Appetitive aggression and adaptation to a violent environment among youth offenders

(1)

Appetitive Aggression and Adaptation to a Violent Environment Among Youth Offenders

Roland Weierstall and Martina Hinsberger

University of Konstanz

Debra Kaminer and Leon Holtzhausen

University of Cape Town

Solomon Madikane

Realistic, Cape Town, South Africa

Thomas Elbert

University of Konstanz

Previous research has demonstrated that those who perpetrate violence and report aggression to be appetitive are less vulnerable to developing posttraumatic stress disorder (PTSD). The current study examined the hypothesis that, among youth offenders living in a high-violence community, an attraction to violence is associated with improved psychosocial functioning and less concern about future threats. A sample of 69 male youth offenders living in two high-violence communities in Cape Town, South Africa, completed measures of trauma exposure, PTSD symptoms, appetitive aggression, reactive aggression, psychosocial functioning, and concerns about future threats. Higher PTSD symptoms predicted lower psychosocial functioning as well as more concern about future threats. However, participants high in appetitive aggression showed better functioning and fewer concerns about future threat.

By contrast, reactive aggression was a signiﬁcant predictor of reduced functioning and increased concerns about threat. These ﬁndings suggest that, for youth offenders living in contexts of ongoing violence exposure, appetitive aggression may serve a psycho- logically self-protective function. Implications for interventions with youth offenders living in high-violence communities are considered.

Keywords:appetitive aggression, youth offenders, PTSD, South Africa

ROLANDWEIERSTALLis an assistant professor of clinical psychology at the University of Konstanz, Germany. His research focuses on the study of the human attraction to violence, using measures on the subjective, behavioral and physiological level. In his work with different combatant populations, he has investigated the functionality of violent and cruel behavior in humans and their relation to trauma- spectrum disorders.

MARTINAHINSBERGERis a graduate student in clinical psychology and psychotherapy at the University of Kon- stanz, Germany. Her research focuses on the determinants and consequences of appetitive aggression in juvenile offenders in South African townships and the development of an integrative treatment approach based on the Narrative Exposure Therapy.

DEBRAKAMINERis a senior lecturer in the Department of Psychology and director of the Child Guidance Clinic at the University of Cape Town, South Africa. She teaches and supervises clinical psychology trainees working with trauma survivors, and she has conducted research on trauma exposure and sequelae in the South African popu- lation.

LEONHOLTZHAUSEN is a senior lecturer in the Depart- ment of Social Development, University of Cape Town, South Africa. He teaches and supervises at both the under- graduate and the postgraduate level, with a speciﬁc focus

on criminal justice practice and substance abuse. His research focuses on intimate partner violence, the process of violentization, criminal justice social work, and transna- tional social work.

SOLOMONMADIKANEis the founder and executive director of Realistic, a community-based organization in Cape Town, South Africa, which supports ex-prisoners and youth at risk with training and life skills to prevent their reoffending.

THOMASELBERTis professor of clinical psychology and neuropsychology at the University of Konstanz, Germany, and an internationally renowned expert in the study of traumatic stress and aggression. Beyond clinical measures, he uses neuroimaging and epigenetic techniques to study the cycle of violence. With his colleagues Frank Neuner and Maggie Schauer, Elbert developed the Nar- rative Exposure Therapy, a culturally universal short- term intervention based on the neurocognitive theory of traumatic memory.

Supported by the Deutsche Forschungsgemeinschaft (German Research council [DFG] EL 101/29-1).

CORRESPONDENCE CONCERNING THIS ARTICLEshould be addressed to Roland Weierstall, Department of Psychol- ogy, University of Konstanz, Feursteinstra␤e 55, 78479 Reichenau, Germany. E-mail: roland.weierstall@uni- konstanz.de

138

Konstanzer Online-Publikations-System (KOPS) URL: http://nbn-resolving.de/urn:nbn:de:bsz:352-262003

https://dx.doi.org/10.1037/a0032489

(2)

Although South Africa has experienced a reduction in levels of politically motivated violence since the demise of the apartheid system in the 1990s, rates of homicide, gender-based violence, and gang-related violence are cur- rently among the highest in the world (Altbeker, 2007; Seedat, Van Niekerk, Jewkes, Sufﬂa, &

Ratele, 2009). Almost a third of adult South Africans have been a victim of violence, and over half have experienced multiple exposures to potentially traumatic events (Kaminer, Grim- srud, Myer, Stein, & Williams, 2008). How- ever, South African youth are at particularly high risk for exposure to violence. Youth between 12 and 22 years of age are twice as likely as adults to be victimized by crime and violence (Burton, 2006). Youth between the ages of 10 and 29 years account for 48% of all violent deaths in the country (Foster, 2012), with rates of youth homicide being especially high in Cape Town’s impoverished townships (Seedat et al., 2009). In addition to the developmental challenges posed by high levels of poverty and inequality, many South African youth also grow up amid high levels of threat and danger to which they must adapt. Although levels of violence exposure among South African youth may be particularly high, it is evident that youth living in other contexts also have to ﬁnd ways to cope with daily exposure to violence (Hunt, Martens, & Belcher, 2011;Kennedy, 2008).

