“Cause That’s the Only Skills in School You Need”: A Qualitative Analysis of Revenge Goals in Poor Urban Youth

Abstract

Revenge goals are correlated with maladjustment, and retaliation is a factor driving youth violence. However, revenge might be an important social tool adolescents use to achieve (self-)justice in environments where institutionalized interventions from authorities such as teachers or police are limited. This qualitative secondary analysis of 50 revenge scenarios from a larger study (N = 358 youth-caregiver dyads) expands one-dimensional knowledge from closed-answer vignettes to the rich real-world experiences of youth (aged 10–16; 30 boys, 20 girls), from an urban community sample. Ten patterns of qualitative differences in cognition and emotion of revenge scenarios emerged and are discussed within the revised model of social information processing (SIP). Importance of reputation, confidence in non-violent solutions, and the influence of intense emotions were among the themes. Gender differences and implications for prevention are discussed.

A large body of research based on Social Information Processing (SIP) Theory (Crick & Dodge, 1994; Lemerise & Arsenio, 2000) shows that children’s social goals influence their subsequent behavior and that aggressive goals are specifically related to beliefs about aggression, aggressive behaviors, and adjustment problems (e.g., Delveaux & Daniels, 2000). Current SIP theory proposes that there are different processing or attribution styles that influence decision making in social interactions (Crick & Dodge, 1994; Lemerise & Arsenio, 2000), and ample research has supported the notion that aggressive children’s SIP differs from that of their non-aggressive peers in substantial ways (e.g., Lochman & Dodge, 1994). Furthermore, highly aggressive children pursue more negative and hurtful social goals than do less aggressive or non-aggressive children (de Castro, Verhulp, & Runions, 2012). Consequently, revenge goals have been examined in part to understand the SIP of aggressive or rejected children (McDonald & Lochman, 2012). At the core of the initial theory, Crick and Dodge (1994) proposed a heuristic model of a circular series of six steps through which social information is processed, evaluated, and a behavioral response is chosen and enacted. In the first and second steps of the model, individuals encode and interpret different cues in social interactions, for example, the multiple hints that allow an individual to differentiate sarcasm from a genuine compliment. In the third step, individuals select and clarify goals for the interaction, that is, what they want to achieve in the interaction. Based on those objectives, possible responses are constructed (Step 4) and evaluated based on perceptions of consequences and expected self-efficacy in enacting them (Step 5). Ultimately a behavioral response is enacted (Step 6). This final step again requires monitoring of the reaction of the social partner, leading to new cues that need to be encoded and the cycle starts again. The revised theory by Lemerise and Arsenio (2000) expanded the framework by integrating the role of emotions and moral values in influencing the cognitive processes at every step of the cycle.

Researchers repeatedly have demonstrated that reactive aggression is linked to proactive aggression (e.g., Camodeca & Goossens, 2005), that beliefs about the legitimacy of different aggressive behaviors uniquely predict engagement in those behaviors (e.g., Werner & Nixon, 2005), and that children with revenge goals have fewer and poorer quality friendships (e.g., Rose & Asher, 1999; for reviews, see Adrian, Lyon, Oti, & Tininenko, 2010; or Erdley & Asher, 1999).

However, not all children with revenge goals are maladjusted, only a few children with revenge goals enact them, and not all aggression is driven by revenge motivations (McDonald & Lochman, 2012). One explanation for inconsistencies in findings is the uniform use of vignette measures and closed-answer formats in quantitative studies. Prompting all participants with the same hypothetical situation does not control for the interpretation of the situation by the participant, nor account for possible environmental constraints in response choice (Farrell et al., 2010). This ultimately results in a very limited standardized measure of goals. Farrell and colleagues’ (2012) series of qualitative studies on the reasons for a particular choice of response with low socioeconomic status (SES) urban middle-school youth confirmed this. The authors showed that choosing non-violent behaviors in response to vignette situations was not necessarily a sign of good adjustment because it did not account for effectiveness and other important characteristics of the response.

Qualitative Studies on Revenge

Numerous qualitative studies have shown that exerting revenge is embedded in a rich context of personal, emotional, cognitive, and instrumental factors which all play into whether and how an individual acts on their desire to retaliate. For example, studies investigating aggression in samples of girls show that some physically aggressive girls appear very similar to physically aggressive boys (Adamshick, 2010; Letendre & Smith, 2011; Ness, 2004). In most research with middle-class samples, physically aggressive girls are described as maladjusted, unpopular, and socially isolated. In contrast, investigations with girls from a high-violence neighborhood (Ness, 2004) and with adjudicated girls in an alternative school (Adamshick, 2010) showed that physical aggression was viewed as normative and encouraged by mothers as a way to instill a sense of independence and ensure self-protection of their daughters through deterrence in both contexts.

Research on retaliation in boys generally has concentrated on older, deviant samples (e.g., Jacobs & Wright, 2010). Such studies mainly focused on the “code of the street” of offender subgroups and the spread of urban violence through perpetuating patterns of violent retaliation (Jacobs & Wright, 2010; Stewart, Schreck, & Simons, 2006). Only one study with girls strived to develop a typology of different types of retaliation by examining their strategies in response to social aggression, and categorizing different volatile events and corresponding responses (Kozlowski & Warber, 2010). However, the typology was limited to a convenience sample of girls, and responses were limited to relational aggression. In a broader context of conflict resolution, Harris and Walton (2009) analyzed narratives from a community sample of fourth through sixth graders from urban schools. Working from a grounded theory approach, the authors coded 364 written accounts of an experienced conflict collected as part of a writing exercise in an intervention in two inner-city elementary schools. In their study, a relationship emerged between response to conflict and narrative skills, mentioning retaliation was associated with a lowered likelihood of discussing internal states of the actors or of displaying a strong narrative form.

Overall, qualitative research has demonstrated that although revenge is destructive, facets of revenge maintain social equilibrium within a group by helping offended individuals cope with feelings of disempowerment or unjust treatment (Elster, 1990). Qualitative studies have given individuals a voice to illuminate situations where following a code of honor is seen as an adaptive strategy to negotiate highly violent environments. However, despite the information gained, the findings are largely descriptive in nature and focus on special sub-samples. Furthermore, this body of research has not connected the vengeful attitudes and behaviors observed in the qualitative narratives to a well-established quantitative theoretical framework of general cognition, such as SIP.

