Abstract
The high prevalence of violent offending amongst gang-involved youth has been established in the literature. Yet the underlying psychological mechanisms that enable youth to engage in such acts of violence remain unclear. One hundred eighty-nine young people were recruited from areas in London, UK, known for their gang activity. We found that gang members, in comparison to nongang youth, described the groups they belong to as having recognized leaders, specific rules and codes, initiation rituals, and special clothing. Gang members were also more likely than nongang youth to engage in violent behavior and endorse moral disengagement strategies (i.e., moral justification, euphemistic language, advantageous comparison, displacement of responsibility, attribution of blame, and dehumanization). Finally, we found that dehumanizing victims partially mediated the relationship between gang membership and violent behavior. These findings highlight the effects of groups at the individual level and an underlying psychological mechanism that explains, in part, how gang members engage in violence.
In England and Wales, young men under the age of 30 are at highest risk of becoming victims of violent crime (Home Office, 2007; Sivarajasingam, Wells, Moore, Morgan, & Shepherd, 2011). Furthermore, two thirds of crime experienced by 10–15 year olds are violence-related (Home Office, 2011), and an overwhelming proportion of violent crime can be attributed to street gang behavior (Home Office, 2008). Past literature has established the link between gang membership and criminal behavior so far as to distinguish gang youth from nongang delinquents by their asymmetrical criminality (Tita & Ridgeway, 2007). The literature has also shown that gang membership escalates general criminal behavior to violence (Battin, Hill, Abbott, Catalano, & Hawkins, 1998; Esbensen & Huizinga, 1993). However, we still lack thorough understanding of why and how young people are able to engage in these acts of violence. So far, we know that there is a relationship between the collective identity of the gang (including group cohesiveness), threats to that identity, and an increase in gang-related violence (Decker, 1996; Decker & van Winkle, 1996). We can also infer from the literature that informal social controls that have been internalized are either discarded in favor of new norms (such as gang rules), or adapted to enable gang members to engage in gang-related crime (see Wood & Alleyne, 2010, for review). Yet the mechanisms directly linking these internalized beliefs and violence are not fully clear. The purpose of the current study was to examine the relationship between gang membership and violent crime by exploring which sociocognitive mechanisms, related to violent offending in comparable contexts, can explain this relationship.
Social Cognition and Criminality
There is an abundance of psychological theory explaining intergroup conflict, for example, social identity theory (Tajfel & Turner, 1986) and its derivative ethnolinguistic identity theory (Giles & Johnson, 1987), social dominance theory (Sidanius & Pratto, 1999), reputation enhancement theory (Emler & Reicher, 1995), etcetera; yet very few have been applied to the context of street gangs (see, however, Goldman, Giles, & Hogg, 2014). Bandura (1986) conceptualized social cognitive theory (SCT) to explain human behavior (prosocial and antisocial) by the reciprocal interactions between (a) behavior, (b) cognitive and personal factors, and (c) environmental events. SCT encapsulates the various factors that contribute to human behavior and their bidirectional relationships. Furthermore, Bandura (2002) argued that everyone develops a moral self and, as part of that self, there is a dyadic moral agency: inhibitive form—the ability to refrain from behaving inhumanely; and proactive form—the ability to behave humanely. However, people experience moral conflicts when they come across valuable benefits that would arise from immoral behavior (Bandura, 1990). In order to attenuate any cognitive dissonance that has resulted from these moral conflicts, we employ moral disengagement strategies that cognitively restructure the immoral/harmful behavior. These strategies were derived from Sykes and Matza’s (1957) neutralization techniques (i.e., techniques that minimize or sanitize personal misconduct). Bandura (1990, 2002; Bandura, Barbaranelli, Caprara, & Pastorelli, 1996) theorized we employ the following eight strategies: (a) moral justification—“the end justifies the means”; (b) euphemistic labelling—relabelling immoral and/or criminal acts by sanitizing the language, for example, “little white lie”; (c) advantageous comparison—comparing your own behavior to far worse acts committed by others, for example, a stabbing compared to genocide; (d) diffusion of responsibility—“where everyone is responsible, no one is really responsible” (Bandura, 1990, p. 36–37); (e) displacement of responsibility—responsibility is attributed to the person giving orders, not the person carrying out the deed; (f) distortion of consequences—disregarding or minimizing the harm done; (g) blaming the victim—the victim deserves the consequences because of their past behavior; (h) dehumanization—the victims are “no longer viewed as persons with feelings, hopes and concerns but as sub-human objects” (Bandura, 2002, p. 109). Some of these strategies have been examined separately in the social psychological literature (e.g., displacement of responsibility was demonstrated in the classic Milgram study where participants inflicted harmful stimuli on others especially when the “experimenter” claimed full responsibility, Milgram [1974]; dehumanization has been examined in relation to violence (including sexual violence) and it has been examined within interpersonal and intergroup contexts, see Castano & Kofta [2009]). Moral disengagement has also been looked at as a composite variable where specific strategies are not distinguished (see following lines).