Although the majority of survivors of single traumas do not develop posttraumatic stress disorder (PTSD; Breslau, 1998), research has indicated a dose-response relationship between trauma exposure and PTSD: Increased exposure to traumatic stressors is associated with an increased risk of PTSD, and beyond a cumulative threshold of exposure almost all survivors will develop the disorder (Kolassa et al., 2010;

Neuner et al., 2004;Schauer et al., 2003;Steel et al., 2009). The increasing likelihood of PTSD with higher exposure to traumatic stressors has been termed thebuilding-block effect. The dose- response effect reported in adults has also been documented for children (Catani, Jacob, Schauer, Kohila, & Neuner, 2008;Catani et al., 2009). However, because not all youth living in conditions of violence develop trauma symptoms and reduced psychosocial functioning, there must also be factors that contribute to resilience against the negative effects of cumulative trauma exposure.

Recent studies have started to investigate the perception and processing of violence among perpetrators. It has been argued that cruel behavior offers an evolutionary advantage and an associated gratification in perpetrators (Jones, 2008;Nell, 2006). If cruel behavior evolved as a beneficial by-product of evolutionary adaptation, there must be a different perception of violence cues among those who perpetrate serious atrocities (Elbert, Weierstall, & Schauer, 2010). The infliction of harm on a victim for the purpose of experiencing violence-related enjoy- ment beyond secondary rewards like status or material benefits has been termedappetitive ag- gression(Elbert et al., 2010). It is motivated by approach behavior toward a positive emotional state aroused by violence cues.

By contrast, its counterpart, calledreactiveor facilitative aggression, occurs in response to a threat or aversive stimulus, and is motivated by an avoidance of a negative emotional state like fear or distress (Weierstall & Elbert, 2012). A number of studies have revealed that appetitive aggression is not only a common phenomenon among former combatants (Weierstall & Elbert, 2011), but is also associated with a reduced risk of trauma spectrum disorders. Across different samples it has been demonstrated that those who report higher appetitive aggression show fewer posttraumatic stress symptoms, accounting for up to 12% of the variance of PTSD symptoms (Weierstall, Schaal, Schalinski, Dus- ingizemungu, & Elbert, 2011;Weierstall, Scha- linski, Crombach, Hecker, & Elbert, 2012).

Hence, those who describe perpetrating violent offenses as appetitive or pleasurable demon- strate a better adaptation to environments characterized by high levels of violence.

It has been reported that reactive aggression can also be increased in those who perpetrate violence. Among war veterans who suffer from PTSD, reactive aggressive behavior and an increase in anger and hostility have been demonstrated, and are often associated with psychosocial problems (Byrne & Riggs, 1996;Jakupcak et al., 2007). However, we suggest that the underlying mechanisms between the two forms of aggression are opposed: Whereas appetitive aggression is related to a positive valence, the heightened aggressive behavior in those who suffer from PTSD is a consequence of a permanent hyperarousal and the increased perception of fear, which makes defense reactions out of

(3)

anger or fear more likely to occur— often in situations in which this type of responding is inappropriate. Consequently, reactive aggression should not improve adaptation to a violent environment, but rather cause further psychosocial problems.

The aim of the present study was to investigate whether appetitive aggression serves as a protective factor for the development of psychosocial dysfunction and the expectation of future danger among male youth offenders living in a high-violence context. It was hypothesized that those who are attracted to perpetrating violence are less likely to develop psychosocial dysfunction and will show a lower concern about experiencing future danger. To investigate differences in the impact of the two different forms of aggressive behavior, we also controlled our analyses for reactive aggressive behavior.

Method Sample

A sample of 70 male participants, residing in two low-income communities in Cape Town, South Africa, participated in the study. They were between 13 and 27 years of age (M⫽20, SD ⫽ 3). The ﬁrst language of all participants was isiXhosa, an indigenous South African language, while levels of English ﬂuency varied within the sample. The demographic data of the sample is presented inTable 1.