Revenge in Adolescence

Adolescence might be a crucial developmental stage to explore revenge goals, because youth are developing moral reasoning that follows principles of reciprocal justice and mutual respect rather than following rules laid down by authority figures (Kohlberg, 1963). In addition, peer status becomes increasingly important and image, friends’ support, and peer pressure are key concerns especially in early adolescence (Steinberg, 2005). Finally, one of the fundamental tasks of this stage is to explore and form identity (Erikson, 1968). Thoughts of revenge occur in the context of incidents that threaten identity or reputation, and retaliation can serve the purpose of restoring or protecting reputation in the eyes of peers and the self (Copeland-Linder, Johnson, Haynie, Chung, & Cheng, 2012). Consequently, the increasing approval of revenge strategies could be an important factor associated with youth violence. Findings with adolescent samples are scarcer, but research has shown that revenge goals are correlated with other forms of maladjustment in adolescents (e.g., Lochman, Wayland, & White, 1993).

The Current Study

The main aim of this secondary analysis of qualitative data was to describe different narratives of revenge scenarios in a sample of poor urban youth from high-violence neighborhoods, and to investigate diverse responses to different forms of victimization. First, our sample is a community sample that includes both boys and girls. Prior empirical studies of revenge in juveniles mostly concentrate on special samples of aggressive youth (e.g., Copeland-Linder et al., 2012; de Castro et al., 2012). Second, our study focuses on reasons youth who have the desire to retaliate may or may not enact that desire (Amjad & Skinner, 2008). We analyzed youth’s narratives surrounding desires to retaliate. Narrative analysis is a valid approach in disentangling children’s management of conflict (Harris & Walton, 2009), and it overcomes some of the shortcomings associated with a vignette-based approach to assessing revenge goals. Examining scenarios of revenge in a sample of poor urban youth from high-violence neighborhoods furthers our understanding of different motives for endorsing retaliatory goals in a community sample. The inclusion of scenarios where youth talked about but decided that revenge was not an option they wanted to pursue allowed us to investigate the reasons for those decisions. Third, we used the SIP model as an underlying framework for the study. This allowed us to frame the results in a way that would facilitate the integration of the rich worlds of juveniles into a well-established theoretical framework that systematically describes the different cognitive steps influencing the decision-making process in social situations.

Method

Design of the Study

The present study was a qualitative secondary analysis of a study building on the underlying framework of Lemerise and Arsenio’s (2000) modified SIP Theory. The method of inquiry for the present study was the constant comparative approach used in grounded theory, as adapted for secondary analysis (Heaton, 2004). This study was a supra-analysis of a primary study, exploring new empirical questions that arose from the primary study, but which transcended the scope of the primary analysis (Heaton, 2004).

The primary study

Data for the present study were drawn from the adolescent participants of Project COPE, a four-wave longitudinal study of community and peer violence and substance use in a sample of dyads of urban adolescents and their primary female caregivers in Richmond, Virginia. The sample was recruited from areas classified as having moderate to high violence according to police statistics. To be eligible, dyads had to consist of a fifth or eighth grade child and his or her primary female caregiver during the first wave of data collection; 63% of approached eligible dyads enrolled in the study. Project COPE was built on a model of risk and resilience, with attention to individual, family-level, and community-level protective factors. (For additional information on the procedures of the primary study, see Reid-Quiñones et al., 2011). For a comparison between adolescent participants of the primary and the current study see Table 1.

Table 1.

Sample Characteristics of Primary and Secondary (Current) Studies.

	Primary study	Current study
Participants (N)	Dyads of youth and maternal caregivers (358)	Youth (50)
Gender (%)	Male (46.4%), female (53.6%)	Male (60.0%), female (40.0%)
Grade level (%)	5th grade (53.4%) 8th grade (46.6%)	5th grade (40.0%), 8th grade (60.0%)
Age (M, SD)	9–16 (12.13, 1.62)	10–15 (12.42, 1.69)
Race (%)	African American (>90%)	African American (>85%)
Live-in caregiver most of the time (%)	Biological mother (>85%)	Biological mother (>85%)
	Biological father (20.4%)	Biological father (14.0%)
Weekly income (Median)	Income (US$401–US$500)	Income (US$401–US$500)
Weekly income (%)	Income of US$200 or less (22.8%)	Income of US$200 or less (20.0%)

Measures

Social Competence Interview (SCI)

The current study analyzed transcripts of the SCI (Ewart, Jorgensen, Suchday, Chen, & Matthews, 2002), a 15 to 20 minute audio-taped interview in which youth were asked to re-experience their most stressful event in the past couple of months. The SCI has demonstrated excellent reliability and validity. Students were prompted to discuss situations that involved witnessing or experiencing violence, including peer victimization. As a guide for choosing a stressful event, students were asked to rank categories of stressors from most to least stressful. Eight categories were provided on index cards and included the following: (a) relational victimization by peers; (b) physical victimization by peers; (c) situations involving drugs; (d) situations involving accidents or breaking and entering; (e) situations involving guns; (f) situations involving threats, hitting, or punching; (g) situations involving serious violence that included knives, muggings, beatings, or wounding; and (h) any other situation where the youth felt frightened, or thought they could get hurt very badly or die. These eight categories were based on items from a measure of peer victimization (Problem Behavior Frequency Scales; Farrell et al., 2000) and a measure of community violence exposure (Richters & Saltzman, 1990). After ranking the index cards, the adolescents were asked if they could identify a recent situation that exemplified the category they deemed the most stressful. This situation was then discussed in detail during the first half of the SCI. In the second half of the SCI, youth imagined to be the director of a movie in which a character was in the situation the adolescent described before, but had the opportunity to create his or her own ending to the situation. Finally, the youth were asked to evaluate their solutions and estimate (a) how confident they felt to actually be able to engage in the chosen strategy in a similar situation in the future and (b) how realistically the solution would have the described result. They were asked to quantify those chances on a scale from 1 to 10. Based on this ranking, they could go back and change their ending to a more realistic scenario if they wanted to.

