control and morality on crime and delinquency, with three alternative speculations proposed, each receiving some empirical support. First, the inhibiting effect of self-control on delinquency becomes smaller when morality levels get higher (e.g., Pauwels et al., 2018). This is the main proposition of SAT which suggests that individuals with strong moral values do not need to rely on self-control to refrain from delinquent activities (Wikström & Svensson, 2010). Alternatively, the preventive effect of self-control on delinquency becomes larger when moral values get stronger (Antonaccio & Tittle, 2008). This means that self-control and morality have a mutual amplification effect on one another. Third, self-control and morality do not moderate each other (Gallupe & Baron, 2014).

No research has assessed the potential interactive effects of self-control and morality on delinquency among Chinese children. Several studies have investigated their respective effects, yielding mixed results. Some found curbing effects of self-control (Jiang et al., 2020; Pyrooz & Decker, 2013) and moral values (Liu & Liu, 2016; Ngai & Cheung, 2005) on crime, while others found no such effects when measures such as social bonds were included in analyses (Lu et al., 2013). The inclusion of different sets of variables and the use of different measurements likely explain these discrepancies in findings. For example, Jiang and colleagues (2020) incorporated unsupervised activities and caretaker monitoring as control variables, while Lu et al. (2013) controlled for neighborhood bonding, involvement, and belief. While Lu et al. (2013) used 12 items from Grasmick et al. (1993) to cover dimensions of impulsivity, risk-seeking, self-centeredness, and temper of self-control, Pyrooz and Decker (2013) used 11 items from the Brief Self-Control Scale to capture dimensions of healthy habits, work ethic, and impulsivity.

Empirical Evidence on Crime Propensity × Criminogenic Exposure

A great deal of research has documented the interplay between personal traits and environmental factors in shaping crime, and this section focuses on that between crime propensity (i.e., self-control, morality, or an integration of both under the SAT framework) and criminogenic exposure (i.e., unsupervised activities and delinquent peers). The evidence, overall, is equivocal. Some found that peer delinquency has a stronger effect among individuals who have higher levels of self-control (Meldrum et al., 2009; Vitulano et al., 2010), and similarly, risk-seeking tendency has a more harming influence on violence among adolescents who have less unstructured socializing (LaGrange & Silverman, 1999). In contrast, most of the SAT literature revealed the opposite that the delinquency-promoting effect of risky exposure increases when crime propensity grows (Pauwels et al., 2018). Still others found no interaction between self-control and unstructured/unsupervised socializing (Hoeben et al., 2016; LaGrange & Silverman, 1999; Maimon & Browning, 2010) or peer delinquency (McGloin & O’Neill Shermer, 2009).

Although no research has explored the interactive effects of crime propensity and situational risks on delinquency among Chinese, scholars have examined these two domains of predictors independently. Some found that self-control is not predictive of delinquency when variables such as delinquent peers were included in the analysis among Hong Kong adolescents (N. W. Cheung & Cheung, 2008). Others revealed that both self-control and peer delinquency persist as significant predictors of delinquency among samples of children in both rural and urban China (Jiang et al., 2020; Ngai & Cheung, 2005). Furthermore, evidence suggests that unsupervised activities influence delinquency after controlling for self-control and peer association (Jiang et al., 2020). In brief, these inconsistent findings regarding the influence of self-control, morality, and criminogenic exposure on delinquency, and the lack of scrutiny on potential moderating relationships between these variables in China prompt the current study.

The Chinese Context

China provides an interesting context to study the subject matters of self-control, morality, and delinquency. Morality and self-control are deeply intertwined in Chinese culture, as the long Confucianism tradition promotes governance through morality (Huang, 2015). For the recent 70 years, the Communist ideology has been promoted and dominant. At the heart of Chinese Communism is support for the ruling of the Communist Party, but Communist ideology also shares many commonalities with Confucianism in regulating behaviors. The current President Xi, on multiple occasions, stressed the importance of preserving and promoting Chinese traditional culture in realizing the Chinese dream of national rejuvenation. Indeed, both Confucianism and Chinese Communism are highly moralistic and controlling. Together, they create a politically moralistic culture for Chinese people, including adolescents.

The efficacy of relying on self-control and morality to govern and promote prosocial behaviors however is unknown. Since the late 1970s, with the start of the open-door policy and market economy, the Chinese economy has expanded exponentially, making the nation the second largest economic power in the world. In 2020, China has 389 billionaires, second only to the United States (Ponciano, 2020). At the other end of the spectrum are 600 million Chinese who earn less than $140 per month (Yuan, 2020). A widening class gap escalates domestic tensions and conflicts, yet conventional moral guidance and behavioral norms are losing their regulating power over the populace. Large-scale urbanization and migration have also undermined the efficacy of traditional control mechanisms in shaping behaviors. Cultural conflicts and anomie are prevalent in China (Zhao et al., 2019), creating an ideological crisis (Kwong, 1994).

Within this background, crime rates, albeit generally low, have dramatically increased in post-reform China (Bakken, 2004; Cheng et al., 2017; Song et al., 2019). Crime rates for youths under 18 have also risen while the age at which delinquency prevalence peaks has decreased (C. Cheung & Liu, 2015). To react, besides waves of strike-hard-against-crime campaigns, Chinese leaders have proposed multiple versions of governing by morality, such as former President Jiang’s “Rule by Morality,” former President Hu’s “Eight Virtues and Eight Shames,” and current President Xi’s “Moral Construction of Citizens in the New Era” (Kubat, 2018). Research on delinquency among Chinese has also grown during the past 25 years, testing a variety of theories including strain, learning, social control, self-control, social disorganization, psychological, labeling, and cultural transmission and subculture (Weng et al., 2016; L. Zhang, 2008). The theoretical interaction between self-control and morality, however, has not been explored. The cultural anomie and moral quandary that many Chinese people face today bring the interplay between self-control, morality, risky environment, and delinquency to the focus of this research.

The Current Study

Drawing upon survey data collected from 2,498 middle school students in seventh to ninth grades (typically 12–15 years old) from two areas of Guizhou Province, this study investigates the relationships between self-control, morality, criminogenic exposure, and both violent and property delinquencies, along with a range of individual demographic, family control, and school bonding variables. Specifically, against the backdrop of literature, this study hypothesizes the following:

Hypothesis 1 (H1): Chinese adolescents who have lower levels of self-control report greater delinquency.

Hypothesis 2 (H2): Chinese adolescents holding stronger moral values against delinquency have lower delinquency.

Hypothesis 3 (H3): The beneficial effect of self-control on delinquency is moderated by morality, with self-control reducing delinquency more evidently among adolescents with lower levels of morality.

Hypothesis 4 (H4): Chinese adolescents who are more frequently exposed to delinquent peers and who are more often involved in unsupervised leisure activities report greater delinquency.

Hypothesis 5 (H5): The positive effect of criminogenic exposure on delinquency is stronger among Chinese adolescents who have greater criminal propensity.

Method

Data

Data for this study were collected from Guizhou, China. Located in Southwest China, Guizhou covers a geographic area of 176,167 square kilometers and has a population of 34.7 million. Unlike large metropolitan areas such as Beijing, Shanghai, Tianjin, and Guangzhou, where most of the existing studies on delinquency in China were conducted, Guizhou is one of the least developed yet most diverse regions in the country. The delinquency levels in Guizhou may be lower due to its lower levels of economic development, nonetheless, we have no particular reasons to expect that the explanatory factors of delinquency would be significantly different for adolescents in Guizhou and other areas of China.

