Abstract
This study explores the importance of organizational structures and formal affiliations with the Hong Kong triads to delinquency among youth street gang members in Hong Kong. More specifically, this study examines the relative importance of the number of organizational structures and triad affiliation to patterns of delinquency in a sample of active members of youth street gangs (N = 201). With the aid of outreach social workers, a convenience sampling method was used to recruit a gender-balanced sample of at-risk youths. Logistic regression analysis of the survey data that was gathered indicated that formal affiliation to Hong Kong triads and the presence of organizational structures significantly increased the odds of delinquency (independently of each other). Suggestions for future research on gang membership and delinquency, with particular reference to the Asian context, are provided.
Introduction
The nexus of youth, delinquency, and gangs has for decades been one of the central focus points of gang research, the findings of which have informed youth policy globally (Chu, Daffern, Thomas, & Lim, 2012; Klein & Maxson, 2006). Largely due to the adoption of the dominant American models to conceptualize youth subcultures and youth street gangs, recent scholarship has shown that gang membership is strongly associated with increased rates of delinquency (Delisi, Barnes, Beaver, & Gibson, 2009; Krohn & Thornberry, 2008). Simply put, being a member of a gang means that an individual is more likely to commit serious acts of delinquency—and to do so more frequently (Barnes, Beaver, & Miller, 2010; Curry, Decker, & Egley, 2002; Thornberry, 1998; Tita & Ridgeway, 2007). Despite this consensus, factors that offset the propensity for gang members to engage in acts of delinquency remain a topic of academic investigation. A considerable number of scholars have noted that there is scant knowledge of the relationship between the organizational characteristics of gangs and gang members’ behaviors (Benson & Decker, 2010; Bouchard & Spindler, 2010; Decker, Katz, & Webb, 2008; Decker, Melde, & Pyrooz, 2013; Sheley, Zhang, Brody, & Wright, 1995). For instance, using samples of school attendees or arrestees, Sheley et al. (1995), Decker et al. (2008), and Bouchard and Spindler (2010) have made a compelling case for reexamining the relationship between gang membership and delinquency from the perspective of gang structures and organization. Although these studies have received critical acclaim, in adopting “gang-centric” assumptions concerning the primacy of gang structures over individual-level factors to explain gang members’ acts of delinquency, organization theorists appear to neglect the relevance of out-group relational dimensions and their link with delinquency. This is of particular concern as recent research has shown that street gang membership and its organizational boundaries are transitory or fluid, rather than enduring or rigid (Bolden, 2012; Felson, 2009). The concepts of gang membership and group organization widely used by organization theorists do not accurately reflect this reality of elastic group boundaries. Co-offending was once thought to involve collaboration between members of the same gang only (never between members of different gangs). However, current research shows that, despite being active members of one street gang, individual gang members can and do collaborate and co-offend with individuals who belong to other gangs—a gang member may even hold an “associate status” with another gang (Benson & Decker, 2010; Bolden, 2012; Densley, 2014). This phenomenon, termed extended co-offending (Felson, 2009), reflects the increasingly dynamic character and nature of current-day street gangs, which have been recorded worldwide—for example, in the United States (Bolden, 2012), the United Kingdom (Densley, 2014), and Hong Kong (Lo, 2012). Therefore, by only measuring influential factors within a gang—such as the gang’s organizational structures or delinquent peers in the gang—to explain the gang members’ patterns of delinquency, the significant impact that out-group affiliations may have on the general pattern of delinquency among the gang members is neglected.
If a street gang’s organization does have an impact on the behavior of an individual gang member, what relative influence does having an external co-offending affiliation have on this gang member’s delinquency? Moreover, what difference does it make if an individual gang member has an out-group affiliation with another gang, in terms of frequency and type of delinquency? To the best of the authors’ knowledge, no study has ever examined the relative influence of out-group gang affiliations and in-group organizational features on gang members’ delinquency at an individual level. Therefore, the purpose of the present study was to fill this research gap, by means of a survey-based investigation conducted in an Asian context, that is, Hong Kong, where extended co-offending is evident among youth gangs and organized crime groups (Chin, 1995; Lo, 2012). Despite Hong Kong’s hard-line stance against gangs, particularly the organized crime groups, both youth gangs and adult organized crime groups have maintained a steady presence in the city. Yet there is a paucity of empirical studies that investigate the network overlaps that the two groups may share and its delinquent or criminal manifestations.
Gang Organization and Extended Co-Offending
Three Major Perspectives on Gang Organization
Understood as being the most common form of organization among offenders, gangs have been studied as organizational entities and for decades, scholars have charted their organizational development over time (Bouchard & Spindler, 2010) and across varying contexts, often in the form of cross-national comparison studies (e.g., Esbensen & Weerman, 2005 on youth gangs in the United States and the Netherlands; Tasgin & Aksu, 2015 on youth gangs in the United States, Europe, and Turkey). In the recent literature, there are three competing views on the organization of gangs (Decker & Curry, 2000).