The study formed part of a pilot project to investigate mental health in former young offenders in two urban townships of Cape Town, South Africa. The South African townships emerged during apartheid, the system of insti- tutionalized racial segregation that characterized South Africa during the rule of the National Party from 1948 –1994. These townships are a legacy of apartheid’s geographic separation of South Africans along racial lines, and remain characterized by high unemployment, informal housing, and a lack of basic infrastructure (Christopher, 2005;Turok & Watson, 2001), as well as high rates of criminal violence (Demom- bynes & Özler, 2005) and domestic violence (Dinan, McCall, & Gibson, 2004).

Participants in this study were members of a community-based intervention program that aims to prevent reoffenses among ex-inmates as well as nonincarcerated former young offenders who have been diverted from the justice system by the state. Over a 6-month period, members of the organization take part in nine programs that offer social and psychological support as well as skills development. The program director, together with the researchers, explained the study aims to all youth attending the organization’s program in the last quarter of 2011 and the ﬁrst quarter of 2012, and all were invited to participate. It was further explained that participation in the study was voluntary and that declining to participate would not impact on participation in the organization’s program, and that participants would receive no ﬁnancial compensation.

None of the potential participants declined to participate.

Measures

Traumatic event load. For the assessment of previous exposure to potentially traumatic or threatening events, a trauma checklist was administered. The 32-item checklist included all trauma exposure items from the UCLA Trauma Checklist (Rodriguez, Steinberg, & Pynoos, 1999) as well as items from the Children’s Exposure to Violence Checklist (CEVC;

Amaya-Jackson, 1998), a self-report checklist that assesses children’s exposure to violence as witnesses or victims. The CEVC has demonstrated excellent internal reliability in both U.S.

(Fehon, Grilo, & Lipschitz, 2001) and South African samples (Fincham et al., 2009). The Table 1

Sociodemographic Characteristics of Sample (N⫽70)

Characteristic M(SD), [range] orn(%)

Age (years) 20.3 (2.9) [13–27]

Marital status (%)

Single 19 (27.1)

Partner 51 (72.9)

Number of children

0 56 (80.0)

1 13 (18.6)

2 1 (1.4)

Years spend in school 9.8 (1.9) [5–14]

Education

No school 1 (1.4)

Primary school 65 (92.9)

Secondary school 4 (5.7)

Imprisonment (yes/no) 36 (51.4)/34 (48.6)

(4)

checklist includes items about direct victimiza- tion or threat to physical integrity (e.g., “Have you been in a bad accident, like a very serious car accident?”, “Have you actually been beaten up by a family member?”, “Has someone out- side your family threatened to kill you?”) and witnessed events (e.g., “Have you seen some- body pull a gun on another person in your neighborhood?”, “Have you seen someone else be forced to do something with their private parts they didn’t want to do?”, “Did you see someone being killed by another person in your neighborhood?”). The checklist was administered verbally in the form of a structured interview. Each item was scored dichotomously, that is, 0 if the participant was not previously exposed to the event and 1 if the participant had been exposed to the event. For the analysis, all events were summed. Participants reported that they had experienced a range of between 8 and 29 different events (M ⫽ 19.8, SD ⫽ 4.1), indicating that all participants were exposed to multiple potentially traumatic events. The reliability of the CEVC in our sample was sufﬁ- cient (Kuder–Richardson alpha⫽ .79).

PTSD symptom severity. For the assessment of PTSD symptom severity, the UCLA PTSD Index (Rodriguez et al., 1999) was administered. This 20-item measure assesses PTSD symptoms according to the diagnostic criteria of theDiagnostic and Statistical Man- ual of Mental Disorders, Fourth Edition (DSM–IVTR), and its validity for African samples has been established (e.g., Ellis, Lhewa, Charney, & Cabral, 2006; Hermenau et al., 2011). It has been speciﬁcally designed for the valid assessment of PTSD in children and adolescents, and it asks for PTSD symptoms during the past month. Participants were asked to select the worst event that they had experienced from those endorsed on the Traumatic Event Load scale, and to answer the symptom items with reference to that speciﬁc event. Items are rated on a 5-point Likert scale, ranging from 0 (none) to 4 (almost every day). This self-report scale can be administered either verbally or in paper- and-pen format. To improve external validity, a clinical psychologist administered the items in a structured interview. The items in the UCLA PTSD Index can be either summed up for a severity score or scored categorically, yielding a likely diagnostic grouping according to the DSM–IV. In this study, the item scores were

summed for a PTSD symptom severity score.

The UCLA PTSD Index has demonstrated strong convergent validity (Rodriguez, Stein- berg, Saltzman, & Pynoos, 2001) as well as a high internal consistency, indicated by alpha coefﬁcients (.90; Layne, Stuvland, Saltzman, Djapo, & Pynoos, 2001). For the present sample, Cronbach’s alpha was .88.