Data were collected from two areas of Guizhou, Guiyang City and Qiannan Buyi and Miao Autonomous Prefecture (shortened as Qiannan), between November and December of 2019.¹ Guiyang is the capital city of Guizhou and one of the economic, commercial, tourist, and big data hubs in Southwest China. Qiannan is located more remotely in the southern part of the province, with 71% of the population as rural residents, and 55% as ethnic minorities such as Buyi and Miao. These two sites were selected because of their distinct physical and social environments, providing access to major subgroups within the population (e.g., urban/rural, ethnic majority/minority, economically better-off/less advantaged). In both sites, we had obtained permission and support for the study from local Bureaus of Education, which are responsible for guiding and administering educational work in the areas.²

We employed a multistage cluster sampling design to select the sample. First, a list of all middle schools in Guiyang and Qiannan, which include seventh, eighth, and ninth grades, was obtained from local Bureaus of Education. From this list of 150 schools, we randomly selected five middle schools from each site for participation. For Guiyang, three schools are in rural or suburban areas and two are in the city. For Qiannan, two schools are in rural or suburban areas and three are in cities. All selected schools agreed to participate. Second, we randomly selected two classes in each grade from the sample schools. These classes had 36 to 52 students. All selected classes agreed to participate. All students from the selected classes were invited to participate. After obtaining oral consent from the teachers and adolescents themselves, we distributed questionnaires to the participants on regular school days, with teachers and school administrators asked not to be present during the survey period. On the spot, the researchers explained to the students the purpose of the project, answered questions, and emphasized voluntary participation. The survey was anonymous and students could withdraw from the study at any time. In total, 2,533 students were invited to participate and 35 declined the invitation, resulting in a sample of 2,498 students.

In terms of demographic traits (Table 1), the average age of sample students was 13.5 years. Boys and girls were equally split. Han students accounted for 44%, with the remaining ethnic minorities.³ About one fifth of the students were urban residents, three quarters were rural residents, and 7% were migrants. Regarding the delinquent behaviors (Table 1), 18.3% of the respondents had committed at least one of the four violent delinquent acts, and 15% had conducted at least one of the five property delinquent acts during the past year. Respectively, 9.5% and 9.6% of the respondents had engaged in one violent and one property delinquent act during the past year. A very small number of the respondents had engaged in all four types of violent delinquency (2.1%) or all five types of property delinquency (0.7%).

Table 1:

Descriptive Statistics of Variables in Multivariate Regression Analyses

Variables	M/%	SD	Minimum	Maximum
Part 1 analysis (all variables)
Violent delinquency	0.33	0.83	0	4
0	81.7%
1	9.5%
2	4.8%
3	1.9%
4	2.1%
Property delinquency	0.25	0.73	0	5
0	85%
1	9.6%
2	2.8%
3	1.1%
4	0.7%
5	0.7%
Self-control	0.00	6.65	−28.33	11.67
Morality against delinquency	0.00	5.66	−33.22	2.78
Male	0.50	0.50	0	1
Age	13.54	1.01	12	16
Han ethnicity	0.44	0.50	0	1
Urban residents	0.18	0.39	0	1
Rural residents	0.75	0.43	0	1
Migrants	0.06	0.24	0	1
Health	3.80	0.87	1	5
Biological parents married	0.80	0.40	0	1
Economic condition	3.14	0.74	1	5
Parents’ education	2.88	0.86	1	6
Family cohesion	19.82	4.90	5	25
Caregiver control	20.68	4.47	5	25
Abuse/neglect	4.46	2.28	3	15
School bonding	27.28	5.86	7	35
Unsupervised leisure activities	0	1	−0.68	6.21
Delinquent peers	0	5.01	−2.87	33.13
Part 2 analysis (independent variables)
Crime propensity	0	11.87	−17.55	54.45
Unsupervised leisure activities	0	1	−0.68	6.21
Delinquent peers	0	5.01	−2.87	33.13

Measures

Dependent Variables

Two dependent variables, violent delinquency and property delinquency, were constructed (Elliott & Ageton, 1980). For violent delinquency, respondents were asked during the past year, whether they had physically attacked other people, participated in a group fight, carried a weapon, and pointed a weapon at others. For property delinquency, respondents were asked whether they had gambled, stolen from a store, stolen from a family member, stolen a bike from a stranger, and damaged others’ property during the past year. Response categories for all these items included 1 (never), 2 (1–2 times), 3 (3–4 times), 4 (5–7 times), and 5 (8 times and more). To construct variety scales of delinquency, responses to each of these items were first recoded to 0 representing not having committed an act and 1 representing engaging in such an act, and then the different items were added up. Variety scales are high in reliability and validity, hence often preferred in study of criminal offending, particularly for measuring individual delinquency, instead of the volume of offending (Sweeten, 2012). The two measures are highly skewed,⁴ and should be treated as count variables with a negative binomial (NB) distribution.

Independent and Control Variables

The independent variables consisted of self-control, morality, and their interaction term. Self-control was formed by an additive scale of 10 items that tap into the dimensions of impulsivity, simple tasks, risk-seeking, self-centeredness, and temper (Grasmick et al., 1993).⁵ The respondents were asked to what degree 10 statements match their behaviors, such as “I often act on the spur of the moment without stopping to think”; “When I am really angry, other people better stay away from me”; and “Sometimes I will take a risk just for the fun of it.” Response categories ranged from completely match my situation (coded as 1) to completely do not match my situation (coded as 5); thus, a higher score suggests greater self-control. Supplemental Appendix S1 (available in the online version of this article) reports the factor analysis (i.e., principal components analysis⁶) and reliability test results of all composite measures in this study.⁷

To signal morality against delinquency, the respondents were asked from a moral perspective, how wrong they think nine different actions such as stealing and physically injuring other people are (H. Zhang et al., 2016). Response options ranged from not wrong at all (1) to very wrong (5). A higher value represents stronger conventional morality. An interaction product was then created between self-control and morality, with both self-control and morality mean centered to make the results more meaningful and interpretable (Hayes, 2017).

A variety of personal traits, parental and household characteristics, family and school variables, and finally criminogenic exposure factors were controlled in the analyses. Personal traits are composed of gender (0 = female, 1 = male), age (in years), ethnicity (0 = ethnic minorities, 1 = Hans), residential status (three dummy variables representing urban residents, rural residents, and migrants), and self-rated health condition compared with peers (1 = very bad, 5 = very good).⁸

Parental and household characteristics consisted of parents’ marital status (0 = biological parents divorced, deceased, or remarried; 1 = biological parents currently married to each other), family economic condition (1 = not good, 5 = very good), and parents’ educational level (an average of father’s and mother’s highest levels of educational attainment with each ranging from 1 = illiterate to 6 = college degree or above).

Family and school variables encompassed family cohesion, caregiver control, guardian abuse/neglect, and school bonding. To capture family cohesion, respondents were asked if they agreed with five statement such as “Family members always provide upmost help and support to each other” and “Family members get along well with one another” (1 = strongly disagree, 5 = strongly agree) (Wen & Lin, 2012). To tap into caregiver control, respondents were asked to what extent their caregivers know their happenings, for example, “They know where I am after school” and “I tell them whom I am with when I am out” (1 = never, 2 = rarely, 3 = occasionally, 4 = often, 5 = always) (Chen et al., 2017). Guardian abuse/neglect was indicated by a scale of three items indicating whether guardians (parents or other adults who lived together and looked after the respondents) have physically harmed, verbally abused, or neglected the respondents (1 = never, 2 = 1–2 times, 3 = 3–4 times, 4 = 5–7 times, 5 = 8 times and more) (Finkelhor et al., 2005). School bonding was assessed using a scale of seven items asking the respondents if they agree “Teachers in my school care about students” and “I like my school’s learning atmosphere,” among others (1 = completely disagree, 5 = completely agree) (Shen & Zhong, 2018).