Previous coding

The SCI was previously coded for the type of event, type of coping strategy, emotions expressed during the narrative, and goal of the solution to the movie scenario. One of the emerging goals was revenge, defined as “getting back at somebody physically or emotionally” (Reid-Quiñones et al., 2011). Previous coding, consistent with most existing literature, did not include an exploration of different types of revenge scenarios. In the current study, coders were blind to the previous coding. However, previous classification was used as one measure of quality control for the coding in the present study. This resulted in a sample for the current study that included all youth discussing revenge scenarios identified before, but significantly expanded the sample by including participants who discussed retaliation but decided that revenge was not an option they wanted to pursue as their main goal.

Data Analysis

The following steps of analysis were performed in the present study: Transcripts were sorted into the current study according to Heaton’s (2004) recommendations for conducting secondary analysis. This technique of manipulating and shaping the data ensures that one is working with a data set that fits the secondary research questions (Heaton, 2004). In a second application, this technique can also serve as an adapted form of the theoretical sampling described in grounded theory for secondary analysis (Heaton, 2004; Strauss & Corbin, 1998). We combined both applications in our data analysis strategy. In the first step, all codes for social goals were removed from Wave 1 transcripts. Then, all transcripts were re-read by multiple research assistant coders and sorted into the following two groups: (a) revenge was not discussed and (b) revenge was discussed. The definition of revenge used was whether transcripts contained any accounts, thoughts, or elaborations on scenarios involving “getting back at a person emotionally or physically.” This definition was consistent with the one-dimensional definition used in the majority of previous studies. This process resulted in a sample of 50 transcripts where revenge was discussed. Those transcripts were the sample for the present study. Subsequently, the transcripts were entered into Atlas.ti (2010), a qualitative data analysis software, and sorted into two categories: (a) youth discussed revenge and identified this as the viable strategy to achieve the desired outcome and (b) youth discussed revenge but decided that revenge was not an option they wanted to pursue at the end of the SCI. Finally, type of situation (e.g., peer victimization, witnessed violence) and gender of child were entered into the Atlas.ti program to create subgroups of transcripts (e.g., boys talking about peer violence who decided to pursue revenge). The first author and a research assistant coder trained in qualitative data analysis co-developed a coding framework, by continually comparing emerging themes within and across different subgroups. Based on comparing codes generated by both coders, a list of codes was generated and definitions of each code were written jointly. Disagreements in coding and definitions were solved through discussion; if necessary, a final decision was made by the first author. This procedure was successively expanded from the most descriptive exemplar transcripts to the entire sample of 50 transcripts to validate, saturate, and refine emerging themes. This form of successive data analysis has been suggested as an adaptation of grounded theory’s theoretical sampling for secondary analysis (Heaton, 2004; Strauss & Corbin, 1998).

The structure of the coding framework was partially pre-determined by the structure of the SCI and the Atlas.ti software. This resulted in a hierarchical system of co-occurring codes. For example, some descriptive auxiliary codes only described the place of a statement in the SCI, while other codes were exclusively based on the content of the statement. The full meaning of a statement was thus conveyed by the temporal auxiliary code (e.g., movie) in combination with a content code (e.g., retaliation in kind).

Next, the resulting lower order codes were grouped into larger categories and definitions were written for these categories. The final coding framework consisted of 115 primary codes, grouped into four hierarchical groups of code categories. Using Atlas.ti (2010), overarching themes were identified by the first author through investigation of the data with “queries,” a program feature that allows exploration and sorting of data with combinations of codes (Friese, 2014). In the present study, data were explored using the different higher order categories. Overarching themes were identified in two basic ways: either a content pattern recurred throughout a transcript (e.g., in the solution to the actual narrated situation and in the solution to the movie and in the solution to a future situation) and was seen in several transcripts, or a content pattern recurred across several youths belonging to a specific group (e.g., boys talking about relational peer victimization). To facilitate the systematic presentation of results, the final identified themes were organized in the SIP model. As the SIP model comprises the full cycle of social decision making embedded in emotional and environmental contexts, all final themes corresponded to at least one aspect of the model. Placement of the themes thus expresses the theme’s main feature as agreed on by the first and second authors.

Results

Overarching Patterns Emerging From the Data

Ten overarching patterns of qualitative differences in revenge narratives emerged from the data. Those patterns also differed in manifestations of distortions or adaptations in SIP. For example, the importance of one’s reputation was connected to a focus on potential challenges of managing one’s reputation within social interactions. The findings were organized in the hypothesized steps of information processing described within the framework of the SIP model adapted by Lemerise and Arsenio (2000) to facilitate a systematic overview of the results. Different youth’s narratives could contain multiple different patterns; as all youth are hypothesized to experience the full circular model described in the SIP, we expected to see multiple patterns in complex narratives. In addition, youth could theoretically take initial biases with them through the entire model. Figure 1 provides a visual overview of the emerging patterns as we located them in the SIP model.

Figure 1.

SIP model as adapted by Lemerise and Arsenio (2000) with integrated result patterns.

Appraisal of the event level in the SIP model

Pattern I: Balanced reciprocity (SIP Steps 1 and 6, encoding, behavior enactment, and peer evaluation)

The majority of transcripts consisted of a balanced reciprocity between the degree of violence in the event and in the solution (e.g., an eye for an eye), and retaliation in kind was the norm in this sample. Thus, the majority of physical responses happened in the context of physical aggression. Although peers often encouraged confrontations, or got involved, this typically occurred on the same level of violence. Accounts where this proportionality was gravely violated were narrated as memorable and shock was expressed when proportionality was broken. This is seen in this example when adversaries threatened to bring guns into a fight at a Boys and Girls Club’s basketball court:

I was at the boys and girls club, playing basketball, and um, people, dudes got mad, so one of them just hit me and then we started fighting, and then another dude jumped in, and then all of us were just fighting, and then they left, they was talking about guns and all that … I never would have thought of it before, I just, saw it in someone’s eyes, that they are not over being shot … over a basketball game. (Boy 44)

Even in accounts that involved substantial violence, one of the most common values stated by the narrator was the rule of strict reciprocity. In the rare event of massive escalation, the narrator (whose behavior was escalating) still perceived this rule to be in effect.