Psychosocial dysfunction. The Strengths and Difﬁculties Questionnaire (SDQ;Goodman

& Scott, 1999) was used to assess psychosocial dysfunction. It is a brief 25-item behavioral screening instrument, designed for children aged 4⫺16 years, with demonstrated reliability (Goodman, 2001). Compared with other measures, it allows for the broad assessment of internalizing and externalizing symptoms. It consists of ﬁve scales that assess emotional, conduct, hyperactivity, peer relations, and social behavior problems. Originally, it was developed in the United Kingdom (Goodman, 2001), but has been adapted for different coun- tries and languages and has demonstrated good cross-cultural validity (Mullick & Goodman, 2001). Symptoms can either be assessed as self- ratings or ratings by teachers and parents. In this study, the items were administered verbally to participants as part of a structured interview.

The severity of each symptom can be scored with 0, 1, or 2 points. The symptoms can either be summed separately for each of the ﬁve scales or into a total sum score for the scale. In this study, we used the total SDQ sum score as an overall measure of psychosocial dysfunction.

Concerns about future threat. As a further measure of the enduring possible consequences of traumatic and threatening events, participants’ concerns about potential future threatening events were assessed using the Traumatic Event Load scale described above.

For each item, in addition to stating whether they had previously experienced the event, participants were also asked if they were wor- ried that this event would happen to them again in the future. As with the items about prior exposure to potentially traumatic events, these items were rated dichotomously and then summed. Participants reported between 2 and 24 different potentially traumatic events that they were concerned would happen to them in the future (M ⫽10.8,SD ⫽5.8).

Appetitive aggression. Attraction to violence was assessed with the Appetitive Aggres-

(5)

sion Scale (AAS; Weierstall & Elbert, 2011).

The AAS was designed to assess the concept of appetitive aggression, and its sound psychometric properties have been established, indicated by a Cronbach’s alpha of .85 and a single-factor structure, accounting for 32% of the variance. It has been administered to more than 2,000 participants, mainly perpetrators of severe violence, in different conflict regions (e.g., Uganda, Rwanda, Colombia). It consists of 15 items (e.g., “Is it fun to prepare yourself for fighting?”, “Does it fascinate you if someone is hurt in a fight?”, Once you got used to fighting, did you want to inflict more and more pain?”).

For every item, participants are asked to re- spond to the statement dealing with their pro- pensity toward violence using a 5-point Likert scale, ranging from 0 (I totally disagree) to 4 (I totally agree). Cronbach’s alpha in this sample was .85.

Reactive aggression. For the assessment of reactive aggression, the reactive subscale from the Reactive⫺Proactive Aggression Question- naire (Raine et al., 2006) was chosen. It has been validated in different populations and proven its factorial validity as well as its satis- factory reliability. It is the only available questionnaire that speciﬁcally addresses reactive aggression. From the 23-item questionnaire, we selected the 11 reactive items. Participants were asked how often they exhibited different types of reactive aggression in the past month. For each item, response options were never (0), sometimes(1), oroften(2). The score for reactive aggression was calculated by the sum of all items, which ranged in this sample from 4⫺21 points (M ⫽ 13.71, SD ⫽ 4.57). Cronbach’s alpha in the sample was .78.

Experiences of violence perpetration.

Participants completed a list of 16 items related to past violence perpetration, which forms the ﬁrst section of the AAS. The perpetration items range from physical assault to murder to sexual assault. An item was scored as 1 if the participant committed the violent act and 0 if he did not. The participant’s answers were summed. In this sample, the summed score ranged from 1⫺16, with a mean score of 10.2 violent acts committed (SD ⫽ 4.1). The reliability of this scale was high (Kuder–Richardson alpha ⫽ .87).

Procedure

Prior to administration of the study, all participants were given a written informed consent letter, which was translated into isiXhosa with the assistance of an interpreter. The interpreter also read the consent letter aloud to participants, due to varying levels of education and literacy in the sample. For participants under the age of 18 years, parents or legal caretakers also gave written informed consent prior to administration.

All questionnaires were translated from Eng- lish into Xhosa by a bilingual speaker and back- translated from Xhosa into English by a second bilingual speaker. The back-translations were compared with the original English versions by a panel of bilingual translators to ensure equiv- alence with the original standardized instru- ments, and any discrepancies were resolved through consensus. The questionnaires were administered to participants by a clinical psychologist, together with a bilingual interpreter, who was trained to translate the structured interviews. Six interpreters assisted with administration, and all received prior training in general mental health issues as well as in the concepts of trauma and aggression.