Finally, criminogenic exposure consisted of unsupervised leisure activities and delinquent peers. Unsupervised activities was a composite measure of two items. The first, unsupervised time, was measured by the number of nights a week on average that are spent alone or with friends hanging out in public places (e.g., shops, streets, squares, parks, around train stations) with nothing to do (Janssen et al., 2016). Response options contained 1 = never, 2 = no more than once per week, 3 = several times per week, and 4 = everyday (weekends included). The second, bar/internet Café visit, asked the respondents during the past year, how often they visited places such as bars and internet cafés (1 = never, 2 = at least once last year, 3 = at least once per month, 4 = at least once per week, 5 = at least 3–4 times per week, 6 = at least once per day). Delinquent peers was a scale of nine items adapted from the National Youth Survey (Elliott, 1985). The respondents were asked during the past year, how many of their friends have committed acts such as beating and injuring others, pointing a knife, stick, or dagger at others, and stealing things from a store, supermarket, or shopping center. Response options included 1 (almost none), 2 (very few), 3 (some), 4 (majority), and 5 (all).

To test the interaction effects of crime propensity and criminogenic exposure, we followed the tradition of SAT research in operationalizing crime propensity by creating a composite scale containing both dimensions of low self-control and low morality (Antonaccio et al., 2017). Then, two products of Crime Propensity × Unsupervised Leisure Activities and Crime Propensity × Delinquent Peers were created. The remaining variables of individual and household characteristics, and family and school variables continued to serve as control variables.

Analyses

Two-part analyses were conducted to gauge the interaction effects first between self-control and morality, and then between crime propensity and criminogenic exposure on delinquency. Within each part, two identical sets of NB regression models⁹ were estimated to explain violent and property victimization, respectively. The interactional effects in NB models are notoriously difficult to test and interpret because due to the functional form and the multiplicative nature, these models are already interactive to an extent,¹⁰ regardless of whether interaction terms are included. Recent studies (e.g., Mize, 2019) find that including an interaction term in the model when there is no actual interactional effect does not harm inferences, but if there is indeed an interaction, “a product term must be included in the model” (p. 99). Following previous research (Hilbe, 2011; Mize, 2019), we plotted and presented average marginal effects (AMEs), estimated by averaging the marginal effects for every case in the sample, to better interpret the interaction effects in NB models.

Full Information Maximum Likelihood algorithm was applied to address missing data and estimate model parameters. No issue of multicollinearity was detected in the models as the variance inflation factors (VIFs) were all below 2.0, under a conservative cutoff point of 3.3 (Kock & Lynn, 2012). All NB regression analyses were conducted in the statistical program Mplus 7.4.

Results

Interaction Effects of Self-Control and Morality on Delinquency

Table 2 reports the Part 1 analysis results on violent delinquency. Model 1 included only the control variables. For gender, a positive coefficient of b = .80 (p < .001) means that the log of expected counts of violent delinquency for males is .80 unit higher than for females, holding other variables constant. The corresponding incidence rate ratio (IRR) is 2.22, meaning that males are expected to have a rate 2.22 times greater than females for violent delinquency. Meanwhile, students from rural areas and students with biological parents married to each other have lower rates of violent delinquency than their counterparts from urban areas (IRR = 0.65, p < .001) and whose biological parents are no longer married to each other (IRR = 0.79, p = .035). Furthermore, students who reported higher levels of control by caregivers (IRR = 0.96, p = .003) or bonding to schools (IRR = 0.97, p < .001) have lower rates of violent delinquency. In contrast, students who engaged in more unsupervised leisure activities (IRR = 1.37, p < .001) or had greater association with delinquent peers (IRR = 1.10, p < .001) have higher rates of violent delinquency.

Table 2:

Interaction Effects of Self-Control and Morality on Violent Delinquency (N = 2,498)

	Model 1		Model 2		Model 3
Variables	b (SE)	IRR	b (SE)	IRR	b (SE)	IRR
Independent variables
Self-control			−.05*** (.01)	0.95	−.06*** (.01)	0.94
Morality against delinquency			−.04*** (.01)	0.96	−.05*** (.01)	0.95
Self-Control × Morality					−.00** (.00)	0.99
Control variables
Male	.80*** (.11)	2.22	.71*** (.11)	2.04	.72*** (.11)	2.05
Age	.05(.05)	1.05	.02(.05)	1.02	.02(.05)	1.02
Han ethnicity	−.14(.10)	0.87	−.11(.10)	0.89	−.11(.10)	0.89
Rural residents	−.44*** (.11)	0.65	−.41*** (.11)	0.66	−.42*** (.11)	0.65
Migrants	−.06(.18)	0.94	−.03(.17)	0.97	−.02(.17)	0.98
Health	.05(.06)	1.05	.12* (.06)	1.13	.11(.06)	1.12
Biological parents married	−.24* (.11)	0.79	−.25* (.11)	0.78	−.25* (.11)	0.78
Economic condition	.00(.08)	1.00	.00(.08)	1.00	.02(.08)	1.02
Parental education	−.07(.06)	0.93	−.04(.06)	0.96	−.05(.06)	0.95
Family cohesion	−.01(.01)	0.99	−.01(.01)	0.99	−.01(.01)	0.99
Caregiver control	−.03** (.01)	0.96	−.02(.01)	0.98	−.02(.01)	0.98
Abuse/neglect	.03(.02)	1.03	.02(.02)	1.02	.02(.02)	1.02
School bonding	−.03*** (.01)	0.97	−.02(.01)	0.98	−.02(.01)	0.98
Unsupervised leisure activities	.31*** (.05)	1.37	.25*** (.05)	1.28	.25*** (.05)	1.29
Delinquent peers	.10*** (.01)	1.10	.07*** (.01)	1.08	.08*** (.01)	1.08

Note. SE = standard error; IRR = incidence rate ratio.

p < .05. **p < .01. ***p < .001.

Results from Model 2 indicated that higher levels of self-control (IRR = 0.95, p < .001) and morality against delinquency (IRR = 0.96, p < .001) are both linked to lower rates of violent delinquency, holding control variables constant. Results from the full model (Model 3) showed that the two-way interaction term between self-control and morality is significant (IRR = 0.99, p = .005), hence the delinquency-containing effect of self-control on violent delinquency is larger among students who have stronger moral values. You can view from Figure 1 that the AME of self-control on violent delinquency is significant and negative when the level of morality is higher than value - 12 on the morality scale. This means that among the very few who have the lowest level of morality, self-control does not affect their rate of violent delinquency. However, for the overwhelming majority of the respondents, the curbing effect of self-control on violent delinquency is significant and increases when the level of morality elevates.¹¹

Figure 1:

The Conditional Effect of Self-Control on Violent Delinquency by Level of Morality

The main effects of self-control (IRR = 0.94, p < .001) and morality (IRR = 0.95, p < .001) are still significant, with adolescents who have lower levels of self-control and morality having greater violent delinquency. Regarding the control variables, across the three models, male, rural residency, biological parents married, unsupervised leisure activities, and delinquent peers exert consistent influences over violent delinquency. Adolescents who are male, have more delinquent peers, and engage in more unsupervised activities have higher rates of violent delinquency. Conversely, adolescents who are in rural areas or have biological parents still married to each other have lower rates of violent delinquency.