Pattern II: Relational victimization for boys (SIP Steps 1–2, encoding and interpretation of cues)

In this sample, a difference in the definition of relational versus physical victimization for boys and girls emerged. While there were no physical components in girls’ accounts of relational violence events, the difference between relational and physical victimization was blurred for boys. One of the salient aspects of relational violence for boys was being ridiculed in front of peers. Repeatedly, this humiliation and exclusion was remembered as happening in the context of playing team sports, such as “while we were playing freeze tag [name] tackled me and he whispered in my ear something mean at me” (Boy 1). However, these physical altercations were not perceived as someone “starting a fight,” but were conceptualized as part of a strategy to humiliate, tease, or “so they can, um, keep me away from friends and make me not have any friends” (Boy 15).

Cognition level of the SIP model

Pattern III: Importance of reputation (SIP Step 3, clarification of goals)

Retaliation and revenge happen in a context, and in this sample where fighting was generally viewed as a normative “necessary evil,” it often also happened as a public event at school. In the transcripts where revenge was discussed as a public event, evaluation by peers was important. For example,

The most important part of it is just showing that I … showing that I have the courage that I’m not going to let nobody push me around anymore…. I mean, that’s what I really want…. Since he embarrassed me, I want to embarrass him. (Boy 32)

In this sense, for some youth, reactions to an insult were part of efforts to communicate strength, build a reputation of being tough, and deter future harm not only by sending a message to the perpetrator but to a wider audience of peers. Some youth were concerned about becoming a regular target if they did not retaliate, especially if they felt they were small, of a different race than the majority, as in this girl’s statement “she thought she could push me around probably because I was White” (Girl 7), or just different. One girl, for example, stated difficulties in her efforts to concentrate on school and abstain from fighting:

I said, “Stop looking at me before I come over there and hurt you.” … And then the girl was like “I’ll punch you in the face, you stupid bitch.” And then I went in my class. So I started shaking and stuff, ’cause I was mad she keep on thinking I’m a joke…. But I snapped out of it because of school is more important. (Girl 27)

The drawback of such a culture is the necessity to engage in fights, preferably with someone who will lose the altercation. Some youth consequently reported an increased pressure to build a reputation. False allegations of a lost fight were thus perceived as stressful:

If I fight him and I lose that is going to be the main thing people are going to be talking about the next day…. That I got into fighting and I lost in front of my friends and after they was gone, it wasn’t gonna upset me because I lost, but see my friends’ friends, they gonna spread rumors (Boy 57)

Pattern IV: Limited generation of response alternatives (SIP Step 4, response access and construction)

Some youth who generated revenge scenarios seemed to have difficulties constructing other response alternatives for their conflicts. However, there were different manifestations in those constraints. Some youth generated scenarios that were comparatively limited in narrating the event or framing the problem. Their narratives indicated a dependence on guidance from adults to regulate their negative emotions and to help them to communicate frustration and conflict resolution within a friendship.

What I really do is try to get them back in a way honestly…. I didn’t let them play. . . . now they started to notice that that was really wrong what they did so they started to apologize … Cause they talked to like the teacher. [That made me feel] really good because they finally noticed that I was mad because they are my best friends so I was waiting for a long time. [And then] we became friends. (Girl 54)

Another group of youth, in contrast, perseverated on their wish to retaliate and often narrowly relived their situation exactly as it had unfolded in real life, including adults interfering and punishing them. For example, this boy who had stated that “I wanted to beat him up so bad, he gets on my nerves” and wanting to “make him all bloody,” restates that talking about the event with his mother and with the interviewer now made him “just want to fight him again.” In his movie and real-life scenario, he would fight the other boy and then “I’ll break them up [by a teacher]” (Boy 29). Narrowly reliving what happened in the event often was combined with perseveration on the desire to retaliate, difficulties in explaining social causality within the narratives, limited use of emotion vocabulary, limited communication on one’s own mental state to the interviewer, and accounts of misunderstandings that suggested problems in recognizing others’ intentions. For example, one girl stated that “’Cause people don’t, most people around here don’t like me, so I’m always trying to find fights” (Girl 49). Another example is this boy’s narrative:

Boy: I should have hit him back … before the teacher came…. Then, after we got on the bus and he kept on punching me [in the shoulder] and I got really mad. And I punched him in the face until he started crying.

Interviewer: You said, you said that they’re cameras on the bus?

B: Mm-hmm.

I: You’re not worried about getting in trouble? …

B: [No] Because I’m defending myself.

I: How is that different than when he was standing in line [waiting for the bus]?

B: The teacher was there. (Boy 52)

Pattern V: Lack of confidence in a non-violent solution (SIP Step 5, response decision)

For those youth who discussed multiple solutions to the event they discussed, a pattern emerged where lack of confidence in non-violent solutions was the determining factor in their decision making regarding their future behavior. A substantial number of adolescents in this study did not believe that appealing to authorities or engaging in other non-violent solutions would yield successful results, although there were statements that involving authority would be the right thing to do. For example, this boy stated,

I tried everything else I could think of … I tried talking to the principal, the teachers, my mom. There was a, I think compliance officers at our school, I tried talking to him … None of that worked [because] he was just … he was the kid at that school that nobody could deal with. (Boy 56)

Overall, girls were more vocal than boys about different retaliation scenarios. A high number of girls discussed peaceful solutions but anticipated that this would not be successful and rationalized how they would have to take things in their own hands:

This is what I’d have happen. I’d have Tasha go to the office, ’cause this way I know is right … If the situation get worse … handle it herself and just take whatever consequences and deal with it, because it’ll blow over … [in reality] I think … we’re just gonna end up fighting. (Girl 42)

The theme appeared for both genders, with peaceful solutions for conflicts involving the same perpetrator in the future rated as the most unrealistic. This is a boy’s statement after successfully ignoring a peer who was insulting and threatening him: “If the same thing happened. You know, I gotta beat him down … Twice in a row. Can’t get none” (Boy 28). Both genders also agreed on the consequences of their actions, with this girl stating that “I will get suspended, probably for ten days, come back to school, be back up there she probably won’t be messing with me” (Girl 34). Even messages from the environment intended to stop bullying seemed to convey the necessity to stand up for one-self. This is an example of a boy’s views on his school’s policy on him fighting back after being provoked by a peer:

She [the teacher] saw it, but that’s how, how they are. They gon be like, “oh he trying to bully you, he trying to bully you.” But if they see that I’m not let him bully me then they won’t suspend me. (Boy 21)

Database level of the SIP model

Pattern VI: Importance of parental messages

As seen in the example above, violence, specifically appropriate retaliation, was experienced by the narrators as being conveyed as appropriate in the school context. In addition, there was substantial support for violence from parents mentioned in the transcripts. It might be the case that youth only discussed school violence with parents who would support their responses to provocation by peers. Still, a pattern emerged connecting messages on fighting or retaliation from parents and other family members and the solutions that were discussed. This is an example of messages of de-escalation:

Then he walked off, and I was trying to go get him, just beat him down. Then I thought about what my brother said, about getting suspended. . . . he said he gonna call me and he gonna make sure that I feel worse than I feel when I get suspended or go to jail. (Boy 28)

The few messages that encouraged violence for boys were rules about how to stand their ground without letting things escalate; however, boys were more reluctant to disclose the nature of social support they received and thus did not talk about maternal messages in depth. In contrast, de-escalation messages for girls only emerged in relational violence events and consisted of mothers’ advice to ignore the mean girls.

Pattern VII: Physical girl violence

A theme repeatedly discussed by the girls in this sample who were involved in physical violence was the normativeness of this experience for them. Accounts of past and future physical altercations with other girls were repeatedly portrayed as things that were just bound to happen once in a while, and there was no stigma attached to girls getting “physical.” Quite the contrary, being a good fighter was discussed as a valued ability. It is important to note that this was confined to fights with other girls; engaging in violence with boys was not part of this. In this example, the girl consulted intensely with numerous friends on her engagement in fighting:

I would through the house all day and talk on the phone. But when I’m on the phone with my friends now, me and my friends don’t like them either. So it’s like somehow I always end up bringing that subject back up because I just want to fight her so bad and get it over with. (Girl 42)

This participant described the importance of having a reputation of being a good fighter, and reported using physical aggression (or at least the talking about it) to express identity. It was not only friends who played an important role in the navigation of physical conflicts, repeatedly, mothers were aware of what was happening, and sanctioned some aggressive behavior. For example, this girl recounted,

[My mom] was like control yourself … as long as she doesn’t let her put her hands on you, and if she does, you know what to do…. Hit her back … She was telling me how to handle my own business. (Girl 27)

Following that advice, she gave the other girl her address to be able to handle her “business” away from the supervised school. Her movie solution was “that she came into my residence, she stepped on my property. We fought and then we solved it out … and the last thing, she’d get beat up” (Girl 27). In another example, the most important thing seemed to be to give the mother a warning before suspension for a fight would happen: “My momma she’ll, she’ll tell me to beat her tail if she hit me…. I just made sure I told her before I was supposed to get to fighting and get suspended too” (Girl 34). In summary, the examples show that overall girls who fought maintained a high level of disclosure with their mothers, who seemed well-informed about their daughter’s lives. Repeatedly, daughters disclosed intentions to their mothers in advance of something happening, as this example shows again: “I already told my momma before that me and [name] weren’t getting really good. And I think something’s gonna happen or whatever” (Girl, 42:11, 68:68). Having a good relationship seemed to consist of talking to their mothers in the early stages of a conflict suggesting the importance of their feedback. Being a good fighter was discussed as a valued part of a girls’ identity and a necessary skill to master.

Emotion processes level of the SIP model

Pattern VIII: Overwhelmed and scared victims

Revenge scenarios in events that elicited fear were discussed differently compared with other events. While fear was referenced only in few school-based events with peers, being scared was the main emotion in events of the few youth recounting community violence or domestic abuse. It was notable that the group of adolescents discussing revenge scenarios in answer to non-peer violence events was dominated by boys. Two boys were talking about experiencing a drive-by shooting. In both instances, the boys were talking about intense fear, but while the first boy was just “happy that I’m still here” because “I could be dead right now” (Boy 6), the second was also talking about anger and feeling “frustration” (Boy 33). Both movie scenarios were similar, in that the shooters would get punishment by the hands of the boys. In the scenario police would come, apprehend the shooters “and then everybody will come over there and beat them up” (Boy 33) before they would be put in jail. Calling the police in real life was explored by both boys, but confidence in that action was very low and they would both instead again “run in the house real fast,” reliving the event. A similar pattern was followed in an event where an older youth in the community took a bicycle from the narrator. Although the events were substantially different, this basic pattern also emerged in the context of domestic violence. A boy talking about being abused by his mother’s new boyfriend stated how he was too scared to tell her what was happening. In his movie, he wanted to stand up to the perpetrator, making himself strong enough to attack him: “Let me see … It’s like, yeah instead of him hitting me, I’m gonna hit him back this time. And, I’m gonna show him what I’m really feeling, you know physically” (Boy 30). The only girl in this group followed that pattern as well, talking about intense fear when her abusive father threatened to shoot her and her mother. Her main goals were the desire to protect her mother and getting out of harm’s way, but she also stated that

I wanted to hurt him ‘cause all the times he hurt me … I wanted him to just feel one time how I felt … tie him up to a tree and just get something and just beat him up or something like that to let him know how I feel. (Girl 51)

Both this girl and the boy mentioned above stated that they had since talked to their mothers who ended the situation. In summary, a small subgroup of youth experienced situations where they were at someone else’s mercy, which made them feel helpless, overwhelmed, and very scared. The revenge scenarios discussed by those youth were fantasies where they would stand up to the perpetrators and shake the role of the victim, while acknowledging that in reality they would not or could not enact such desires.