According to the first view of gang structure and organization, a gang is an instrumental-rational entity: A gang is considered to be purposefully structured in a way that facilitates the pursuit of rational goals of self-interest; it features a high degree of organization—such as a hierarchical structure and centralized power—and its economic operations are strategically organized (Benson & Decker, 2010). The gang, in this view, furthermore, is thought to be cohesive and rigid in its group boundaries; it is able to define its group values and instill them in its members, thus ensuring their loyalty (Decker & Curry, 2000). Thus, gangs are taken to be instrumental-rational entities; they are seen as being unified, cohesive groups that are able to articulate group goals; and gang members are thought to be required to meet group expectations and to have clearly defined roles. This view is most often the one taken when adult organized crime groups, such as the Sicilian mafia, the Russian mafia, the Japanese yakuza, and the Hong Kong triads (Varese, 2017), are being described.
The second view of gangs is that they are diffused and unstructured entities. According to this informal-individual perspective, a gang has “difficulties organizing activities, producing consensus, and controlling profits” (Decker & Curry, 2000, p. 474) because its members tend to pursue individualistic goals. This view is the one commonly taken of youth street gangs, where individual motivations and interests trump the articulation of a unified vision of the group as a whole (Densley, 2014). According to this perspective, such gangs lack shared attitudes, norms, and goals; this leads to elastic group boundaries. As Klein (2004) comprehensively puts it, “Street gangs are only loosely structured, with transient leadership and membership, easily transcended codes of loyalty, and informal rather than formal roles for the members” (p. 59).
A third viewpoint repositions the debate within a natural evolution perspective of gang organization. From this perspective, the organizational structures of a gang are not viewed as being instrumental-rational or informal-individual but as existing within a continuum of evolution or natural progression. Thrasher’s (1927) seminal study of 1,313 gangs in Chicago found that gangs could evolve from diffused and disorganized groups into rigid and cohesive ones. In a recent qualitative investigation of the evolution of gang life, gang structures, and gang activities in London, Densley (2014) found a similar natural progression whereby street gangs were able to evolve from diffused and loosely organized structures to more rigid, enterprise-like structures. Based on interviews with 52 active gang members and 17 gang associates, Densley (2014) found that there was support for both instrumental-rational and informal-individual portrayals of gangs within a three-stage evolutionary sequence: recreation, criminal, and governance. As in previous studies (e.g., Hagedorn, 1994; Thrasher, 1927), situational and contextual circumstances were found to be strong determinants of the natural progression of gangs, as they moved from opportunistic crimes to more organized financial activities, such as drug smuggling. However, this linear process of development does not always run smoothly: Densley (2014) found evidence that, having reached a certain stage, a gang might then regress, stagnate, or dissolve. At the culmination of the evolutionary process, the gangs take on the form and function of an organized crime group. For example, over a period of two decades since President Richard Nixon announced the “war on drugs,” Sánchez-Jankowski (2003) argued that implemented changes in the criminal justice system, rather than serving to dismantle organized criminal entities and their activities, unintentionally provoked an organized presence and structure for prison gangs, such that they became “direct extensions of organized crime” (p. 207) and paved way for unifying youth street gangs and adult organized crime groups. According to Sánchez-Jankowski (2003), “Youthful gang members do not want ‘older boys’ controlling them; adolescent rebelliousness has meant that authority imposed on them . . . even if these members were part of the greater gang society, was resented” (p. 207). Yet due to the “aggressive” situational circumstances that increased the likelihood of incarcerating young gang members and passing longer prison sentences, youth street gangs could no longer refrain from affiliating with adult organized crime groups as they feared for their safety within the prison contexts. Thus, paradoxically, the aggressive and harsher criminal justice system gave way to a nonlinear trajectory for youth street gangs to become more organized under the control of adult organized crime groups (Sánchez-Jankowski, 2003).
Despite the differences, there are a number of significant similarities in these three perspectives on gang organization. First, all three perspectives emphasize the importance of the “work environment” (Bouchard & Spindler, 2010, p. 922) or organizational structure to facilitating crime and delinquency. Proponents of all three believe that the organizational complexities of a gang, whether in a hierarchical or diffused arrangement, provide an environment that is conducive to delinquent or criminal behavioral tendencies among its members (Bouchard & Spindler, 2010; Decker et al., 2008). Second, organization theorists such as Bouchard and Spindler (2010) have continued to frame the study of individual delinquent or criminal behaviors within the context of the gang. This consistency is informed by the implicit importance that they attribute to gang membership and gangs’ organizational boundaries to facilitating key group processes among its gang members—for example, the development of group cohesion, within-group intimacy, and individual gang members’ sense of identification with the group (Decker et al., 2008; Decker et al., 2013). Thus, organization theorists assume that to understand why acts of delinquency are committed by gang members, it is necessary to take account of organizational structures and the processes occurring in the “work environment.” Recent research has found strong evidence in support of these hypotheses (Bouchard & Spindler, 2010), with some researchers even going so far as to suggest that the presence of a few organizational structures tends to result in an increase in delinquency among its gang members (Decker et al., 2008). As delinquent groups vary in size, composition, and structure, a gang’s organization is viewed as being a matter of degree—from simple to complex (Benson & Decker, 2010). The relationship between organizational structures and delinquency is often made based on analyzing “whether the level of organization exhibited by a delinquent group is associated with higher productivity levels by members” (Bouchard & Spindler, 2010, p. 923). Therefore, the more structures a gang has, the more organized it is, and the greater is the influence it has on the delinquency of its members (Decker et al., 2008; Sheley et al., 1995).