Participants were interviewed at the organization’s premises. Interviews took 90⫺120 min to complete. In cases in which suicidality or acute mental health problems were reported by participants, they received further counseling or referrals to the local psychiatric health service.

Data Analysis

Linear statistical modeling was used to test our hypothesis and to explore the relation between aggressive tendencies, exposure to potentially traumatic events, symptoms of traumatic stress, and the two dependent variables of psychosocial dysfunction and concerns about future threat. Akaike information criterion (AIC;

Akaike, 1987) was used to select the best-ﬁtting models. For the full models, all possible two- way interactions and squared terms were added as predictors. To mitigate multicollinearity, variables were mean-centered before interactions were calculated. To detect signiﬁcant R² changes, analyses of variance were calculated.

Data analysis was conducted using SPSS for Mac, Version 20.0.

(6)

Results

Table 2 shows the two-way correlations between all dependent variables that were assessed in the sample. Self-committed violence correlated with traumatic event load and showed the strongest correlations with the two measures of aggression, but had no signiﬁcant correlation with the other two measures of current symptomatology. Except for the relation between appetitive aggression and concerns about future threat as well as PTSD symptom severity and reactive aggression, all the other variables were signiﬁcantly and positively correlated. To test our research hypothesis and to discover the complex relations between these variables in the prediction of psychosocial dysfunction and concerns about future threat, two linear regression analyses were calculated.

Predicting Psychosocial Dysfunction From Aggressive Behavior, Trauma Exposure, and PTSD Symptoms

A multiple linear regression analysis was performed to test whether appetitive aggression could reduce the risk for psychosocial dysfunction as a consequence of potentially traumatic events. Psychosocial dysfunction, indicated by the SDQ sum score, was therefore reduced on the two trauma-associated variables—PTSD symptom severity and traumatic event load—as well as the three aggression-related variables—

appetitive aggression, reactive aggression, and self-committed violence—and all possible two- way interactions. For the ﬁnal model (seeFig- ure 1), all simple effects as well as both two- way interactions between appetitive aggression with self-committed violence and reactive ag-

gression with self-committed violence were kept in the model. The selected model explained 47% of the variance in psychosocial dysfunction,Radj2 ⫽.47,F(7, 62)⫽9.89,p⬍.001,f²⫽ .89, and, because of the large effect size, also had sufﬁciently high statistical power, even though the sample size was comparably small, (1⫺ ␤)⫽.99 (Faul, Erdfelder, Lang, & Buch- ner, 2007). The regression model revealed a main effect for PTSD symptom severity and traumatic event load. As expected, participants who reported more severe trauma symptoms also reported greater psychosocial dysfunction.

The same was true for trauma exposure: Partic- ipants who had experienced more potentially traumatic events also reported greater psychosocial problems, independent of PTSD symptoms.

The relationship between the aggression variables and psychosocial functioning was more complex. On the one hand, there was a main effect for the factor reactive aggression, a significant positive interaction by trend between self-committed violence and reactive aggression, as well as a significant negative trend for self-committed violence. Thus, reactive aggression and self-committed violence appear to cancel each other out: Participants who committed more violent acts did not show better psychosocial functioning when they also showed more reactive aggression. Instead, with an increase in reactive aggression—as indicated by the main effect—psychosocial dysfunction also increases. On the other hand, there was no main effect for the factor appetitive aggression, a significant negative interaction between self- committed violence and appetitive aggression,

Table 2

Means, (Standard Deviations), and Correlations (R_p) for All Variables

Measure M(SD) 2 3 4 5 6 7

1. Traumatic event load 19.7 (4.1) .36^ⴱⴱ .41^ⴱⴱ .38^ⴱⴱ .27^ⴱ .31^ⴱ .35^ⴱ

2. PTSD symptom severity 22.0 (16.5) .63^ⴱⴱ .42^ⴱⴱ .28^ⴱ .23 .21

3. Psychosocial dysfunction 16.3 (8.3) .33^ⴱ .24^ⴱ .36^ⴱ .16

4. Concerns about future threat 18.7 (9.2) .07 .29^ⴱ .11

5. Appetitive aggression 30.7 (15.3) .63^ⴱⴱ .77^ⴱⴱ

6. Reactive aggression 13.7 (4.6) .56^ⴱⴱ

7. Self-committed violence 10.2 (4.2) Note. PTSD⫽posttraumatic stress disorder.

ⴱCorrelation signiﬁcant at the .05 level (two-tailed).^ⴱⴱCorrelation signiﬁcant at the .001 level (two-tailed).