Pattern IX: Intense rage

A major influence on choice of solution was the experience of intense anger. In answer to the question about how a certain event made the youth feel, numerous youth stated things such as “It made me want to beat them up,” “I felt like killing him/her,” “I want to go and cut her up.” This did not necessarily mean that the youth did or would follow up with those plans, but in context appeared as expression of intense anger, most appropriately translated into “It made me really mad.” In the events involving higher amounts of physical violence, almost all girls and several boys were talking about problems with controlling their temper. One effect of intense rage was that youth would retaliate without considering consequences, with this boy stating that when “just steam was coming [I was] so mad I ain’t even care about [getting suspended]” (Boy 21). The second effect was expressed desire to retaliate despite beliefs that this would be wrong: “I know [fighting them] wouldn’t be a solution, but it’s just something I wanted to do at the point when I was real mad” (Boy 5). Another example is this girl:

People come and touch me, I tell them, “Don’t touch me, … Don’t say nothing else to me until I calm down.”

Because you don’t have to do nothing to me, I’ll snap on you when I’m mad…. I have to think for a while. Sit down. You think you feel better about doing it, but when you look at it and think about it, you shouldn’t’ve did it. You should have just let it slide. (Girl 10)

Physically violent boys reported intense anger mainly without mentioning any attempts to control or regulate their emotions; they just “went” for it without much verbal reflection in their narratives. Girls, in contrast, repeatedly talked about attempts to regulate their emotions. For them, feelings of extreme rage were often connected to statements about loss of control and either needing help with emotion regulation or experience of failure even at assisted attempts of emotion regulation. For example,

I was saying don’t kill her, just beat her, but my mind was saying kill her … I don’t know what’s wrong with me I just I don’t know cause can’t nobody calm me down when I am ready to fight. . . . ’cause when I get mad, I go crazy. (Girl 34)

Pattern X: Emotional numbness

A very small proportion of youth seemed overwhelmed and helpless, and in their narratives this was complemented by emotional numbness. Evident throughout the narratives were an overarching absence of affect, severe difficulties in naming emotions, passivity instead of action, and, subsequently, an indifference to outcomes and absence of goals, which seemed to lead to an inability to think about goal-driven behavior. The most striking components of those narratives were the mention of isolation and complete indifference; the whole narrative consisted mainly of purely passive reaction, and the impression of a deep hopelessness formed: “I don’t know why I’m unhappy ’cause I don’t need no friends. God didn’t put me on this earth for to have friends, so … ” (Girl 12).

Discussion

The main aim of this study was to describe different types of revenge in youth’s narratives of solutions to stressful situations and to investigate diverse responses to different forms of victimization. The 10 overarching distinct patterns of revenge goals to emerge from the data were connected to different points in the SIP model. Qualitative data analysis was not influenced by the SIP model; rather, the underlying framework was used once analysis was complete to assign the emerging themes to their logical place within a theoretical model of social decision making. This approach fulfilled two purposes: (a) it facilitated a systematic discussion of the results and (b) it allowed the qualitative findings to be integrated into a body of well-established quantitative research.

Missing Link of Revenge Goals to Maladjustment

The analysis revealed that a majority of youth were following the general principle of proportionality. The youth who skillfully used this principle seemed better adjusted, that is, did not appear to be socially isolated, suffer from depressive thoughts, or experience uncontrollable violent outbursts of intense anger based on their narratives, compared with their peers who violated this principle and escalated the amount of violence in their scenarios. Most previous studies have focused exclusively on whether youth would engage in retaliatory behavior; SIP research has only investigated the type of retaliatory behavior youth endorse in a limited way.

One partial exception is the study by Amjad and Skinner (2008), which differentiated between two subtypes of beliefs about the amount of retaliation—excessive retaliation beliefs and equal retaliation beliefs—and showed that endorsing different sets of beliefs was connected to the expected difference in self-reported aggressive behavior. However, that study made no connections to the type of event in which specific aggressive behaviors were exhibited. In the present study, youth who engaged in excessive, escalating, violent retaliation, compared with youth who retaliated proportionally, had different beliefs and their narrative abilities were less well-developed. Misinterpretation of the rule of reciprocity, and rigidity in applying it, repeatedly co-occurred with accounts of rejection by peers and comparatively more difficulties in identifying emotions or generating non-violent response alternatives in the narrative. This finding might indicate that the youth who engage in this misinterpretation and rigidity have difficulties appraising social events accurately.

Interestingly, a potentially adaptive form of retaliation scenarios emerged from the narratives of youth who experienced victimization over long periods of time, or in situations where they experienced themselves as helpless. The adult literature suggests that the desire for revenge can be a maladaptive coping reaction in response to injustice that is experienced (Orth, Montada, & Maercker, 2006). In the narratives of this group of juveniles, however, revenge was confined to the imaginary world, forming the impression that imagining themselves in the role of an avenger might enable them to reconcile the incapacitated self with an image of the self as worthy agent in control of its own destiny. As this group consisted mainly of boys, this might be especially important in reestablishing masculinity. Most youth in this group mentioned that they had taken action and worked on resolving the situation to the best of their abilities since the actual occurrence of the event, suggesting that such revenge fantasies might be an expression of cognitive restructuring that is adaptive in helping to re-establish agency in the face of potentially traumatizing experiences.

Youth’s Evaluation of Different Ending Scenarios

In contrast to the leading methodology in the SIP field, whether or not youth decided that revenge was the option they wanted to pursue in their scenarios was of limited importance in this sample; why youth would choose one ending over another ending, or any ending at all, emerged as the question of much greater significance. A significant number of youth discussed non-violent solutions, but gave such solutions low confidence ratings and evaluated them as unrealistic. Not surprisingly, perceptions of institutionalized interventions from authorities such as teachers or police as being non-effective was intimately connected to generating solutions that favored taking matters into one’s own hand. Adolescents’ confidence in the success and feasibility of a non-violent solution was of crucial importance in the choice of such a solution, supporting both earlier SIP research on the importance of beliefs and outcome expectations (e.g., Adrian et al., 2010) and qualitative findings about perceived obstacles to engage in non-fighting versus fighting behavior (Farrell et al., 2010).