Fluidity of Gang Membership and Extended Co-Offending
Although the organizational approach to understanding delinquency has gained great traction, it is limited to explaining delinquency solely within the gang context. Unexamined by the above-mentioned scholars, co-offending has long been found to exist beyond the context of individuals’ own gangs (Klein & Maxson, 2006; Thrasher, 1927). Smith, Rush, and Burton (2013), extending a social network-based perspective to gang dynamics, conceptualized the phenomenon in this way: “gangs, organized crime groups, and terrorist groups all play in the same space, the criminal space” (p. 14). The line that once used to demarcate youth gangs and organized crime groups is now becoming increasingly blurred as the two begin to develop alliances for the purpose of co-offending (Bolden, 2012; Densley, 2014; Fleisher, 2002). In recent years, the empirical studies just cited, and others, have yielded findings that support the argument that there is a need to revise the assumptions of organization theorists concerning the gang–delinquency relationship. For instance, a qualitative study conducted by Bolden (2012) found that the traditionally conceived boundaries and rules associated with gang membership no longer applied to the current gang landscape; face-to-face interviews with 15 former gang members in San Antonio, Texas, revealed how initiation rituals were no longer insisted upon and punishment for leaving a gang was no longer enforced. Rather, Bolden (2012) found that individuals freely switched between different gangs for pragmatic reasons—a gang member might switch to a different gang, for instance, because it could yield greater personal benefit—without strict disciplinary action or punishment for so doing. In relating this finding that a gang is a “fluid social network” (Bolden, 2012, p. 220) to the preconception that gangs are organizations with rigid boundaries, Bolden succinctly commented that the “preoccupation with the belief of solid organizational structures of gangs seems to be more irrelevant and inaccurate, as the gang member’s affiliation and status appears much more amorphous and adaptable” (p. 221). Other studies, such as ones by Decker and Curry (2000) and Densley (2014), have found evidence to support the view that gang membership is fluid and the group boundaries are amorphous: individual gang members were neither punished for having out-group affiliations to other street gangs or to organized crime groups nor even discouraged from forming such affiliations.
Contrary to the claims of organization theorists, therefore, from an out-group relational perspective, co-offending can be understood as transcending the conceptual bounds of gang membership and organizational boundaries (Fleisher, 2002). A relevant concept here is that of extended co-offending, which is rooted in the recognition that gang membership is fluid and the organizational structure of gangs, which help form the gang’s group boundaries, is permeable (Felson, 2009). According to Felson (2009), co-offending should not be thought of as involving co-operation only between the members of a single gang. Rather—in line with the plethora of studies attesting to the changing nature of gang’s structures—gang membership and gang structures are volatile and loose: participation in gang activities is now fluid (i.e., members of different gangs freely co-offend); gangs are not the rigid, structured organizations they once might have been (or were once thought to be); and the behavior of gang members is not rigidly controlled (Felson, 2009). Felson (2009) argued that co-operation between offenders extends across time and space to encompass a variety of possibilities, such as being able to involve youth gangs in the protection rackets and loan-shark operations of adult organized crime groups. Felson calls such co-operation as “extended co-offending” (Felson, 2009, p. 160). Extended co-offending has a flexible dynamic in which offenders “can see one another informally in the context of daily activities; or offenders can converge, then act illegally together, even without common symbols” (Felson, 2009, p. 161). Thus, provided that suitable co-offenders can be found and persuaded to join in a criminal endeavor, extended co-offending reveals the possibilities of alliances formed across different groups of individuals that share a common motivation to converge, despite differing gang memberships. Indeed, if affiliating with an organized crime group can help an individual fulfill his or her needs or open doors to new opportunities for him or her, then it makes sense to see this as a possible influential factor in delinquency. Recent studies support this as members of youth street gangs have been found to be more individualistic than collectivistic in their goals when joining or participating in a gang (Decker & Curry, 2000; Decker et al., 2008). Therefore, an individual belonging to a youth gang may purposely get involved in an organized crime gang if he or she believes that the available pool of co-offenders there can offer him or her greater rewards than his or her own gang provides (Bouchard & Spindler, 2010).
The reality of extended co-offending in delinquent or criminal offenses (i.e., between members of different gangs/groups) should not be ignored: It is a mistake to assume that acts of delinquency or criminal behavior by a gang members are committed exclusively within the “working environment” of his or her own gang only. This is especially true in Asian societies, such as that of Hong Kong, where there is considerable evidence of alliances between youth gangs and serious criminal enterprises. Hong Kong’s gang landscape has long featured various types of delinquent or criminal gangs. Although there is no hard data on the number of youth gangs operating in Hong Kong (Lee, 2005a), it is widely recognized that members of youth gangs (or at-risk youths in general) and adult organized crime groups native to Hong Kong (i.e., the Hong Kong triads) are not isolated from one another. Rather, they share the same social space, which the Hong Kong triads use to recruit members of youth gangs or at-risk youths through a process called “triadization” (Lo, 2012). The Hong Kong triads offer lucrative opportunities, absent in youth gangs acting alone, that facilitate the “big brother-followership relationship” (Lo, 2012, p. 560). Such opportunities available to these youths encourage serious delinquency. For example, in recent years, young people—some of the children as young as 13 years old—have been found to be involved in the following Hong Kong triad activities: violent crimes (Hong Kong Police Force, 2017; South China Morning Post, 2016), drug trafficking (South China Morning Post, 2013), and prostitution (Chu, 2000). Although not all youths involved in Hong Kong triads belong to youth gangs, Lee (2005a) found that most members of youth gangs have triad affiliations. Having conducted a survey of 153 outreach social workers in Hong Kong and holding in-depth interviews with 19 of them, Lee (2005a) found that youth gang members were predominantly male and aged between 13 and 15 years old, that gangs were composed of five to eight members on average, and that all the gangs had a three-tier hierarchy, with a leader (or leaders) at the apex, other core members in the middle, and fringe members at the bottom (Lee, 2005a, p. 17).