The role moral disengagement has in facilitating criminality has been explored in past studies. For example, prior research has found that moral disengagement is both correlated with and predictive of offending (Shulman, Cauffman, Piquero, & Fagan, 2011), and it acts as a mediator for the following relationships: neighborhood impoverishment–antisocial behavior, empathy–antisocial behavior (Hyde, Shaw, & Moilanen, 2010); bullying–perceived importance of social status (South & Wood, 2006), and bullying–prison gang-related activity (however only partially; Wood, Moir, & James, 2009). Furthermore, Bandura et al. (1996) found a link between moral disengagement and aggressive behavior. In a sample of 10- to 15-year-old high school students, they found that participants who self-reported high moral disengagement, behaved more aggressively (verbally and physically), and were more likely to exhibit thought patterns supporting aggressive behavior (i.e., hostile rumination and irascibility). More specifically, it was found that moral reconstrual, obscuring responsibility, misrepresenting consequences, and vilifying or dehumanizing victims were linked to delinquent and aggressive behavior. In line with this, Elliot and Rhinehart (1995) found that proclivity to morally disengage predicted both minor and serious assaults in adolescents. Therefore, the sociocognitive theory of moral disengagement (Bandura, 2002) sets an explanatory backdrop for examining and understanding gang involvement.
Social Cognition in Gang Members: Existing Literature
Vigil (1988) stated that for young people who join gangs, “the gang norms, its functions, and its roles help shape what a person thinks about himself and others, and the gang provides models for how to look and act under various circumstances” (p. 421). Decker and van Winkle (1996) conceptualized this process, an interaction between the social environment and personal beliefs and attitudes, by a means of pushes and pulls. Young people are pushed into gang membership as a result of perceived threats and/or social pressures; they are pulled into gang membership as a result of the attractive benefits they acquire (e.g., increased social status). This has been exemplified in the literature whereby youth with less confidence and low self-esteem, and weak bonds with a prosocial environment and social network (i.e., schools and family) are more likely to look towards gangs than youth who are more confident (Dukes, Martinez, & Stein, 1997). Furthermore, self-esteem has a dynamic relationship with gang membership. It plays a central role in whether a young person joins a gang, participates as a member, and decides to leave the gang (Dukes et al., 1997). To illustrate, a young person with low self-esteem could look towards a gang for support and, consequently, as the group esteem goes up (due to success in delinquent and antisocial activities) that individual’s esteem parallels.
Since the literature supports the reciprocal relationship between the self and the social environment, it can be argued that similar sociocognitive processes would be employed in gang members. For example, Esbensen and Weerman (2005) conducted a study with adolescents from the US and the Netherlands. They examined moral disengagement as a whole construct (i.e., they did not examine specific strategies) and found that gang members scored significantly higher on moral disengagement items than nongang youth in both the US and the Netherlands. Subsequently, additional research has examined the specific moral disengagement strategies in relation to gang membership. For example, in the UK, Alleyne and Wood (2010) found that gang membership was linked with euphemistic labelling, attribution of blame (in fully fledged gang members), and displacement of responsibility (in peripherally involved youth). Unexpectedly, yet nonetheless interesting, moral disengagement as a composite measure was not related to gang involvement (Alleyne & Wood, 2010) nor did it predict gang-related crime (Alleyne & Wood, 2013). Alleyne and Wood (2010) argued that despite the evidence that specific strategies are employed by gang-affiliated youth, their findings also indicate that young people are aware of the consequences of their behavior which is alarming when you consider the extreme forms of violence they engage in. Their findings support that moral disengagement strategies are highly sensitive and context-dependent which warrants further independent study. Therefore, in order to fully understand the sociocognitive processes enabling gang-related crime (and especially violence) it is vital to examine the strategies of moral disengagement specifically and the roles they play.