In contrast to most SIP research, youth in this study had to generate the solutions and possible outcome goals themselves, and results showed that there were youth who had difficulties in this step. Those difficulties were related to limited verbal skills, accounts of peer rejection, or references to sadness and emotional numbness. This finding mirrors that of Harris and Walton (2009), who noted that children narrating revenge scenarios had overall lower narrative skills and were less likely to report on emotional states than other youth. The SIP research is inconclusive, but some early studies found that aggressive children generate fewer solutions compared with their peers (Lochman & Dodge, 1994). Although recently there has been increased attention on the role of the response evaluation and decision step in the SIP model (e.g., Fontaine & Dodge, 2006), the ability to generate different responses as a necessary prerequisite for this step has largely been ignored. Similarly, the other emerging subgroup of youth who had difficulties in generating alternatives of actions combined with an apparent lag in their ability to develop efficient self-directed non-violent solutions has not been discussed before. These youth might represent a distinct group of vulnerable adolescents who have been overlooked in the literature.

Adolescence and the Code of the Street

Early adolescence is a developmental stage that is marked by an increase in concerns about peer status and reputation, making adolescents especially susceptible to peer influences (Steinberg, 2005). Confirming research conducted with a similar sample (Farrell et al., 2010), in this study peers played a major role in influencing responses to conflicts, and youth generally described a peer climate where fighting was supported. Revenge often was discussed as a public event directed not only at the perpetrator but at witnesses and peers in general; trying to walk away from a conflict was complicated by a simultaneous need to withstand peer pressure. Overall, there was a strong impression in this sample that fighting, revenge, and maintaining a reputation of being a good fighter were a normative experience for participants and their peers. Concern with reputation was a major theme in the revenge scenarios. It was connected to low confidence levels for peaceful solutions because youth believed that non-violent solutions would simply postpone the inevitable. Furthermore, youth endorsed the idea that maintaining a tough reputation would help avoid fighting in the future. This pattern almost exactly replicated findings from a recent study with a diverse sample of rural and urban adolescents who found that 41% of adolescents believed that if they would not stand up for themselves in response to particular types of provocation they would be considered weak and would be subject to continued harassment and victimization (Farrell et al., 2012). The “code of the street” pattern of similar beliefs has been qualitatively investigated as appropriate adaptation for particular high-violence environments. A recent longitudinal study, however, found that African American adolescents endorsing this mentality were engaged more often in violence, and, instead of deterring future harm, were at an increased risk of becoming victims of retaliation (Stewart et al., 2006). Furthermore, Bettencourt and Farrell (2013) found that well-adjusted youth from the same school environment did not share perceptions regarding the necessity of aggression with youth who were concerned with preserving a tough reputation. Results of the present study dovetail well with this research. Youth who were most concerned with their reputation most often discussed physically violent scenarios in their narratives, and referenced more problems with chronic involvement in physical violence compared with youth for whom reputation was less salient. It seems likely that in the present sample, importance of reputation beliefs also placed youth at an increased risk of repeated engagement in aggression, thus increasing their risk of continuing victimization and for experiencing sanctions and jeopardizing their academic achievement.

While the importance of peer messages about fighting was a main influence, peers were not the only source of messages about fighting. Numerous narratives indicated the high importance of parental messages, especially as source of support for abstaining from retaliatory behavior. A substantial number of youth reported that their caregivers knew of or even encouraged their retaliatory behavior, while almost all youth abstaining from retaliation referenced a close family member who would support this decision. Although it is well-established that parental influence declines with the beginning of adolescence as peers gain importance (Steinberg, 2005), findings in the current study are in accord with other research on retaliatory attitudes in adolescents: Copeland-Linder and colleagues (2012) found in their study of assault-injured African American adolescents that adolescent’s perceptions of parental attitudes about fighting had the strongest impact on retaliatory attitudes. Similarly, Farrell et al. (2010) found that the most often mentioned deterrent to fighting were parental values against fighting.

Normative Girl Violence

While some gender-specific patterns emerged in the study, it was notable that the level of physical violence for a stable subgroup of girls did not differ from the level of physical violence talked about by boys in this sample. A prevailing theme for the girls involved in physical violence in the present study was the normativeness of the experience for them. There was no reference to any stigma attached to girls getting “physical.” To the contrary, accounts of past and future physical altercations with other girls were repeatedly portrayed as an inevitable part of navigating their environments. In addition and in contrast to boys, however, mothers appeared as well-informed actors who encouraged their daughters to stand their ground and engage in physical fights when necessary. In qualitative studies, similar patterns of normative fighting among girls were discovered, often similarly closely connected with positive maternal messages about fighting (Adamshick, 2010; Kozlowski & Warber, 2010; Letendre & Smith, 2011; Ness, 2004). For example, in an ethnographic study of street-fighting in girls from poor neighborhoods in Philadelphia, mothers were reported to play an integral role in their daughters’ use of violence (Ness, 2004). In those neighborhoods, girls were “socialized from a young age to stand up to anyone who disrespects them and hold their own” (p. 37), and mothers encouraged aggression as a way to instill a sense of independence and ensure self-protection of their daughters.

Other qualitative studies on revenge and fighting in girls concentrated on triggers for fighting and special topics that would ignite retaliation (Kozlowski & Warber, 2010; Letendre & Smith, 2011). In the present sample, the importance of loyalty and friendship identified in the Letendre and Smith (2011) study was repeated, although more in the context of relational violence than as trigger for physical fights.

Emotions and Goal-Directedness of Aggression

A subgroup of youth in the present study with increased violent scenarios and endings recounted uncontrollable rage, loss of control, and disregard of consequences. Sometimes these youth developed endings with harmful results for the self or even talked about revenge despite this being in conflict with their beliefs. This is interesting in the light of newer research that has called the goal-directedness of aggressive behavior in question, at least on the subjective level of experience. de Castro and colleagues (2012) found that both highly aggressive and a normal comparison group of boys explained their aggressive responses in provocation situations as driven by, in their words, feelings of uncontrollable rage and failure to control those feelings, rather than referencing a goal they wanted to achieve. de Castro and colleagues’ sample was restricted to younger boys, but it seems their findings translated well to the girls in the present sample: Almost all physically violent girls discussed feelings of extreme anger and the inability to control their own emotions. Reports of intense anger have been documented as driving factors leading to excessive retaliation and a disregard for consequences in several qualitative studies of retaliation of males (Jacobs & Wright, 2010) and females (Ness, 2004).

Another group of youth where emotions strongly influenced retaliatory endings was those who seemed numb and disconnected from the stressful events they described. Those youth also seemed indifferent to possible outcome goals, had difficulties describing any alternative endings, and in their solutions, instead, they relived the events in which they had reacted violently as they had happened.