Table 3 reports the results of Part 1 analysis on property delinquency. Both self-control (IRR = 0.94, p < .001) and morality against delinquency (IRR = 0.94, p < .001) inhibit property delinquency. Their interaction product is also significant (IRR = 0.99, p = .002), showing that except for the few individuals who have extremely low levels of morality, for the predominant majority of the adolescents, the negative effect of self-control on property delinquency is more apparent among those who have stronger moral values (Figure similar to Figure 1, thus not shown).

Table 3:

Interaction Effects of Self-Control and Morality on Property Delinquency (N = 2,498)

	Model 1		Model 2		Model 3
Variables	b (SE)	IRR	b (SE)	IRR	b (SE)	IRR
Independent variables
Self-control			−.06*** (.01)	0.95	−.06*** (.01)	0.94
Morality against delinquency			.05*** (.01)	0.95	−.06*** (.01)	0.94
Self-Control × Morality					−.00** (.00)	0.99
Control variables
Male	.68*** (.12)	1.97	.60*** (.12)	1.83	.60*** (.12)	1.82
Age	.04(.06)	1.04	.02(.06)	1.02	.01(.06)	1.01
Han ethnicity	−.25(.13)	0.78	−.22(.13)	0.80	−.24(.13)	0.79
Rural residents	−.31* (.15)	0.73	−.32* (.15)	0.73	−.32* (.15)	0.73
Migrants	−.17(.24)	0.85	−.16(.24)	0.85	−.11(.24)	0.89
Health	−.07(.06)	0.93	.03(.06)	1.03	.00(.06)	1.00
Biological parents married	−.12(.13)	0.88	−.13(.13)	0.88	−.13(.13)	0.87
Economic condition	−.05(.10)	0.95	−.05(.10)	0.95	−.03(.10)	0.97
Parental education	−.24** (.09)	0.79	−.22** (.09)	0.80	−.21* (.08)	0.81
Family cohesion	.01(.01)	1.01	.00(.01)	1.00	.00(.01)	1.00
Caregiver control	−.02(.01)	0.98	−.01(.01)	0.99	−.01(.01)	0.99
Abuse/neglect	.04(.02)	1.04	.03(.02)	1.03	.03(.02)	1.03
School bonding	−.05*** (.01)	0.95	−.03* (.01)	0.97	−.02* (.01)	0.97
Unsupervised leisure activities	.22*** (.06)	1.24	.15* (.06)	1.16	.15* (.06)	1.17
Delinquent peers	.09*** (.01)	1.09	.06*** (.01)	1.06	.06*** (.01)	1.07

Note. SE = standard error; IRR = incidence rate ratio.

p < .05. **p < .01. ***p < .001.

Several control variables affect adolescent property delinquency in a significant and consistent manner across the models. In the full model (Model 3), while male students have higher expected rates of property delinquency than female students, rural students have lower rates than their urban counterparts. Students whose parents have higher levels of education are associated with lower incidence rates of property delinquency. Meanwhile, bonding to school is linked to lower rates of property delinquency. In addition, students involved in more unsupervised leisure activities have greater chances of property delinquency, so do those who reported involvement with more delinquent peers.

Interaction Effects of Crime Propensity and Criminogenic Exposure on Delinquency

Five models were estimated to assess the interaction effects of crime propensity (a combination of low self-control and low morality) and criminogenic exposure on violent delinquency (Table 4). In the full model (Model 5), crime propensity has a significant main effect on violent delinquency (IRR = 1.06, p < .001), with children who have a stronger propensity toward crime reporting higher rates of violent delinquency. Meanwhile, both unsupervised leisure activities (IRR = 1.37, p < .001) and delinquent peers (IRR = 1.10, p < .001) are associated with increased rates of violent delinquency. Furthermore, the two interaction terms are statistically significant. For a compelling majority of students, the effect of crime propensity on violent delinquency is significant and positive, and such effect decreases in magnitude when levels of unsupervised leisure activities (interaction IRR = 0.99, p < .001) and delinquent peers (interaction IRR = 0.99, p < .001) increase (Figures 2 and 3). In other words, the harmful impact of crime propensity on violent delinquency is larger in situations where criminogenic risks are lower.¹² Only among a very small number of individuals exposed to an extremely risky environment, crime propensity does not affect their rates of violent delinquency. With respect to the control variables, their effects remained largely the same as in Part 1 analysis, with males continuing to have higher rates of delinquency while rural residents and students who have both biological parents married to one another showing lower rates of violent delinquency.

Table 4:

Interaction Effects of Crime Propensity and Criminogenic Exposure on Violent Delinquency (N = 2,498)

	Model 1		Model 2		Model 3		Model 4		Model 5
Variables	b (SE)	IRR	b (SE)	IRR	b (SE)	IRR	b (SE)	IRR	b (SE)	IRR
Independent variables
Crime propensity			.05*** (.00)	1.05	.06*** (.01)	1.06	.06*** (.01)	1.06	.06*** (.01)	1.06
Unsupervised leisure activities			.23*** (.05)	1.26	.34*** (.05)	1.40	.24*** (.04)	1.27	.32*** (.05)	1.37
Delinquent peers			.07*** (.01)	1.07	.07*** (.01)	1.07	.10*** (.01)	1.11	.10*** (.01)	1.10
Propensity × Activities					−.01*** (.00)	0.99			−.01*** (.00)	0.99
Propensity × Peers							−.00*** (.00)	0.99	−.00*** (.00)	0.99
Control variables
Male	.98*** (.12)	2.67	.69*** (.11)	1.98	.69*** (.11)	1.99	.70*** (.11)	2.01	.69*** (.11)	1.99
Age	.09(.05)	1.10	.03(.05)	1.03	.01(.05)	1.01	.02(.05)	1.02	.01(.05)	1.01
Han ethnicity	−.21(.11)	0.81	−.12(.10)	0.89	−.08(.10)	0.92	−.08(.10)	0.92	−.06(.10)	0.94
Rural residents	−.55*** (.11)	0.58	−.43*** (.11)	0.65	−.41*** (.11)	0.67	−.41*** (.11)	0.66	−.39*** (.10)	0.67
Migrants	−.22(.18)	0.80	−.01(.17)	0.99	−.02(.17)	0.98	.01(.17)	1.01	.02(.17)	1.02
Health	.02(.06)	1.02	.12* (.06)	1.13	.13* (.06)	1.14	.11(.05)	1.11	.12(.05)	1.13
Biological parents married	−.23* (.11)	0.80	−.23* (.11)	0.79	−.20(.11)	0.82	−.25* (.11)	0.78	−.23* (.10)	0.79
Economic condition	.03(.08)	1.03	.01(.07)	1.01	.02(.07)	1.02	.04(.07)	1.04	.03(.07)	1.03
Parental education	−.07(.06)	0.93	−.03(.06)	0.97	−.04(.06)	0.96	−.06(.06)	0.94	−.06(.06)	0.94
Family cohesion	.01(.01)	1.01	−.00(.01)	0.99	−.00(.01)	0.99	.00(.01)	1.00	.00(.01)	1.00
Caregiver control	−.06*** (.01)	0.94	−.02(.01)	0.98	−.02(.01)	0.98	−.02(.01)	0.98	−.02(.01)	0.98
Abuse/neglect	.09*** (.02)	1.10	.02(.02)	1.02	.02(.02)	1.02	.02(.02)	1.02	.02(.02)	1.02
School bonding	−.06*** (.01)	0.94	−.02(.01)	0.98	−.02(.01)	0.98	−.01(.01)	0.99	−.02(.01)	0.98

Note. SE = standard error; IRR = incidence rate ratio.

p < .05. **p < .01. ***p < .001.