The Current Study
Traditionally, street gangs have been viewed as an American phenomenon. This perceptual bias was coined as the Eurogang Paradox—the tendency throughout Europe to claim that the “American gang problem” (based on inaccurate stereotypes) is nonexistent in Europe (Klein, 1996; Klein, Kerner, Maxson, & Weitekamp, 2001). However, there is growing evidence supporting the existence of gangs across Europe generally (e.g., Esbensen & Weerman, 2005; Klein et al., 2001; Klein, Weerman, & Thornberry, 2006) and in the UK specifically (e.g., Densley, 2013; Mares, 2001; Sharp, Aldridge, & Medina, 2006). In the UK, examples of prevalence rates from community-based samples include 6% claiming membership to a “delinquent youth group” (Sharp et al., 2006) and 7% claiming street gang membership (Alleyne & Wood, 2010, 2013). Yet these young people commit a disproportionate amount of the violent crime (Bennett & Holloway, 2004; Sharp et al., 2006) and when compared to nongang youth, they engage in significantly higher levels of crime overall and violence specifically (Alleyne & Wood, 2010, 2013). Comparisons between gang-involved youth and nongang youth in an urban city where violent crime is highly prevalent (i.e., London, UK; Smith & Allen, 2004) is a fruitful way to elucidate underlying processes facilitating gang-related violence (Klein, 2006). The current study was not meant to be an exhaustive investigation into the sociocognitive processes related to gang-related violence. Such an investigation is too large for one study to achieve. Based on the literature reviewed and the sociocognitive theory of moral disengagement, the purpose of the current study was to empirically test the extent to which moral disengagement strategies play a facilitative role in violent behavior carried out by gang-involved youth. As a cross-sectional study, we cannot determine causality. However, we hypothesized that gang-involved youth would endorse moral disengagement strategies significantly more than nongang youth. We also hypothesized that at least one of the moral disengagement strategies would act as a mediator, explaining the relationship between gang membership and violent behavior. However, no firm hypotheses were made about which strategies due to the exploratory nature of this study.
Method
Participants
Participants were recruited from four youth centres and one secondary school in London, UK (N = 189). The age of participants ranged from 12 to 25 years old (M = 15.26, SD = 2.82). There were 152 (81%) male participants and the ethnicities reported by participants included White UK/Irish/European (27%), Black/Black British (40%), Asian (17%), mixed ethnicity (14%), and other (2%). Based on the criteria outlined next, 25 participants were identified as gang members (see Table 1 for further details on demographic characteristics).
Demographic characteristics of the total sample, nongang youth, and gang members.
Note. Figures in parentheses are the proportions (in percentages) of young people in each subgroup.
Measures
The Youth Survey: Eurogang Program of Research
The Youth Survey (Weerman et al., 2009) consists of 89 items including questions on demographic characteristics (e.g., age, gender, ethnicity), group membership and group characteristics, and criminal behavior (e.g., violent behavior). This instrument was also designed to identify those who do and do not belong to a gang, according to the Eurogang definition (see next for criteria), and contains further measures of risk and protective factors (see Weerman et al., 2009, for a more detailed description).
Gang membership
Group affiliations were first assessed: for example, “In addition to any such formal groups, some people have a certain group of friends that they spend time with, doing things together or just hanging out. Do you have a group of friends like that?” Participants who responded “yes” were then asked questions assessing gang membership. In accordance with the Eurogang definition the following four components were measured: (a) youthfulness—that is, all members of the group were under the age of 25; (b) durability—the group had been together for more than 3 months; (c) street orientation—responding “yes” to the item “Does this group spend a lot of time together in public places like the park, the street, shopping areas, or the neighborhood?”; (d) group criminality as an integral part of the group identity—responding “yes” to the items “Is doing illegal things accepted by or okay for your group?” and “Do people in your group actually do illegal things together?”