Unfortunately, there has been no research on the impact of extended co-offending on patterns of delinquency among members of youth gangs in the Hong Kong or even the Asian context. The influence of organizational structures on offending behaviors is well understood, but it cannot account for the offending process resulting from extended co-offending opportunities. Therefore, to better understand patterns of delinquency among members of present-day youth gangs, it is important to account for and address both possible explanations. The first question that arises in the Hong Kong context is this: What is the relative importance of the influence that a youth gang’s organizational features and out-group affiliations to an organized crime group have on the delinquency of an active member of the gang? Second, what difference does it make to a gang member’s pattern of delinquency if he or she has an out-group affiliation to an organized crime group? We hypothesize that a lack of organizational structures within a youth street gang increases the likelihood of out-group affiliations. Without an organizational structure—one that promotes “shared normative orientation, group identity, and ties among gang members” (Decker et al., 2008, p. 384) or provides “some order, some planning, or some coordination to their criminal activities” (Bouchard & Spindler, 2010, pp. 922-923)—an individual gang member’s first allegiance may not be to his or her gang, and his or her main “working environment” may not be that of his or her gang. In such a case, out-group affiliation to an organized crime group would have a higher relative importance to predicting the individual’s general pattern of delinquency than would the organizational structures of the youth gang to which the individual belonged. Furthermore, we also hypothesize that the presence of an out-group affiliation to an organized crime group increases the likelihood of delinquency, in terms of both the frequency and the types of acts of delinquency. Not all group behaviors of youth gangs are confined to delinquent or criminal behaviors, and it may be the case that—with regard to each individual gang—such behaviors are engaged in by only a small network of individuals who form the gang’s core group (Bolden, 2012). Therefore, individuals who seek to meet their material needs, advance their criminal career, or achieve other ambitions that cannot be achieved in their original gangs may opt to affiliate themselves with more severely delinquent social groups that can afford them greater opportunities in these respects (Densley, 2014).
Method
Participants
For the purposes of this empirical study, active youth gang members formed the sample group. Over a period of 2 months, we surveyed 270 at-risk youth gang members from a public housing estate in one of Hong Kong’s new towns. According to Luk (2002) and (Lo, 2012), as well as at schools, commercial areas, and parks or playgrounds, a high concentration of youth gangs can be found loitering around public housing estates, which are subject to the presence and influence of Hong Kong triads. The response rate was 74.4%. Therefore, the final sample consisted of 201 young gang members, aged between 12 and 24 years (M = 16.8 years, SD = 2.49 years). The gender distribution of the sample was roughly equal: 109 males (54.2%) and 92 females (45.8%). Regarding duration of membership in their current youth gang, 153 respondents (78.5%) reported that they had been in their gang for 1 year or more; 28 respondents reported that they had been with their gang for 6 months to less than 1 year; and 14 respondents reported that they had been with their gang for less than 3 months. Ten respondents did not mention the duration of their gang membership.
Although an age range of 12 to 24 years may seem rather wide, it is reflective of Hong Kong’s current gang landscape. As mentioned above, individuals as young as 13 years old have been found to be involved in youth gangs and to have working relationships with the Hong Kong triads. Moreover, research participants as young as 12 or 13 years old have been included in prior studies of gang organization and gang processes—see, for example, Bouchard and Spindler (2010; a study conducted in Canada); Decker and Curry (2000; a study conducted in the United States); and Densley (2014; a study conducted in the United Kingdom). Therefore, this study adheres to the practices of the international academic community, as well as the normal ethical standards involved in researching youth gang members (Bennett & Holloway, 2004).
Measures
The survey questionnaire had three parts; all three were translated into Chinese. Backward translation was done by the first author, and forward translation was done by the collaborating outreach youth social workers. The first part gathered sociodemographic information about the respondents, including their age, gender, education level, and student status.
The Offending, Crime, and Justice Survey (OCJS) instrument (Sharp, Aldridge, & Medina, 2006) formed the second part. It is widely used to collect data on the groups of friends respondents have, the characteristics and activities of such groups, and organizational structures of such groups (Sharp et al., 2006). We used the OCJS instrument to gather information about the organizational structures of the respondents’ gangs, the delinquent acts committed by the respondents, and the respondents’ triad affiliations. The instrument withholds the use of the word “gang” until the end of the survey, at which point a respondent is asked the question “Would you consider the group of people we have been talking about a ‘gang’?” Following the advice of the collaborating frontline social workers, several items were added to the instrument so as to reflect the youth gang context that is unique to Hong Kong. This helped to localize the survey. For example, the items “Are any members of your group associated with the triads?” or “Do you have other friends who are not part of this group?” were added to the survey on the social workers’ advice.