Figure 2:

The Conditional Effect of Crime Propensity on Violent Delinquency by Activities

Figure 3:

The Conditional Effect of Crime Propensity on Violent Delinquency by Delinquent Peers

Finally, five models were estimated to determine the interaction effects of crime propensity and criminogenic exposure on property delinquency (Table 5). Results in the full model (Model 5) again indicate that the elevating effect of crime propensity on property delinquency is larger in situations where adolescents engage in fewer unsupervised activities (interaction IRR = 0.99, p = .003) and have fewer delinquent peers (interaction IRR = 0.99, p = .002; Figures similar to Figures 2 and 3, thus not shown). Concerning the control variables, while male students have higher rates of property delinquency, students who have better-educated parents and who are bonded more to school have lower rates of property delinquency.

Table 5:

Interaction Effects of Crime Propensity and Criminogenic Exposure on Property Delinquency (N = 2,498)

	Model 1		Model 2		Mode 3		Model 4		Model 5
Variables	b (SE)	IRR	b (SE)	IRR	b (SE)	IRR	b (SE)	IRR	b (SE)	IRR
Independent variables
Crime propensity			.06*** (.01)	1.06	.06*** (.01)	1.06	.06*** (.01)	1.07	.07*** (.01)	1.07
Unsupervised leisure activities			.12* (.06)	1.13	.26*** (.06)	1.30	.12* (.06)	1.13	.24** (.07)	1.27
Delinquent peers			.06*** (.01)	1.06	.06*** (.01)	1.06	.09*** (.01)	1.09	.08*** (.01)	1.08
Propensity × Activities					−.01*** (.00)	0.99			−.01** (.00)	0.99
Propensity × Peers							−.00*** (.00)	0.99	−.00** (.00)	0.99
Control variables
Male	.84*** (.12)	2.31	.56*** (.12)	1.75	.56*** (.12)	1.74	.56*** (.12)	1.75	.55*** (.12)	1.73
Age	.10(.06)	1.10	.02(.06)	1.02	.01(.06)	1.01	.02(.06)	1.02	.01(.06)	1.01
Han ethnicity	−.33* (.13)	0.72	−.23(.13)	0.79	−.20(.13)	0.82	−.21(.12)	0.81	−.19(.13)	0.83
Rural residents	−.39** (.15)	0.68	−.32* (.14)	0.73	−.28(.14)	0.76	−.32* (.14)	0.74	−.29(.14)	0.75
Migrants	−.19(.25)	0.82	−.15(.24)	0.86	−.13(.24)	0.87	−.12(.23)	0.89	−.10(.23)	0.91
Health	−.13(.07)	0.88	.03(.06)	1.03	.03(.06)	1.03	.01(.06)	1.01	.02(.06)	1.02
Biological parents married	−.13(.13)	0.88	−.12(.13)	89	−.08(.13)	0.92	−.13(.13)	0.88	−.11(.13)	0.90
Economic condition	−.01(.11)	0.99	−.04(.10)	0.96	−.03(.10)	0.97	−.02(.10)	0.98	−.02(.10)	0.98
Parental education	−.24** (.09)	0.78	−.19* (.09)	0.82	−.21* (.09)	0.81	−.22* (.09)	0.80	−.22* (.09)	0.80
Family cohesion	.02(.01)	1.02	.01(.01)	1.01	.01(.01)	1.01	.01(.01)	1.01	.01(.01)	1.01
Caregiver control	−.04** (.01)	0.96	−.00(.01)	0.99	−.00(.01)	0.99	−.00(.01)	0.99	−.00(.01)	0.99
Abuse/neglect	.09*** (.02)	1.10	.03(.02)	1.03	.03(.02)	1.03	.03(.02)	0.03	.03(.02)	1.03
School bonding	−.07*** (.01)	0.93	−.03* (.01)	0.97	−.03** (.01)	0.97	−.02* (.01)	0.98	−.03* (.01)	0.97

Note. SE = standard error; IRR = incidence rate ratio.

p < .05. **p < .01. ***p < .001.

We will end this section with a brief discussion about the effect sizes of the coefficients. Given that our sample size is large (N = 2,498), we reported the significance level (i.e., p value) as well as the IRR, the latter of which is recommended as a good effect size measure (Olivier et al., 2017). For categorical measures, IRRs of 1.22, 1.86, and 3.00 are considered small, medium, and large effect sizes, respectively (Olivier et al., 2017). Applying these criteria, all the significant categorical variables in Tables 2 and 3 (male, rural residents, biological parents married) have small to medium effect sizes. For continuous predictors (e.g., self-control), in our knowledge, no specific criteria are proposed for NB models. These relations, however, appear to be practically meaningful. For example, in Table 2, for self-control (ranging from −28 to 12), every unit increase in it leads to a 5% decrease in the incidence rate of violent delinquency (ranging from 0 to 4), which means that when a child increases their self-control from the lowest to the middle level, their violent delinquency will decrease by half.

Discussion and Conclusion

As a more recent theory, SAT has a promising potential as a general explanation of crime that integrates both individual and environmental factors, and sees no national or cultural boundaries. This study sets out to be the first to test two major propositions of SAT, regarding the interaction effects of self-control and morality, and of crime propensity and criminogenic exposure on both violent and property delinquencies in a previously unexamined setting of China. Our analyses rely on a large, representative sample of middle school students from two areas of Guizhou, and utilize NB regression models to estimate the interplay of a range of individual and environmental variables on delinquency. Some key findings deserve further discussion (in the order of the hypotheses).

Our first finding regards the effects of self-control and morality on delinquency. The significant inhibiting effect of self-control on delinquency is clear, supporting H1 and consistent with a large amount of previous research worldwide (Pratt & Cullen, 2000; Vazsonyi et al., 2017). After controlling for a number of individual demographic and household characteristics, family cohesion, monitoring and abuse variables, school bonding, unsupervised activities, and delinquent peers, adolescents who have lower levels of self-control report greater delinquency, both violent and property, than their counterparts with higher levels of self-control. This result deviates from some prior studies conducted in China, finding that self-control failed to predict delinquency after social variables were included in the analysis (e.g., N. W. Cheung & Cheung, 2008). In this study, the simultaneous inclusion of two important cognitive-developmental elements, self-control and morality, has substantially attenuated the effects of such social factors as caregiver control and monitoring, abuse/neglect, and school bonding on delinquency. This observation does not imply that these social factors are not important; quite the contrary, adolescents’ ability to exercise self-control is likely a result of early socializing experience with family and school, particularly parents’ child-rearing practices (Gottfredson & Hirschi, 1990). Nonetheless, these results remind us the value of cultivating children’s capability to mindfully resist impulses and temptations and delay gratification for future rewards, in all cultures.