Gang members were identified as such if they met all four criteria of the Eurogang definition. 1
Violent crime
Three items were used to assess the extent to which participants engage in crimes that involve assaulting people (as outlined by Esbensen & Weerman, 2005), and this was labelled violent crime. Using a 5-point Likert-type scale (i.e., “never,” “once or twice,” “3–5 times,” “6–10 times,” and “more than 10 times”), participants were asked how often they engaged in the following behaviors: “hit someone with the idea of hurting them,” “attacked someone with a weapon,” and “used a weapon or force to get money or things from people.”
Mechanisms of Moral Disengagement Scale
This scale (Bandura et al., 1996) is a 32-item instrument assessing the extent to which participants are willing to set aside their moral standards by cognitively restructuring their malevolent behavior. This process, in turn, reduces cognitive dissonance attributed to the behavior. Using a 5-point Likert-type scale (i.e., “strongly agree” to “strongly disagree”), participants were asked how much they agreed with statements corresponding to each of the moral disengagement strategies. The scale is comprised of eight subscales (four items each) measuring Bandura et al.’s (1996) eight moral disengagement strategies. Examples include: “It is alright to fight when the respect of your group is threatened” (moral justification); “Slapping and shoving is just joking around” (euphemistic labelling); “It is okay to insult a classmate because beating him/her is worse” (advantageous comparison); “People cannot be blamed for misbehaving if their friends pressured them to do it” (displacement of responsibility); “It is unfair to blame a person who had only a small part in the harm caused by a group” (diffusion of responsibility); “Insults among peers don’t hurt anyone” (distorting consequences); “People who get mistreated usually do things that deserve it” (attribution of blame); and “Some people deserve to be treated like animals” (dehumanization).
Procedure
After receiving ethical approval from the School of Psychology Ethics Committee, letters were sent to the parents of young people aged under 16 who attended the youth centres and secondary school. Parents were given time (range of 2 to 3 weeks) to reply to the letters if they did not want their children to participate. Also, the gatekeepers at the youth centres (programme directors) and secondary school (classroom teacher) signed consent forms, in loco parentis, allowing the researchers to recruit from their establishments. Young people who were 16 and older were able to provide consent themselves when approached. Questionnaires were administered to the young people in groups (to be completed independently) following a full verbal briefing regarding the purpose of the study. Participants were instructed that the questionnaires evaluated the nature of their friendship groups (both prosocial and antisocial). All participants were informed that participation was voluntary which meant they could leave the study at any time without penalty and were informed that their responses were confidential. They were also informed that their responses would have a code, which would be given to them on their debrief sheet so that their data could be identified and destroyed if they chose to withdraw. Following this briefing, participants were given the opportunity to leave the study if they wished to do so. However, in this case, all participants who were approached agreed to take part. Questionnaires were administered with a research assistant present to provide help if needed. Questionnaires took approximately 60 minutes to complete, after which participants were debriefed verbally and provided with a debriefing sheet which reiterated the purpose of the study, provided information on how to withdraw their data if they chose to do so, and offered the researchers’ contact details should they have further questions.
Results
Data were entered into IBM SPSS Statistics Version 20 where analyses were conducted using a p < .05 level of significance. Reliability analyses were conducted on each scale included in the following section. The scales included: violent crime (i.e., crimes against the person; Esbensen & Weerman, 2005), α = .52; and the Mechanisms for Moral Disengagement subscales (Bandura et al., 1996; moral justification, α = .76; euphemistic language, α = .64; advantageous comparison, α = .86; diffusion of responsibility, α = .76; displacement of responsibility, α = .72; distortion of consequences, α = .77; attribution of blame, α = .64; dehumanization, α = .77). All scales had a reasonable to high internal consistency suitable for publishing applied research (Kline, 1999) with the exception of the scale for violent crime which had a poor internal consistency. 2
Membership
Employing the Eurogang’s definition aforementioned, 25 (13%) participants fit the criteria and were identified as street gang members. The remaining participants (n = 164, 87%) were considered to be nongang youth.
Demographic Characteristics
Gang members and nongang youth were compared on age, gender, and ethnicity. In terms of ethnicity, we split the participants into White (27%) and non-White (73%) for analyses. We conducted a one-way ANOVA comparing the two groups and found no significant differences on age (gang: M = 15.08, SD = 1.89; nongang: M = 15.29, SD = 2.94; F(1, 187) = 0.12, p = .726), gender F(1, 186) = 0.18, p = .669, and ethnicity F(1, 187) = 2.73, p = .100.