Organizational structures
To measure the complexity of a gang’s organizational structure, we adopted the index method commonly used in previous research on gangs. According to various scholars (Benson & Decker, 2010; Decker & Curry, 2000; Decker et al., 2008), the following seven organizational structures are commonly found in gangs: (a) group exclusivity, (b) territorial possession, (c) recognized group leader(s), (d) hierarchy and promotion, (e) rules and codes, (f) use of force, and (g) norms of compatriot and loyalty. Following Decker et al. (2008), we developed an index based on the seven features listed above. Reflecting on their own gangs, the respondents could report either the presence or absence of each of the listed organizational structures by affirming (1 = Yes) or denying (0 = No) the relevant statement on the questionnaire. In this index, labeled “count of structures,” the sum of the positive responses to the items concerning each organizational structure reflects the overall extent of organizational complexity. The minimum possible index score was 0 (indicating no organizational complexity) and the maximum possible score was 7 (indicating the greatest level of organizational complexity the index allows for).
Delinquency
An index was created to assess the level of delinquency of each respondent; this was labeled as “counts of delinquency” (Decker et al., 2008). The following acts of delinquency were included: (a) intimidating others, (b) using violence against others, (c) drawing graffiti, (d) damaging property, (e) theft, (f) robbery, (g) carrying a knife, (h) carrying a firearm, (i) using illicit drugs, and (j) drug trafficking. As in previous studies, the assumption made was that “a higher score on the index reflects exposure to a diversity of crime opportunities and a deeper commitment to delinquency” (Bouchard & Spindler, 2010, p. 924). A respondent’s score on the counts of delinquency index was created by summing his or her separate scores for each type of criminal or delinquent activity. Thus, the minimum index score possible was 0 (no reported delinquency) and the maximum score was 10 (highly delinquent). A dichotomous variable, accounting for either the presence (1 = yes) or absence (0 = no) of at least one positive response to a criminal or delinquent activity committed within the previous 12 months, was also created.
The CRIME-PICS II questionnaire was used for the third part of the survey (Frude, Honess, & Maguire, 2009). It was used to assess the respondents’ attitudes toward delinquency and criminal acts (Chui & Chan, 2011; Chui, Wu, Kwok, & Liu, 2017). The original instrument is divided into four parts: The first part covers an offender’s current status, history of past offending, and demographic characteristics; the second part is the 20-item CRIME-PICS II questionnaire; the third part is the 15-item problems checklist; and the fourth part is a summary of the offender’s scores from the questionnaires. As the focus of the present study was on delinquency and involvement in crime, only the General Attitude to Offending, Victim Hurt Denial, and Evaluation of Crime as Worthwhile subscales of the CRIME-PICS II questionnaire (part 2) were used.
General Attitude to Offending
The subscale consists of 17 items. It measures the extent to which a person believes that offending is an acceptable way of life. A low score indicates that the respondent believes that offending is not an acceptable way of life. The items on the 5-point Likert-type scale were reverse-scored (with 1 = strongly disagree and 5 = strongly agree). A sample question is “Crime has now become a way of life for me.” The respondents’ total scores ranged from 22 to 65. In our study, the Cronbach’s alpha of the scale was .670.
Victim Hurt Denial
The subscale consists of three items. It measures the degree to which a respondent accepts or rejects the fact that his or her crimes have had an adverse effect on the victims. A low score indicates acceptance of the fact that offending does have adverse effects on the victims. The items on the 5-point Likert-type scale were reverse-scored (with 1 = strongly disagree and 5 = strongly agree). A sample item is “My offenses have no victims.” The respondents’ total scores ranged from 3 to 15. We found that the Cronbach’s alpha of the scale was .522. As this reliability measure was quite low, an interitem correlation test was conducted. This yielded an interitem correlation value of .27, which is within the acceptable range of .20 to .40 (Piedmont, 2014).
Evaluation of Crime as Worthwhile
The scale consists of four items. It measures the degree to which a respondent accepts or rejects the view that crime is a useful way of obtaining goods or of getting excitement. A low score indicates a rejection of the view that crime has benefits that outweigh the costs. The items on the 5-point Likert-type scale were reverse-scored (with 1 = strongly disagree and 5 = strongly agree). A sample item is “In the end, crime does pay.” The respondents’ total scores ranged from 4 to 17. We found that the Cronbach’s alpha of the scale was .550. As with the Victim Hurt Denial subscale, the reliability measure was relatively low; however, as before, an interitem correlation test yielded an acceptable score of .24 (Piedmont, 2014).
Procedure
Field-based data collection—for instance, on-site interviews with respondents in their natural environments—is highly valued in gang research and other research work that involves deviant populations (Bolden, 2012; Densley, 2014). This data collection practice is particularly advantageous for elucidating aspects of social life that are often overlooked by off-site data gathering practices, such as analyzing school records or official reports on arrests made by law enforcement agencies (Fleisher, 2002). However, field-based data collection requires substantial knowledge of the researched locale and access to the target population. The authors of this study therefore enlisted the aid of outreach youth social workers to identify a purposive sample. Outreach youth social workers in Hong Kong are well informed about the current youth gang landscape and subcultures as, while implementing their prevention and intervention strategy, they regularly meet, and frequent the hangouts of, youth gang members (Lee, 2005a). Thus, there were three key advantages to involving them in this study as recruiters of active members of youth street gangs. First, with their experience and training, the outreach social workers were best equipped to access the population and gain their trust (Chui & Chan, 2012). Second, as active gang membership was required for inclusion in this study, so, apart from relying only on self-nomination by the respondents (Bennett & Holloway, 2004; Decker & Curry, 2000; Decker et al., 2013), the outreach social workers were able to confirm whether or not the respondents fulfilled this inclusion criterion. Finally, having outreach social workers involved in the sampling process helped us to achieve a high response rate (Bennett & Holloway, 2004).