Second, moral values against delinquency have a constraining effect on both violent and property delinquencies, similar to what previous studies conducted in China have found (Liu & Liu, 2016; Ngai & Cheung, 2005). Therefore, H2 is supported. This result is well expected, yet it is worth mentioning that prosocial morality can involve multiple spheres and levels (Shelton & McAdams, 1990) and different groups may hold diverse moral and ideological values. This study tests morality against delinquency, which likely has the most direct and strong connection to law-breaking behaviors. Although there is disagreement regarding appropriate conceptualization of morality,¹³ future research should continue to explore measures of morality beyond the narrow definition of values against the law. Morality may be universal or context-dependent, group- or individual-differentiating, and general or law-relevant.

Importantly, this study offers evidence for an interaction relationship between self-control and morality in shaping delinquency. We find that the ability to exercise self-control works better in inhibiting crime for juveniles who have stronger prosocial values. Similarly, a stronger delinquency-suppressing effect of morality occurs among adolescents who have higher levels of self-control. These findings, suggesting a mutually amplifying influence of self-control and morality, echo Antonaccio and Tittle’s (2008) observation that self-control helps those with higher-than-average levels of morality to reduce risk of misconduct. They, however, counter the mainstream SAT’s proposition that self-control matters more for individuals who have lower levels of morality (Pauwels et al., 2018), disproving H3. Acknowledging the differences between this study and SAT studies (e.g., measures of delinquency, types of regression, control variables), the interactive patterns uncovered in this study are noteworthy as they call attention to the situation of “double failures” for individuals who have both low levels of self-control and morality. For these individuals, the beneficial effect of morality on constraining delinquency is the weakest, even though their risks of delinquency are the highest due to fragile self-control.

We propose some possible explanations for the lesser role that self-control plays in curbing delinquency among adolescents who have lower levels of morality. For these adolescents, perhaps their weaker moral constraints make them less likely to connect delinquent behaviors to negative consequences and the collateral displeasing emotions such as guilt and shame (Li et al., 2014). Consequently, they are less inclined to use self-control to override delinquent impulses or desires. Similarly, morality works better against delinquency among youths who have higher levels of self-control. Perhaps for these individuals, moral values have a greater tendency to translate into rightful behaviors, thanks to their stronger ability to resist impulses and desires (Kotabe & Hofmann, 2015). That is, self-control can help them overcome the difficulty of activating social morals when making decisions. Self-control and morality, thus, add to each other’s strength in controlling delinquency in China, a society that heavily emphasizes self-discipline and moral governance.

With the significant effects of both self-control and morality established in the Part 1 analysis, it is not a surprise that propensity toward crime, a composite measure of low self-control and low morality, is predictive of delinquency net of all controls in the Part 2 analysis. What deserves more discussion is the critical roles that criminogenic factors play in delinquency. After controlling for crime propensity, criminogenic exposure variables, including unsupervised leisure activities and delinquent peers, persist to have significant, independent influences on delinquency, supporting H4. Adolescents who spend more unsupervised time and who associate with more delinquent peers, holding their degree of crime propensity constant, have greater delinquency than their counterparts with lower exposure to risky environment.

Furthermore, crime propensity and criminogenic exposure have an interactional effect on delinquency. Specifically, the promoting effect of crime propensity on both types of delinquency decreases in riskier settings. That is, when adolescents are exposed to stronger criminogenic environment consisting of many unsupervised activities and delinquent peers, crime propensity makes less difference in determining delinquency. In the meantime, the detrimental effect of criminogenic exposure on violent and property delinquencies also declines as crime propensity increases. That is, when adolescents are more crime-prone, the negative influence of risky situations, including unsupervised leisure activities and delinquent peers, is less profound. Conversely, when adolescents are crime-averse, the harmful influences of unsupervised activities and delinquent peers are greater. H5 is not supported.

These results align with some previous studies detecting a stronger effect of peer delinquency among individuals who have higher levels of self-control (Meldrum et al., 2009; Vitulano et al., 2010), but contrast with evidence from the SAT literature showing that the delinquency-promoting effect of exposure increases when crime propensity grows (Pauwels et al., 2018). These results also deviate from studies that did not find any significant interactive relationships between criminogenic exposure and self-control in affecting delinquency (e.g., Hoeben et al., 2016; McGloin & O’Neill Shermer, 2009). Clearly, evidence regarding the interactive effect of crime propensity and risky exposure on delinquency is mixed and often contradictory. It should be recognized though that, compared with significant main effects, significant interaction effects are much less likely to replicate (Open Science Collaboration, 2015), mostly due to the small amount of the variance in outcome that interaction effects can explain. Future research should continue to explore both main and interaction effects of theoretically important variables on crime and justice, as “a single study almost never provides definitive resolution for or against an effect and its explanation” (Open Science Collaboration, 2015, p. 6251). Meanwhile, considering that prior research shows that the effect of susceptibility to peer influence on delinquency is greater among adolescents who have higher levels of self-control (Meldrum et al., 2013), it is not completely surprising that exposure to criminogenic situations, different from susceptibility to criminogenic situations, may also be more consequential for adolescents lower in crime propensity. In contrast, for adolescents already crime-prone, environmental risks play a smaller role, probably because they require less stimulation or inducement to engage in delinquency. In any case, future research should continue to use an integration of individual difference theories and structure/process theories as the basis of research on delinquency.

Finally, the effect of children’s residency deserves discussion. Notably, this study found that migrant students have similar rates of delinquency to urban students, and moreover, rural students have significantly lower rates of delinquency than their urban counterparts. These results counter the government’s argument that migrant workers contributed to the rising crime rates in cities. Official statistics often indicate that migrant workers are heavily involved in the criminal justice system as both offenders and victims in such major migrant destination cities as Shanghai and Guangzhou (Xu, 2014). Scholars, however, pointed out that the rates of crime by rural migrants are generally overestimated (e.g., Xu, 2014). Our findings add to this line of argument by showing that migrant children are not more prone to delinquency than urban children, and rural children are even less inclined to delinquency. Migrant children from rural areas thus should not be treated as primary targets of urban policing.

Findings of this study should be understood against the backdrop of some limitations. To start, although we test two main propositions of SAT, our study does not represent or intend to be a comprehensive testing of SAT. One component missing from this study is deterrence (i.e., perceived risk of sanction), which SAT postulates can interact with crime propensity to shape crime and delinquency, and has received some empirical support (Svensson, 2015). Future research, if data are available, should investigate the role that deterrence plays as an external control and its coaction with the internal control of crime propensity in affecting delinquency, in addition to the critical moderating relationships that this study already examines. Relatedly, despite the large number of predictor and control variables that we have incorporated in this study, additional factors that may potentially act as confounders of the studied relationships are missing, such as the individual-level attribute of mental state and the societal-level process of labeling and formal control. Future studies should either include these variables or employ an experimental design to control for unobserved heterogeneity.

Regarding the research design, sample, and data, if possible, future research should adopt longitudinal research design to further understand the causal relationships of SAT. The cross-sectional nature of the study precludes the establishment of temporal orders of the variables. It is, for instance, possible that experience of delinquency may affect an adolescent’s perception of wrongness. It is also anticipated that dismal family factors such as ineffective and abusive parenting can reduce a child’s capacity to exercise self-control. In addition, data for this study, although based on a large, representative sample of students, are limited to two areas in one province. Considering that China is a huge country with enormous area differences, future research should collect data from more geographic areas, if not nationwide, so that findings can better generalize to the juvenile population in the country. Future research should also take upon the challenging tasks of recruiting non-student samples into study of delinquency. School drop-outs, delinquents who are institutionalized, and street children are especially at high risks of delinquency and warrant attention. Finally, as similar projects that utilize self-report data, this study has the potential issue of social desirability bias. Adolescents may be unwilling to reveal information about their attitudes and behaviors that are considered unconventional, sensitive, or incriminating, despite efforts to ensure anonymity and confidentiality. Therefore, it will be beneficial for future studies to incorporate additional sources of data such as those from parents, teachers, and official record (e.g., school disciplinary record and police record) to supplement adolescents’ self-report data.