Distinct Street Gang Characteristics
The Eurogang Youth Survey (Weerman et al., 2009) contains items that assess the characteristics of groups participants that are members of but not specific to gangs. Participants were asked to indicate whether each item was a characteristic of their chosen group. Presumably, gang members’ responses were in relation to their gang. We conducted chi-square inferential tests to see whether certain characteristics were endorsed more by gang members than members of other types of groups (e.g., sports teams, clubs, or other types of offending groups that did not fit the Eurogang definition; see Klein, 2006, for further discussion). We found that gang members and nongang youth differed significantly on the following group characteristics: recognized leaders (χ2 [1, N = 143] = 7.13, p = .012, φ = .22); symbols (χ2 [1, N = 143] = 8.30, p = .006, φ = .25); specific rules or codes (χ2 [1, N = 143] = 16.29, p < .001, φ = .34); initiation rituals (χ2 [1, N = 143] = 9.70, p = .004, φ = .26); and special clothing (χ2 [1, N = 143] = 6.54, p = .018, φ = .21). On the other hand, gang members did not differ significantly from nongang youth on: gender roles (i.e., boys and girls do different things in the group; (χ2 [1, N = 143] = 1.13, p = .306, φ = .09); regular meetings (χ2 [1, N = 143] = 1.77, p = .252, φ = .11); and tattoos (χ2 [1, N = 143] = 1.33, p = .590, φ = .10). See Table 2 for group proportions of responding.
Gang members’ and nongang youth’s proportion of responses on their group’s characteristics.
*p < .05; **p < .01; ***p < .001.
Violent Crime
We conducted a one-way ANOVA to see whether violent crime varied as a function of gang membership. We found that gang members (M = 6.40, SD = 2.14) were more likely than nongang youth (M = 4.91, SD = 2.17) to report that they had engaged in some form of violent crime within the past 6 months (F[1, 187] = 10.16, p = .002).
Moral Disengagement Strategies
We conducted a final ANOVA to see whether the eight moral disengagement strategies varied as a function of gang membership. We found that gang members were significantly more likely than nongang youth to employ the following strategies: moral justification (F[1, 187] = 10.12, p = .002, η2 = .05); euphemistic language (F[1, 187] = 8.87, p = .003, η2 = .05); advantageous comparison (F[1, 187] = 12.88, p < .001, η2 = .06); displacement of responsibility (F[1, 187] = 6.15, p = .014, η2 = .03); attribution of blame (F[1, 187] = 5.26, p = .023, η2 = .03); and dehumanization (F[1, 187] = 6.06, p = .015, η2 = .03). We found no significant differences between gang members and nongang youth on these remaining strategies: diffusion of responsibility (F[1, 187] = 1.69, p = .200, η2 = .01); and distortion of consequences (F[1, 187] = 0.54, p = .464, η2 = .003). See Table 3 for means and standard deviations.
Means and standard deviations for moral disengagement strategies.
*p < .05; **p < .01; ***p < .001.
Dehumanization as a Mediator Between Gang Membership and Violent Crime
In order to better understand how gang members justify engaging in violent crime we proposed that one of the moral disengagement strategies would best explain the gang member–violence relationship by mediating the relationship. We examined the results of each mediation model (i.e., using the strategies significantly related to gang membership—moral justification, euphemistic language, advantageous comparison, displacement of responsibility, attribution of blame, and dehumanization) and found that the only model that was a good fit for the data was with dehumanization as a partial mediator. We used the Preacher and Hayes (2008) INDIRECT macro in SPSS to conduct the mediation analysis whereby the indirect path was bootstrap tested with 1,000 resamples (bias corrected and accelerated). The total effect (B = 1.49, p = .002) was significant, yet reduced, when the mediator, dehumanization, was included (B = 1.17, p = .011). The mediating path (B = 0.32, p = .036) was significant and the confidence interval did not include zero ([.07, .80]; bootstrapped 95%). This confirms that dehumanization partially accounts for the relationship between gang membership and violent crime. That is, being a gang member is linked to dehumanizing others which partially explains engagement in violent crime.