Seven outreach youth social workers were trained by the first author of this study on how to administer the self-report instrument. They then administered it face-to-face with the potential respondents on a one-on-one basis. Before administering the questionnaire, one of the outreach social workers would provide a potential respondent with an information sheet that outlined the aim of the study; the social worker would then seek the potential respondent’s informed consent. After this consent was given, the respondents could start to complete the questionnaire. The outreach social workers were available to help the respondents whenever they required clarification. On average, it took the respondents approximately 45 min to complete the questionnaire.
Data Analysis
The independent variables are the presence of triad affiliation; count number of organizational structures (group exclusivity, territorial possession, recognized group leader(s), hierarchy and promotion, rules and codes, use of force, and norms of compatriotism and loyalty); and pro-offending attitudes (general attitude to offending, victim hurt denial, and evaluation of crime as worthwhile). The dependent variables are the count number of delinquent behaviors (intimidating others, using violence against others, drawing graffiti, damaging property, theft, robbery, carrying a knife, carrying a firearm, using illicit drugs, and trafficking drugs) and the presence of delinquency.
SPSS Version 22.0 was used to analyze the data. Univariate and correlation tests were conducted to identify the sociodemographic characteristics of the respondents and to examine the relationships between the key variables. As the sample was nonnormal, Spearman’s Rho correlation analysis and the Mann–Whitney U test were used to determine the significance of the relationships between the key predictor variables and reported delinquency in the previous 12 months. Logistic regression analysis was then used to measure the relative contribution of triad affiliation, pro-offending attitudes, and organizational structures on the presence of delinquency. Various gang scholars have found logistic regression useful for “ascertaining individual predicted probabilities of antisocial outcomes or exposures” (Decker et al., 2013, p. 375), such as delinquency (Bouchard & Spindler, 2010).
Results
Table 1 shows that the distribution of the sample—in terms of gender, age, education, student status, and triad affiliation—was fairly balanced. Most of the respondents belonged to youth gangs with at least one of the organizational structures (98%; n = 197). Most of them reported that their gang had either two of the organizational structures (35.8%, n = 72) or three of them (22.9%, n = 46). The most frequently reported organizational structures were norms of loyalty (157 respondents), territorial possession (98 respondents), and rules and codes (84 respondents). The least frequently reported organizational structure was having a hierarchy and promotion system (19 respondents). With regard to out-group affiliations, 144 respondents (72%) reported having friends who were not part of their own gang; cross-tabulation of these friends outside of their own gang and triad affiliations indicated a significant overlap between the two. As one respondent did not respond to the question on friends outside of the gang, out of 64 positive responses for triad affiliation, 59 of the respondents with triad affiliation (92.2%) were considered to be friends outside of their own gang. In terms of triad affiliation, there were no significant observable differences related to gender. Of the 65 respondents who reported having a triad affiliation, 35 were males and 30 were females. Similarly, of the 136 respondents who did not report having a triad affiliation, 74 were male and 62 female. About a third of the sample (33.8%; n = 68) reported committing one or more acts of delinquency in the previous 12 months.
Sociodemographic Characteristics of the Respondents (N = 201).
Note. F.1 through F.7 is the equivalent of Grade 7 through Grade 13 in the U.S. education system. Yi Jin is an alternative educational pathway for those who leave school before F.7.
Table 2 shows that victim hurt denial and the evaluation of crime as being worthwhile were not significantly correlated to the presence or counts of delinquency; they were therefore excluded from further analysis. No significant correlations were found for any interaction terms between the independent variables of interest. The Mann–Whitney U test was conducted for the triad affiliation, counts of structures, and general attitude to offending variables to assess the distribution of the presence of delinquency among the key independent variables. The results showed that the triad affiliated and nontriad affiliated subsamples significantly differed in terms of the reported number of acts of delinquency. The respondents belonging to gangs with a triad affiliation (Mdn = 1) reported committing more acts of delinquency than did the respondents belonging to gangs with no triad affiliation (Mdn = 0), U = 2,323.5, p = .000. Delving deeper into the differences, a Fisher’s exact test showed that intimidating others (χ2 = 27.19, df = 1, p = .000), using violence against others (χ2 = 11.09, df = 1, p = .001), damaging property (χ2 = 19.00, df = 1, p = .000), and using illicit drugs (χ2 = 10.93, df = 1, p = .001) were significantly associated with having formal affiliations with the Hong Kong triads. The subsample of respondents whose gangs lacked any organizational structures (Mdn = 0) did not significantly differ in terms of delinquency from the subsample of respondents who did report belonging to gangs with organizational structures (Mdn = 1), U = 258.0, p = 1.60. Last, it was found that those who reported committing acts of delinquency (Mdn = 1) differed significantly from the subsample of respondents who did not report committing any acts of delinquency (Mdn = 0) in terms of their scores on the general attitude to offending instrument, U = 3,188.5, p = .23.