These limitations notwithstanding, our study represents a meaningful contribution to the literature by extending SAT to an East Asian population that has not yet been examined and uncovering directions of moderation relationships different from what SAT has argued. Our findings shed new light on the interaction between self-control and morality, and between crime propensity and criminogenic exposure on violent and property delinquencies, net of several theoretically prominent and empirically verified correlates of delinquency. The results reconfirm that individual and situational factors interlock in determining delinquency, hence the need for testing theoretical interactions. Essentially, this study reaffirms the value of empirical testing across cultures to validate and possibly improve general theories of crime.

We are ending the article with some final thoughts on policy. Given the importance of self-control and morality in reducing delinquency, we recommend programs that teach children important qualities of self-discipline, responsibility, and consideration of others. Public and private sectors should also provide early education and effective child care to families with preadolescent children, and training that can enhance adult knowledge and skills of monitoring children to reduce adolescent involvement with delinquent peers and unsupervised activities. This is important for families that have children of lower crime propensity as well, because children of lower propensity appear to be more susceptible to the negative influence of risky situations than their counterparts of higher propensity are. These micro- (e.g., individual, parental, and school) and meso-level (e.g., interaction between individual, family, and school) efforts, however, are only part of the broad landscape of delinquency prevention (Weng et al., 2016). Children learn values and behavioral rules from home, peers, communities that are meanwhile operating in the larger society, and adolescents constantly refer to their social environment to determine what is ethically right or wrong, legally permissible or forbidden, practically achievable or unrealizable, and actually rewarded or punished. Macro-level structure and processes thus can determine the development of micro-level and meso-level characteristics that affect individual propensity and criminogenic exposure. In a society increasingly marked by overwhelming materialistic goals, widening inequality, intense cultural anomie, and weakened informal control, top-down campaigns to restore the traditional culture of Confucianism and enforce sets of morality guidelines (Kubat, 2018) may not be a panacea to social problems including delinquency. How to motivate the populace to exercise self-control and internalize morals voluntarily while protecting them from criminogenic environment remain a challenge, for the East and the West alike.

Supplemental Material

sj-docx-1-cjb-10.1177_00938548211034840 – Supplemental material for Explaining Chinese Delinquency: Self-Control, Morality, and Criminogenic Exposure

Supplemental material, sj-docx-1-cjb-10.1177_00938548211034840 for Explaining Chinese Delinquency: Self-Control, Morality, and Criminogenic Exposure by Yuning Wu, Xiaojin Chen and Jia Qu in Criminal Justice and Behavior

Footnotes

The project was supported by a theoretical innovation grant awarded by the Guizhou Social Sciences Association (GZLCLH-2019-011).

ORCID iDs

Yuning Wu

Xiaojin Chen

Supplemental Material

Supplemental Appendix S1 is available in the online version of this article at

Notes

Yuning Wu is a professor in the Department of Criminology and Criminal Justice at Wayne State University in Detroit, MI. Her research interests include public perception of crime and justice, police attitudes and behaviors, and victimization.

Xiaojin Chen is an associate professor in the Department of Sociology at Tulane University in New Orleans, LA. His research interests include life-course criminology, the well-being of socially and economically disadvantaged children, and the application of advanced statistical techniques.

Jia Qu is an assistant professor in College of Public Administration at Guizhou University of Finance and Economics in Guizhou, China. Her current research interests include juvenile delinquency and victimization, policing, law, and social research.

References

Antonaccio

Botchkovar

Hughes

(2017). Ecological determinants of situated choice in situational action theory: Does neighborhood matter? Journal of Research in Crime and Delinquency, 54(2), 208–243.

Antonaccio

Tittle

(2008). Morality, self-control, and crime. Criminology, 46(2), 479–510.

Bakken

(2004). Moral panics, crime rates and harsh punishment in China. Australian & New Zealand Journal of Criminology, 37(1), 67–89.

Bernard

Snipes

Gerould

(2010). Vold’s theoretical criminology. Oxford University Press.

Chen

Liang

Ostertag

(2017). Victimization of children left behind in rural China. Journal of Research in Crime and Delinquency, 54(4), 515–543.

Cheng

Liu

Wang

(2017). Domestic migration, home rentals, and crime rates in China. The Journal of Chinese Sociology, 4(8), 1–21.

Cheung

Liu

(2015). Revisiting the subculture: Understanding deviant students in China. Asian Education and Development Studies, 4(4), 409–422.

Cheung

N. W.

Cheung

Y. W.

(2008). Self-control, social factors, and delinquency: A test of the general theory of crime among adolescents in Hong Kong. Journal of Youth and Adolescence, 37(4), 412–430.

Chui

W. H.

Chan

H. C.

(2016). The gendered analysis of self-control on theft and violent delinquency: An examination of Hong Kong adolescent population. Crime & Delinquency, 62(12), 1648–1677.

10.

Elliott

(1985). National youth survey [United States]: Wave I, 1976. Inter-University Consortium for Political and Social Research.

11.

Elliott

Ageton

(1980). Reconciling race and class differences in self-reported and official estimates of delinquency. American Sociological Review, 45(1), 95–110.

12.

Finkelhor

Hamby

S. L.

Ormrod

Turner

(2005). The Juvenile Victimization Questionnaire: Reliability, validity, and national norms. Child Abuse & Neglect, 29(4), 383–412.

13.

Gallupe

Baron

(2014). Morality, self-control, deterrence, and drug use: Street youths and situational action theory. Crime & Delinquency, 60(2), 284–305.

14.

Gerstner

Oberwittler

(2018). Who’s hanging out and what’s happening? A look at the interplay between unstructured socializing, crime propensity and delinquent peers using social network data. European Journal of Criminology, 15(1), 111–129.

15.

Gottfredson

Hirschi

(1990). A general theory of crime. Stanford University Press.

16.

Grasmick

H. G.

Tittle

C. R.

Bursik

R. J.

Jr. Arneklev

B. J.

(1993). Testing the core empirical implications of Gottfredson and Hirschi’s general theory of crime. Journal of Research in Crime and Delinquency, 30(1), 5–29.

17.

Hayes

(2017). Introduction to mediation, moderation, and conditional process analysis: A regression-based approach. Guilford Press.

18.

Hilbe

(2011). Negative binomial regression. Cambridge University Press.

19.

Hoeben

Meldrum

Walker

Young

(2016). The role of peer delinquency and unstructured socializing in explaining delinquency and substance use: A state-of-the-art review. Journal of Criminal Justice, 47, 108–122.

20.

Hofmann

Meindl

Mooijman

Graham

(2018). Morality and self-control: How they are intertwined and where they differ. Current Directions in Psychological Science, 27(4), 286–291.

21.

Huang

(2015). Morality and law in China, past and present. Modern China, 41(1), 3–39.

22.

Janssen

Eichelsheim

Deković

Bruinsma

(2016). How is parenting related to adolescent delinquency? European Journal of Criminology, 13(2), 169–194.

23.