Discussion
The aim of this study was to empirically test the sociocognitive theory of moral disengagement by examining the extent to which violent behavior carried out by gang-involved youth is accounted for by their use of moral disengagement strategies. In our sample, there was a relatively high proportion of gang members identified when compared to other UK-based research sampling from the community (e.g., Alleyne & Wood, 2010; Sharp et al., 2006). This can be explained by the recruitment of participants from community-based programmes targeting gang-involved youth and youth at risk of joining gangs. There were no differences between gang members and nongang youth in terms of ethnicity, in line with previous work which has suggested that ethnic composition of gangs is representative of the community from which it comes (e.g., Alleyne & Wood, 2010; Bullock & Tilley, 2008). The two groups also did not differ in terms of gender, which, similar to other recent research, further confirms that girls are involved with gangs (e.g., Esbensen, Deschenes, & Winfree, 1999; Miller, 2001). Furthermore, gang members and nongang youth in the sample were found not to differ on age, although this contrasts to previous research which has suggested that there may be a developmental process in gang membership (e.g., Alleyne & Wood, 2010; Thornberry, Krohn, Lizotte, Smith, & Tobin, 2003).
As expected, and in line with previous research (Alleyne & Wood, 2010, 2013; Battin et al., 1998; Thornberry et al., 2003), it was found that gang youth were more likely to report involvement in violent crime than nongang youth. It was also found that gang members, when compared to nongang youth, more often reported that the groups they belonged to had recognized leaders, symbols, specific rules or codes, initiation rituals, and special clothing. Presumably, the groups gang members were describing were their gangs and these characteristics are indicative of a collective identity with group norms that literature would suggest facilitates an increase in gang-related violence (Decker, 1996; Decker & van Winkle, 1996). These characteristics could also be considered forms of communicating this collective identity to outgroups such as authority or rival gangs (Goldman et al., 2014). Nonetheless, they indicate, most importantly, the commitment and cohesiveness of group members. On the other hand, the two groups did not differ in terms of tattoos or gender roles. In fact, few participants in either group reported differing gender roles, so this may suggest that females do not have different roles to males in the gang and are not subordinates in gangs, contrary to arguments seen in previous literature (e.g., Thornberry et al., 2003).
It was found that gang youth employed the use of some moral disengagement strategies more than nongang youth. These strategies were moral justification, euphemistic language, advantageous comparison, displacement of responsibility, attribution of blame, and dehumanization. This supports previous research that has already demonstrated a correlation between gang membership and use of euphemistic labelling, attribution of blame, and displacement of responsibility (Alleyne & Wood, 2010). Gang and nongang youth were not found to differ in terms of diffusion of responsibility or distortion of consequences, contrasting to some previous research which had linked distortion of consequences to gang membership (Alleyne & Wood, 2010). However, Alleyne and Wood (2010) found that it was only peripheral gang youth that displaced responsibility. Perhaps if we had differentiated between the levels of embeddedness amongst the gang sample in our study differences may have arisen.
When we examined the extent to which moral disengagement strategies account for the relationship between gang membership and violent behavior, the findings suggested an underlying process mechanism. It was found that being a gang member is linked to dehumanizing others, which, in turn, partially explains violent behavior. These findings fit neatly within the current social psychological literature on the facilitative role of dehumanization. To put plainly, we treat those we perceive as similar with moral concern, thus empathizing if they are mistreated (Giner-Sorolla, Leidner, & Castano, 2012). However, in order to cause harm we strip away uniquely human qualities from our victims and engage in animalistic dehumanization—a mechanism specific to intergroup conflict as outlined by Haslam (2006, 2014). This disinhibiting process has been found in aggressive and violent contexts including armed conflict (Castano & Giner-Sorolla, 2006), willingness to torture (Staub, 2005; Viki, Osgood, & Phillips, 2013), sexual aggression (e.g., Rudman & Mescher, 2012), and not surprisingly yet important to note, animal cruelty (Gullone, 2012), to name a few.
Although no other moral disengagement strategy was found to account for the relationship between gang membership and violent behavior, this does not mean that we should disregard our findings that gang members used more moral disengagement strategies, not only that of dehumanizing victims, than nongang youth. This is, in fact, very interesting. This makes it seem unlikely that the function of moral disengagement strategies employed by gang members is to cope with the extremity of gang membership—in other words, violence—as has previously been suggested (Alleyne & Wood, 2010). Future research could investigate what other functions moral disengagement strategies serve in buffering gang members from cognitive dissonance. And, most importantly, future research could compare gang-involved youth with nongang youth on measures of dissonance which would clarify whether gang members are even in need of cognitive coping mechanisms and it would clarify the utility of employing these moral disengagement strategies.