Descriptive Statistics and Correlations of the Key Variables of Interest (N = 201).
p < .05. **p < .01. ***p < .001.
Based on the results from the bivariate analyses, logistic regression analysis was performed to determine the relative importance of the organizational dimension and the out-group relational dimension to the presence of delinquency. Table 3 presents the results of the logistic regression analysis, using significantly correlated variables to predict the presence of delinquency in the previous 12 months. The variables were entered in three steps; in the first step, the demographic variables were entered; in the second step, a composite variable—of the number of organizational structures reported to exist in the respondents’ gangs—and the scores on the general attitude to offending instrument were added; and in the third step, the triad affiliation variable was added.
Results of the Logistic Regression Analysis of the Reported Acts of Delinquency Committed in the Previous 12 Months (N = 201).
Note. CI = confidence interval.
Regression coefficient.
Odds ratio.
p < .05. **p < .01. ***p < .001.
Including age, gender, education level, student status, triad affiliation, number of organizational structures, and general attitude to offending variables in the model, results from Table 3 indicated that the overall model was statistically significant in predicting reported delinquent behavior. Triad affiliation, number of organizational structures, and education level were found to be the most statistically significant predictors in the model. The presence of triad affiliation was found to increase the odds of committing acts of delinquency by 5.9 times (Wald statistic = 20.384, p = .000). Not as significant as triad affiliation, the number of organizational structures was found to increase the odds of committing acts of delinquency by 1.3 times (Wald statistic = 4.043, p = .044). Having a high level of education was found to decrease the odds of committing acts of delinquency by 0.6 times (Wald statistic = 7.427, p = .006). In terms of model fit, the Nagelkerke’s R2 value showed that the model resulting from the third step was able to account for 38.4% of the variance.
Discussion and Conclusion
The aim of this study was to examine the relative influence of out-group affiliation and in-group organizational characteristics to youth gang members on an individual level of analysis. The respondents were active gang members in an Asian context, and the results of this study provide evidence of the presence of out-group affiliations among individual youth gang members. Most of the out-group affiliations that the respondents reported having were found to be linked to the Hong Kong triads. This supports recent findings on the fluidity of gang membership and organizational structures (Bolden, 2012; Densley, 2014), which allow individual gang members to develop affiliations with other social groups or entities (such as organized crime groups) in pursuit of more excitement, more money, more prestige, and so on than their own gangs provide them. Regarding the presence and levels of delinquency among those with out-group affiliations to the Hong Kong triads, we found that the presence of such affiliations drastically increased the odds of a young individual committing such delinquent acts as intimidating others, using violence against others, damaging property, and using illicit drugs. These findings provide a unique contribution to understanding the delinquency or criminality of younger persons involved in deviant social groups. Youth gangs are often demonized for their propensity to engage in delinquent or criminal group behaviors (Bennett & Holloway, 2004; Bouchard & Spindler, 2010). Yet, as researchers have argued, co-offending is not a hallmark group behavior of youth gangs (Bennett & Holloway, 2004; Decker et al., 2013) nor is it engaged in by all gang members (Bolden, 2012; Lee, 2005b). Rather, youth gangs may be subject to influence from other socially deviant groups, such as organized crime groups, which operate and move in the same social spaces as they do. The findings of the present study support the view that intersections between different socially deviant groups do take place and that the delinquency or criminality of individual youth gang members may be the result of the dynamics and intersections that take place in that space (Densley, 2014; Fleisher, 2002; Smith et al., 2013). For example, with respect to the concept of triadization, the findings support the notion that youth co-opted to join the Hong Kong triads often enter them as “blue lanterns” or the followers of “big brothers” in the triads (Lo, 2012). The recruiting Hong Kong triads often expect these young, low-status members to demonstrate their willingness to commit acts of violence and to take illegal drugs (Lo, 2012).
Furthermore, the findings of this study confirm the importance of an out-group relational understanding to delinquency among gang members on an individual level. By comparing those with out-group affiliations to the Hong Kong triads and those without such ties, we were able to test the relative importance of organizational complexity and out-group affiliations to acts of delinquency. The results revealed positive influences from both organizational structures and triad affiliation toward reported delinquency. This confirmed the findings of studies conducted by organization theorists (Decker et al., 2008) as well as out-group relational theorists (Chu et al., 2012; Lo, 2012) on the importance of organizational structures and out-group affiliations, respectively, to patterns of delinquency. However, comparing organizational structures and triad affiliation against each other for their influence on delinquency, the finding of the present study is that out-group affiliations to the triads appear to have a much greater influence on individual-level delinquency than do the organizational structures of a gang. Moreover, the influence that triad affiliations have seems to be independent of the effect of a gang’s organizational complexities. As do the findings of a number of other recent studies, our findings cast doubt on the assumption that the delinquency and co-offending of gang members can be best understood as deriving solely from group features of the gangs to which they belong. Recent studies have argued that the phenomenon of co-offending cannot be confined to the conceptual bounds of group membership only. Rather, conceptualizing group membership and organizational boundaries to be fluid and permeable, extended co-offending and movement between gangs are factors that may help explain variations in delinquency and criminal behaviors (Bolden, 2012; Felson, 2009). In the case of the present study, individual variations in delinquency were best predicted by the presence of out-group affiliations to the Hong Kong triads, rather than by the organizational complexity of the youth gangs of which the respondents were active members.