Overall, the findings highlight the importance of including emotional states into SIP research (Lemerise & Arsenio, 2000). The emotional patterns also confirm that subjectively, emotional states of intense anger might lead to aggressive behavior that overrides directed outcome orientation, while an inability to experience emotions or numb disconnect leads to behavior that seems void of outcomes. Findings showed that dysregulated emotional states even lead to aggressive behavior despite the awareness of outcomes that are harmful for the self. This validates a new line of SIP research across both genders and points to potential for further investigation of the connection between absence of experienced emotions and goal-orientation.

Limitations

The present study had several limitations. The structure of the SCI limited the type of information that was elaborated on by the juveniles, and might have forced them to conceptualize the stressful events they were choosing to talk about in a certain way; for example, the focus on the detailed emotional experience might have forced attention to negative affective states as opposed to coherence in the narrative.

Due to its nature as secondary data analysis, there were varying degrees of saturation across the different patterns, and there was one emerging pattern where limited saturation was a substantial concern: It seemed that there was a theme of special maternal messages of fighting with boys emerging, but due to the limited explicit focus and boys’ overall reluctance to elaborate on their social support–seeking behaviors (in contrast to some girls), there was not enough depth in the interviews of the boys to fully develop that theme. Still, based on the limited information, it seemed that mothers of physically violent boys were more likely to relay messages intended to limit fighting behavior in comparison with the messages coming from mothers of physically violent girls. Given the higher levels of serious threat perceived by the boys in this sample, it could be that this threat was more salient for mothers of boys as well. Thus, mothers could be trying to communicate more messages of de-escalation out of fear of their boys’ getting caught in serious violence or gangs, and becoming victims of retaliation.

Another limitation was that sometimes the main focus of the narrative was on hypothetical actions, resulting in limited explicit information regarding the existence of different goals. In addition, the scenarios were self-reported and might have been influenced by adolescents’ desire to inflate status or fulfill aggrandizing fantasies that compensate for experienced vulnerability. While this poses challenges from an SIP perspective, it can be argued that this is approximating the natural context of creating narratives. Similarly, whether and which different alternative solutions were discussed were in part a result of the particular interaction with the interviewer. However, there was high internal consistency of most themes, and youth repeatedly came back to how they framed the problem they were discussing in the very beginning of the interview throughout their narrative. Limitations in narrative skills of some youth made it hard to understand how they constructed the explanation for their behavior in detail. However, this limitation corresponds with findings that children talking about revenge showed poorer narrative skills (Harris & Walton, 2009), and with research that shows impaired SIP in aggressive youth and youth endorsing revenge goals (Adrian et al., 2010; Crick & Dodge, 1994; Dodge & Rabiner, 2004; Lochman et al., 1993; McDonald & Lochman, 2012). Overall, although the SCI structure was limiting in some instances, it also provided a standardized format for describing events, emotions, and outcome expectancies in future situations, allowing the comparison of interviews for salient themes across youth. The structured format made differences in narrative skills more apparent, allowed for a detailed analysis of different emotions connected to an event, and highlighted the relative ability to talk about emotions of a participant. Finally, despite these limitations, the SCI’s structure made it possible to match the findings with the SIP model. Locating the themes that emerged within the SIP model facilitated linking these qualitative findings to an extensive body of quantitative research while still preserving the richness of the qualitative data.

Summary and Directions for Future Research and Prevention

Overall, the analyses suggested that risk of negative adjustment in this sample was most evident in adolescents who had difficulties managing their emotions and staying within the boundaries of strict reciprocity in kind. Furthermore, adolescents whose narrative skills were limited also appeared to be at risk of negative adjustment.

Despite its limitations as secondary data analysis, this qualitative analysis validates previous research about deficits in SIP of aggressive youth and youth with revenge goals (for a review, see Adrian et al., 2010). However, it also points to new directions and highlights gaps in the existing SIP literature: More assessment of the context and content of revenge goals seems warranted.

At minimum, future research on revenge goals should assess the extent to which a youth wants to “get back” at another person. Is the youth talking about retribution in kind or escalating the infliction of pain? If predictions about adjustment are to be made, goals should be assessed in relation to their conformity with a given culture, using an environment-fit model. Parental messages about fighting seem of crucial importance in specific cultural contexts and should be included in further research. Namely, further exploration of the undersaturated parental messages about fighting could combine in-depth qualitative interviews with the development of a questionnaire measure in a mixed-methods approach. Results of this study could further serve as a valuable starting point for future quantitative work. Specifically, the inclusion of underlying motives in studies of social and revenge goals would allow researchers to explore the distribution of the patterns identified in this study in a new population. Similarly, the predictive relevance of the patterns identified in this study for well-established measures of adjustment could be investigated quantitatively.

The present study points to the importance of school-based prevention initiatives, as this was overwhelmingly the most common stage for stressful events and retaliatory behavior. It also seems crucially important to assess and integrate both youth’s and parents’ beliefs about fighting, and the school climate around fighting, into all prevention programming. The present study further points to the importance of assessing peers’ perspectives, as peers were the primary intended audience and judges of retaliatory behavior. Finally, using narratives to explore the nuanced construal of justifications for vengeful behavior, as we did in the present study, could provide valuable starting points for more effective interventions. After all, success in altering cognitions that lead to a behavior is more likely if interventionists target the cognitions that actors perceive as causes for their problematic behavior.

Biographies

Lena Jäggi is a Swiss citizen and received her master of law in criminology and criminal law from the University of Berne, Switzerland, in 2010. In 2014, she received her MS in psychology from Virginia Commonwealth University, where she is currently pursuing her PhD in applied developmental psychology. Her research interests include program evaluation in the context of aggressive behavior, violence, and delinquency prevention for youth, as well as the impact of exposure to violence, aggression, delinquency, and incarceration on adolescent adjustment across different cultural settings.

Wendy Kliewer is professor and chair of the psychology department at Virginia Commonwealth University. A stress and coping researcher with developmental and prevention interests, her work examines the negative sequelae associated with chronic stressors such as poverty and community violence, and the role of the family in mitigating stress responses in youth.