There are some limitations to this study. For instance, the use of self-report questionnaires could have resulted in bias due to common method variance. However, as the study intended to measure participants’ experiences and perceptions, self-report was deemed to be the most appropriate method for this purpose (see Chan, 2009). Furthermore, in relation to the drawbacks of the use of self-report measures, this study only measured how willing the participants were to set aside their moral standards by cognitively restructuring their violent behavior in theory. In reality, it may be very different. However, measuring this in practice would likely prove both impractical and unethical.
Additionally, the sample was made up of a combination of secondary school students and young people who attend youth centre programmes, and there are issues involved with using such populations. For example, the sample does not include those individuals that were not in attendance because of sickness or, in the case of the schools, were truant. This may have led to an underrepresentation of gang members in our sample as previous research would suggest that gang youth are indeed likely to truant (Young, Fitzgerald, Hallsworth, & Joseph, 2007). Furthermore, it is possible that those who attended youth centres may not have been representative of those gang youth who were most disengaged.
The poor internal consistency for the violence scale is another issue, but this could go some of the way to explain why dehumanization is only a partial, as opposed to a full, mediator of the relationship between gang membership and violence. The debate on the utility of Cronbach’s alpha is an ongoing one, but we do know that “the greater the number of items in a test then the more reliable it will be” (Coaley, 2014, p. 143). Future research could examine this relationship by making use of lengthier, more robust measures to further evaluate this relationship.
Finally, in terms of limitations, since moral disengagement was measured after the violent acts were committed, we cannot say whether gang members morally disengaged before carrying out acts or if they have done so to make themselves feel better following the act. A longitudinal design would be best suited in examining this relationship. For example, a study could follow youth from early adolescence to measure moral disengagement prior to gang membership or before violent crimes have been enacted.
More research into the role that moral disengagement strategies play for gang members is clearly needed before we are able to develop empirically based interventions. Still, the findings of our study have implications for future research, and, perhaps ultimately, intervention and prevention. The sociocognitive theory of moral disengagement has significant utility in highlighting targets for treatment when working with gang members. For example, these strategies could be working as postevent justifications, and practitioners should target these to further unveil to gang members the harm they caused. Also, the findings of this study suggest that techniques that anthropomorphize the victims of the crime could lead to greater empathetic concern, thus feelings of remorse and guilt. And as a preventative measure, it may indeed be of merit to work with at-risk youth or gang members on reducing the use of dehumanization techniques with the aim to reduce violent behavior. Finally, the behaviors that other moral disengagement strategies may be linked to amongst the gang member population provide an avenue for future research.
The findings of this study highlight the potential added value to be gained by applying alternative theoretical approaches. Wood and Alleyne (2010) argued for an integrated theoretical approach towards the study of gangs. So, for example, gang-related violence, however partially explained by dehumanization, can also be explained by Emler and Reicher’s (1995) reputation enhancement theory whereby gang members are motivated by the desire to build their reputation. There is already some support for this in the literature (e.g., Anderson, 1999). Similarly, future research could look at how gang-related violence could be explained, at least in part, by intergroup processes as outlined by Sidanius & Pratto’s (1999) social dominance theory (SDT). SDT—a theoretical explanation for why and how individuals behave in a way that enhances, or at least reinforces, their place (and the place of their group) within an overarching social hierarchy—has been found to be an important predictor of prison gang involvement (Wood, Alleyne, Mozova, & James, 2013), thus an avenue for future research amongst street gang members.
This study, which examined the extent to which moral disengagement accounts for violent behavior carried out by gang youth, was the first of its kind. Results suggest that gang youth may be morally disengaging from the acts they commit by dehumanizing their victims. It does, however, cast doubt over whether other moral disengagement strategies play a role in facilitating violent crime by gang members, and calls into question what purpose these other strategies have for members of gangs. Most importantly, the findings suggest that “designations of others in terms that humanize them can serve as an effective corrective against aggression” (Bandura, Underwood, & Fromson, 1975, p. 267; see also Bandura, 2002; Čehajić, Brown, & González, 2009). Therefore, our findings show that underlying cognitive processes associated with gang membership and gang violence offer potential targets for intervention.
Footnotes
Funding
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.