(7)

as well as the negative trend for self-committed violence already described.

Thus, the more violent acts a participant com- mits under an appetitive perception of violence, the better their psychosocial functioning. More- over, an increase in the AAS score is not suffi- cient to reduce the SDQ symptom score. As indicated by the significant interaction between appetitive aggression and self-committed violence, only when the violence is acted out does the attraction to violence become noticeable and the participants experience less psychosocial problems. This effect accounts for at least 9% of the variance. A residual diagnostic performed to establish whether the proposed model fulfilled all formal criteria revealed that there were no statistical limitations (maximum variance infla- tion factor [VIF]⫽3.13, maximum Cook’sd⫽ .22, Kolmogorov⫺Smirnov test for residual differences from normal distributionZ⫽.56,p⫽ .912).

Relation Between Appetitive Aggression and Concerns About Future Threat

A further multiple linear regression investigated which variables contributed to the vari-

ance in the number of perceived future threats about which participants were concerned. As in the previous model, the score for concerns about future threat was regressed on the two trauma-associated variables—PTSD symptom severity and traumatic event load—as well as the three aggression-related variables—appetitive aggression, reactive aggression, and self- committed violence—and all possible two-way interactions. AIC was used for the selection of the best-ﬁtting model. Figure 2 shows the selected model, accounting for 43% of the variance in participants’ concerns about future threats,Radj2 ⫽.43,F(4, 65)⫽13.77,p⬍.001.

On the one hand, PTSD symptom severity as well as trauma event load both increased concerns about potential future threats. On the other hand, appetitive aggression reduced concerns about future threats. By contrast, reactive aggression was a marker for increased concerns about future threats. Neither self-committed violence nor any of the two-way interactions improved model ﬁt (ps⬎ .05). The effect size as well as the statistical power were sufﬁcient, f² ⫽ .75, (1 ⫺ ␤) ⫽ .99 (Faul et al., 2007).

Analyses of the residuals revealed that the re-

Figure 1. Linear regression model predicting psychosocial dysfunction as a function of trauma-related variables (PTSD symptom severity and trauma event load) and aggression- related variables (appetitive aggression, reactive aggression, self-committed violence) and both two-way interactions between the latter one with the two forms of aggressive behavior.

PTSD⫽posttraumatic stress disorder.

(8)

gression model fulﬁlled all necessary quality criteria (maximum VIF ⫽ 1.71, maximum Cook’sd⫽.15, Kolmogorov⫺Smirnov test for residual differences from normal distribution Z ⫽.41,p⫽ .996).

Discussion

The aim of the present study was to investigate whether appetitive aggression is associated with a reduced risk of psychosocial dysfunction and concerns about future threats when living under conditions of continuous exposure to violence. As one would expect, the number of experienced traumatic events and current PTSD symptomatology were signiﬁcant predictors of both psychosocial dysfunction and concerns about future threat. However, the results further revealed that those who reported violence to be fascinating and rewarding and who also acted out violent behavior showed less psychosocial dysfunction and reported fewer concerns about future threats. For both outcomes, appetitive aggression accounts for a signiﬁcant proportion of the variance. Interestingly, this inverse relation is limited to the interactions of appetitive aggression with trauma symptoms, but does not apply to other measures. In particular, the counterpart to appetitive aggression, reactive aggressive behavior, was associated with increased psychosocial dysfunction. Thus, it is not the perpetration of violence or aggressive behavior

in general that contributes to a better adaptation to a threatening environment, but the rewarding perception of violence. How can these results be explained?

The more different traumatic event types a person experiences, the greater the probability that daily cues will trigger elements of the traumatic events and cause a fear response. During a traumatic event, the sensory, emotional, and physiological cues an individual experiences are burned into memory. However, because most traumatic events share elements—as, for example, heart racing, blood, or fear—traumatic events are not stored in the brain as separate entities, but rather every traumatic event also triggers former traumatic events. Following principles of Hebbian learning that neurons that ﬁre together also wire together, experiences are stored in an interconnected neural network (El- bert, Rockstroh, Kolassa, Schauer, & Neuner, 2006), which may establish a “fear network”

(Lang, 1979, 1984). This fear or trauma network encompasses sensory, cognitive, physiological, and emotional experiences and includes the action disposition related to the experience.

It grows with every new stressful experience, thereby increasing the probability of developing PTSD (Schauer, Neuner, & Elbert, 2011). In PTSD, this sensory⫺perceptual representation has lost its usual association to the contextual system. The brain can no longer distinguish which representation belongs to which experi-

Figure 2. Linear regression model predicting distressing concerns as a function of trauma- related variables (PTSD symptom severity and trauma event load) and aggression-related variables (appetitive aggression, reactive aggression). PTS⫽posttraumatic stress disorder.