Jiang

Chen

Zhuo

(2020). Self-control, external environment, and delinquency: A test of self-control theory in rural China. International Journal of Offender Therapy and Comparative Criminology, 64, 1696–1716.

24.

Kock

Lynn

(2012). Lateral collinearity and misleading results in variance-based SEM: An illustration and recommendations. Journal of the Association for Information Systems, 13(7), 546–580.

25.

Kort-Butler

L. A.

(2017). Health-related strains and subsequent delinquency and marijuana use. Youth & Society, 49(8), 1077–1103.

26.

Kotabe

Hofmann

(2015). On integrating the components of self-control. Perspectives on Psychological Science, 10(5), 618–638.

27.

Kubat

(2018). Morality as legitimacy under Xi Jinping: The political functionality of traditional culture for the Chinese Communist Party. Journal of Current Chinese Affairs, 47(3), 47–86.

28.

Kwong

(1994). Ideological crisis among China’s youths: Values and official ideology. British Journal of Sociology, 45(2), 247–264.

29.

LaGrange

Silverman

(1999). Low self-control and opportunity: Testing the general theory of crime as an explanation for gender differences in delinquency. Criminology, 37(1), 41–72.

30.

Dang

Zhang

Guo

(2014). Internet addiction among Chinese adolescents: The effect of parental behavior and self-control. Computers in Human Behavior, 41, 1–7.

31.

Liu

(2016). Are children of rural migrants more delinquent than their peers? A comparative analysis of delinquent behaviors in the City of Guangzhou, China. Crime, Law and Social Change, 66(5), 465–489.

32.

Ren

Marshall

I. H.

(2013). Exploring the utility of self-control theory for risky behavior and minor delinquency among Chinese adolescents. Journal of Contemporary Criminal Justice, 29(1), 32–52.

33.

Maimon

Browning

C. R.

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

34.

McGloin

O’Neill Shermer

(2009). Self-control and deviant peer network structure. Journal of Research in Crime and Delinquency, 46(1), 35–72.

35.

Meldrum

Miller

Flexon

(2013). Susceptibility to peer influence, self-control, and delinquency. Sociological Inquiry, 83(1), 106–129.

36.

Meldrum

Young

Weerman

(2009). Reconsidering the effect of self-control and delinquent peers: Implications of measurement for theoretical significance. Journal of Research in Crime and Delinquency, 46(3), 353–376.

37.

Mize

(2019). Best practices for estimating, interpreting, and presenting nonlinear interaction effects. Sociological Science, 6(4), 81–117.

38.

Ngai

Cheung

(2005). Predictors of the likelihood of delinquency: A study of marginal youth in Hong Kong, China. Youth and Society, 36(4), 445–470.

39.

Olivier

May

W. L.

Bell

M. L.

(2017). Relative effect sizes for measures of risk. Communications in Statistics-Theory and Methods, 46(14), 6774–6781.

40.

Open Science Collaboration. (2015). Estimating the reproducibility of psychological science. Science, 349(6251), 943–950.

41.

Pauwels

Svensson

Hirtenlehner

(2018). Testing situational action theory: A narrative review of studies published between 2006 and 2015. European Journal of Criminology, 15(1), 32–55.

42.

Ponciano

(2020, April 8). The countries with the most billionaires. Forbes. https://www.forbes.com/sites/jonathanponciano/2020/04/08/the-countries-with-the-most-billionaires-in-2020/#2475fd2d4429

43.

Pratt

Cullen

(2000). The empirical status of Gottfredson and Hirschi’s general theory of crime: A meta-analysis. Criminology, 38(3), 931–964.

44.

Pyrooz

D. C.

Decker

S. H.

(2013). Delinquent behavior, violence, and gang involvement in China. Journal of Quantitative Criminology, 29(2), 251–272.

45.

Sampson

R. J.

Bartusch

D. J.

(1998). Legal cynicism and (subcultural?) tolerance of deviance: The neighborhood context of racial differences. Law and Society Review, 777–804.

46.

Schepers

Reinecke

(2018). Conditional relevance of controls: A simultaneous test of the influences of self-control and deterrence on criminal behaviour in the context of situational action theory. European Journal of Criminology, 15(1), 77–92.

47.

Shelton

McAdams

(1990). In search of an everyday morality: The development of a measure. Adolescence, 25(100), 923–943.

48.

Shen

Zhong

(2018). Rural-to-urban migration and juvenile delinquency in urban China: A social control perspective. Asian Journal of Criminology, 13(3), 207–229.

49.

Song

Yan

Jiang

(2019). Can the rise in housing price lead to crime? An empirical assessment of China. International Journal of Law, Crime and Justice, 59, 100341.

50.

Stams

Brugman

Deković

Van Rosmalen

Van Der Laan

Gibbs

J. C.

(2006). The moral judgment of juvenile delinquents: A meta-analysis. Journal of Abnormal Child Psychology, 34(5), 692–708.

51.

Svensson

(2015). An examination of the interaction between morality and deterrence in offending: A research note. Crime & Delinquency, 61(1), 3–18.

52.

Sweeten

(2012). Scaling criminal offending. Journal of Quantitative Criminology, 28(3), 533–557.

53.

Tabachnick

Fidell

(2007). Using multivariate statistics (5th ed.). Allyn & Bacon/Pearson Education.

54.

Ursachi

Horodnic

I. A.

Zait

(2015). How reliable are measurement scales? External factors with indirect influence on reliability estimators. Procedia Economics and Finance, 20, 679–686.

55.

Vazsonyi

A. T.

Mikuška

Kelley

E. L.

(2017). It’s time: A meta-analysis on the self-control-deviance link. Journal of Criminal Justice, 48(1), 48–63.

56.

Vitulano

Fite

Rathert

(2010). Delinquent peer influence on childhood delinquency: The moderating effect of impulsivity. Journal of Psychopathology and Behavioral Assessment, 32(3), 315–322.

57.

Wen

Lin

(2012). Child development in rural China: Children left behind by their migrant parents and children of nonmigrant families. Child Development, 83(1), 120–136.

58.

Weng

Ran

Chui

(2016). Juvenile delinquency in Chinese adolescents: An ecological review of the literature. Aggression and Violent Behavior, 31(6), 26–36.

59.

Wikström

P.-O.

(2004). Crime as alternative. In McCord

(Ed.), Beyond empiricism: Institutions and intentions in the study of crime (p. 1–37). Transaction Publishers.

60.

Wikström

P.-O.

Svensson

(2010). When does self-control matter? European Journal of Criminology, 7(5), 395–410.

61.

(2014). Urbanization and inevitable migration. In Cao

Sun

Hebenton

(Eds.), The Routledge handbook of Chinese criminology (pp. 209–223). Routledge.

62.

Yuan

(2020, June 11). China’s street vendor push ignites a debate: How rich is it? The New York Times. https://www.nytimes.com/2020/06/11/business/china-street-vendors-stall-economy.html

63.

Zhang

Zhao

Ren

Zhao

(2016). Subculture, gang involvement, and delinquency: A study of incarcerated youth in China. Justice Quarterly, 34(6), 952–977.

64.

Zhang

(2008). Juvenile delinquency and justice in contemporary China: A critical review of the literature over 15 years. Crime, Law and Social Change, 50(3), 149–160.

65.

Zhao

Zhang

Zhao

J. S.

Wang

(2019). When the west meets the east: Cultural clash and its impacts on anomie in a sample of Chinese adolescents. Deviant Behavior, 40(10), 1187–1205.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.04 MB