The results of this study provide a useful resource for future researchers to consult and build upon. First, as having an affiliation to an organized crime group was not found to be widespread in our sample of members of youth gangs, it is important to understand the decision-making factors involved in establishing a connection with an organized crime group. For example, are they influenced by in-group or out-group favoritism (Densley, 2014), availability of desired crime opportunities (Bouchard & Spindler, 2010), or perhaps a desire for reliable protection from rival gangs (Lo, 2012)? Furthermore, it is important to connect the dots between the influence of the organizational structures of a young individual’s own gang and the influence of his or her out-group affiliation on his or her general pattern of delinquency. How was the quality of the organizational structures in the youth gangs that had members with out-group affiliations (Densley, 2014)? As the present study examined gang structures only by means of an index of a number of organizational structures, our findings—like those of many other studies (Benson & Decker, 2010)—contribute nothing to an understanding of the influence of the quality of organizational structures on delinquency.
To conclude, gang scholars have for decades sought to understand the processes involved in shaping and maintaining the patterns of delinquency or criminality among gang members through the concept of gang membership and gangs’ organizational structures. In recent years, however, this framework has been challenged by studies that suggest that a gang is a “fluid social network” (Bolden, 2012; Fleisher, 2002). Rather than fixed and rigid, gang membership and the gang organizational structures, which help to form the group’s boundaries, are conceptualized to be transitory and permeable. Thus, a gang member’s pattern of delinquency can be influenced by opportunities that allow him or her to co-offend with a member or members of another gang. Unfortunately, previous studies have not investigated the relevance and importance of the organizational structure thesis and extended co-offending thesis together for their relative influence on delinquency. In this study, we surveyed a sample of active (male and female) members of Hong Kong youth gangs. Out-group affiliations with the Hong Kong triads were evident among both the male and female respondents. We found, as have the researchers of previous studies, that affiliation with the Hong Kong triads appears to have a significant influence on the odds of exhibiting delinquency, increasing the propensity to threaten others, to use violence against others, to damage property, and to use illicit drugs. Moreover, this influence was found to be independent of and far greater than the influence of gangs’ organizational structures. In light of these findings, it is clear that co-offending should be conceived as a group behavior that can exist beyond the confines and rules of a single youth gang. To understand delinquency among youth gang members on an individual level of analysis, we suggest that there needs to be a closer examination of not just the organizational dimensions but also the out-group relational dimensions involved in gang life. This would help us to better understand the nature and character of present-day youth gangs within a larger social space of similar, interacting groups (Bennett & Holloway, 2004). For practitioners involved in antigang efforts, such as outreach youth workers, the reality of “extended co-offending” among youth gangs, particularly in contexts where adult organized crime groups are omnipresent, signals a need to adopt a comprehensive intervention strategy that aims to prevent and protect the youth from transitioning to more serious acts of delinquency. This would not only imply vigilant attendance to the youths’ developmental and identity needs (Lee, 2005b) but also the stability of the bonds they have with the youth gangs they are already a part of and the dynamic interactions that may exist between youth gangs and adult organized crime groups. Attention to these details, for example, through observations and relationship-building with the youth gang members, would shed light on potential “push” and “pull” factors associated with extended co-offending, thereby informing effective prevention strategies, such as purposeful groupwork and casework intervention.
As with a number of previous studies on gangs (e.g., Bouchard & Spindler, 2010; Decker et al., 2008), the cross-sectional design of the present study means that the findings may have limited generalizability. This study makes no claims for the generalizability of its findings. Another possible limitation may arise from the selection process by which outreach youth social workers selected the potential respondents to complete the survey (Bennett & Holloway, 2004). In this study, a minority of the male and female youths surveyed reported having committed at least one count of delinquent behavior in the previous 12 months. This may seem somewhat anomalous, considering how the Commissioner’s Operational Priorities in Hong Kong continues to recognize youth gangs’ propensity for violent crimes as a possible threat to society (Hong Kong Police Force, 2017). It is possible that the outreach social workers helped to recruit only youth gang members with whom they were familiar to ensure their participation and to secure a roughly equal gender distribution in the sample. In other words, the purposive sampling may have led to the recruitment of mainly gang-involved youths who had received guidance from or counseling by outreach social workers. This may explain why the number of acts of delinquency reported by the respondents was low. Furthermore, this may also explain why the results of the study found no significant gender differences in delinquency and triad affiliation. It may be possible that those who were approached by the outreach social workers are undergoing the same prevention or treatment intervention. Finally, this study did not operationalize network embeddedness and positionality of the researched within their respective gangs (for more on social embeddedness and networks, see Granovetter, 1973, 1983). Therefore, among other things, the findings are not able to shed light on who or under what circumstances youth gang members participate or do not participate in extended co-offending with adult organized crime groups. Recent research studies have found that gang embeddedness influences delinquency and the stability of gang membership (Hughes, 2013; Pyrooz, Sweeten, & Piquero, 2013). Future studies may consider adopting such lines of inquiry to further elucidate the salience of gang embeddedness in contexts of extended co-offending.
Footnotes
Acknowledgements
The authors are grateful to the anonymous reviewers for their constructive comments on earlier drafts of the manuscript.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The work described in this article was partly supported by a grant offered by the Hong Kong Young Women’s Christian Association.