(9)

ence. Environmental stimuli (e.g., a smell or noise) and internal cues (e.g., a thought) can still activate the fear or trauma structure, but the trauma survivor has difﬁculties with autobio- graphical memory; that is, they are unable to orient the fear associated with the events appro- priately in time and space or to clearly chronologize these traumatic events. The information regarding “where” and “when” gets lost.

Sometimes even minor and implicit daily cues can thus arouse severe trauma symptoms. The permanent fear response arouses ongoing stress and impairs psychosocial adaptation, causing further behavioral problems.

This was also observed in the present study:

PTSD symptom severity was related to both psychosocial dysfunctioning and concerns about future threat. However, we suggest that the avoidance producing (negative valent) “fear network”

competes with a “hunting network” with positive valence, that is, approach rather than avoidance for the integration of violence cues (Elbert et al., 2010). According to the appetitive perception of violence, violence cues that are processed from a perpetrator’s perspective during the commission of atrocities are associated with an appetitive instead of an aversive arousal. Consequently, cues like blood or heart racing do not become associated with a fear response, but rather develop a desirable association. These cues no longer en- large the fear network, but are integrated into an appetitive network. Moreover, cues that formerly belonged to the fear network now become appetitive. This leads to a reduction of the frequency of fear responses and thus may improve daily functioning.

Furthermore, we have found in other populations that appetitive aggression also increases status in the ingroup (Crombach, Weierstall, Hecker, Schalinski, & Elbert, in press). One might specu- late that psychosocial adaptation depends on one’s social status in the ingroup: Those who behave more violently experience fewer threats within their own group and therefore display less psychosocial dysfunction. This element seems to be cru- cial in the South African context as well. Gang- related urban violence in Cape Town has a long history in South Africa and goes back to the early 20th century. It has been estimated that in this area, up to 100,000 inhabitants, most of them males living in townships, are organized in street gangs (Kinnes, 1995). Gang membership is often associated with a gain in security and a sense of

community, which are often lacking in daily life, but is also associated with a rise in social status (Covey, 2003).Sathiparsad, Taylor, and Dlamini (2008) stated that dominance, assertiveness, and aggression are key features of masculinity in the highly patriarchal society of South Africa. How- ever, societal changes in the past years in South Africa have challenged masculine identities, lead- ing to “instability and uncertainty over social role and identity, sexuality, work and personal rela- tionships”—a situation that has been called a crisis in masculinity (Frosh, Phoenix, & Pattman, 2002, p. 85;Walker, 2005).

In economically disadvantaged contexts such as South Africa’s townships, the favelas of Bra- zil, or the inner cities of the United States, where power and social status are difﬁcult to achieve economically, engagement in street gangs and the related creation of a violent norm of masculinity may help to create new homes and families and to retain traditional masculine roles (Hagedorn, 2008;Kynoch, 1999). In line with these theoretical assumptions, our results support the view that, in an adverse environment, dominated by violence and threat, fearful processing of the enduring threats and of past exposure to violence compromises psychosocial functioning, while violent enactments and an appetitive perception of violence may be an effective coping strategy. We suggest that appetitive aggression not only alters the perception of violence but also might have an impact on status and acceptance in the social environment. The possible relation between appetitive aggression, norms of masculinity, and social status in those who perpetrate violence should be investigated in future studies.

Of course, in the long term, this coping strategy may result in further risk of harm to self and others, incarceration, and cycling of violence. This asks for the development of more appropriate coping strategies and resiliencies among young men who live in high-violence communities, while rec- ognizing their need to maintain physical safety, psychological status, and meaningful masculine roles within their communities.

To address appetitive aggression in populations with histories of violence perpetration, we have recently developed FORNET (Narrative Exposure Therapy for Forensic Offender Reha- bilitation;Elbert, Hermnau, Hecker, Weierstall,

& Schauer, 2012), an adaptation of Narrative Exposure Therapy (Schauer et al., 2011), a

(10)

treatment that has proven its effectiveness for the treatment of trauma victims in randomized controlled trials (Robjan & Fazel, 2010). In FORNET, a chronological narrative of the participant’s life is constructed, focusing on the detailed exploration of the traumatic experiences and the committed offenses. A signiﬁcant reduction in the perpetration of violence after participation in FORNET has been demonstrated (Hermenau, Hecker, Schaal, Maedl, &

Elbert, in press). However, even though aggressive behavior can effectively be reduced on an individual level, it remains highly functional in socioeconomically marginalized communities as a means to gain resources, status, and safety.

Effective programs of socioeconomic empow- erment are needed at a community level as well.

As long as the organization in street gangs offers more powerful incentives than abstinence from violent behavior, the long-lasting transfer of therapeutic gains from institutional settings into the real world of the community remains a challenge for those youth offenders who try to refrain from violence.

Study Limitations

This study focused on a group of former young offenders in a South African context, and external validity is therefore limited to this sub- group. Another limitation is that the age range of participants was heterogeneous. However, there was no inﬂuence of participants’ age in any of the analyses. A future research focus on different age groups, following their development over a longer time course, would provide fruitful insights into the underlying mechanisms of aggression and traumatization in youth and help to identify sensitive periods. It is further possible that participants in this study may have attempted to present themselves to the researchers in a way that reﬂected masculine gender norms that value aggression and denigrate fear or anxiety responses. Moreover, motivation to successfully complete the program in which they were enrolled may have limited the truth- fulness with which participants answered the questions. However, even these limitations do not fully account for the rather complex relations between the assessed constructs that are in line with results from other studies. In particular, the differential effects of reactive and appetitive aggression on our main outcome vari-

ables appear to lend support to the validity of our ﬁndings.

References

Akaike, H. (1987). Factor analysis and AIC. Psy- chometrika, 52, 317–332. doi:10.1007/

BF02294359

Altbeker, A. (2007). A country at war with itself:

South Africa’s crisis of crime. Jonathan Ball Pub, Johannesburg, SA.

Amaya-Jackson, L. (1998). Child’s Exposure to Vi- olence Checklist. Durham, NC: Trauma Evalua- tion, Treatment and Research Program, Centre for Child and Family Health.

Breslau, N. (1998). Epidemiology of trauma and posttraumatic stress disorder. In R. Yehuda (Ed.), Review of psychiatry(Vol. 17, pp. 1–30). Wash- ington, DC: American Psychiatric Association.

Burton, P. (2006). Easy prey: Results of the national youth victimisation study.SA Crime Quarterly, 16, 1– 6.

Byrne, C. A., & Riggs, D. S. (1996). The cycle of trauma: Relationship aggression in male Vietnam veterans with symptoms of posttraumatic stress disorder.Violence and Victims, 11(3), 213–225.

Catani, C., Jacob, N., Schauer, E., Kohila, M., &

Neuner, F. (2008). Family violence, war, and nat- ural disasters: A study of the effect of extreme stress on children’s mental health in Sri Lanka.

BMC Psychiatry, 8, 33, doi:10.1186/1471-244X- 8-33

Catani, C., Schauer, E., Elbert, T., Missmahl, I., Bette, J.-P., & Neuner, F. (2009). War trauma, child labor, and family violence: Life adversities and PTSD in a sample of school children in Kabul.

Journal of Traumatic Stress, 22, 163–171. doi:

10.1002/jts.20415

Christopher, A. J. (2005). The slow pace of desegre- gation in South African cities, 1996 –2001.Urban Studies, 42, 2305–2320. doi:10.1080/

00420980500332122

Covey, H. C. (2003). Street gangs throughout the world. Springﬁeld, IL: Charles C. Thomas.

Crombach, C., Weierstall, R., Hecker, T., Schalinski, I., & Elbert, T. (in press). Social status and appetitive violence experiences: A study with former child soldiers of Uganda.Journal of Aggression, Maltreatment & Trauma.

Demombynes, G., & Özler, B. (2005). Crime and local inequality in South Africa.Journal of Devel- opment Economics, 76, 265–292. doi:10.1016/j .jdeveco.2003.12.015

Dinan, B. A., McCall, G. J., & Gibson, D. (2004).

Community violence and PTSD in selected South African townships. Journal of Interpersonal Violence, 19, 727–742. doi:10.1177/

0886260504263869

(11)

Elbert, T., Hermenau, K., Hecker, T., Weierstall, R.,

& Schauer, M. (2012). FORNET: Behandlung von traumatisierten und nicht-traumatisierten Gewalt- tätern mittels Narrativer Expositionstherapie. In J.

Endrass, A. Rossegger, F. Urbaniok, & B. Bor- chard (Eds.),Interventionen bei Gewalt- und Sexu- alstraftätern: Risk-Management, Methoden und Konzepte der forensischen Therapie (pp. 225–

276). Berlin, Germany: Medizinisch Wissen- schaftliche Verlagsgesellschaft.

Elbert, T., Rockstroh, B., Kolassa, I. T., Schauer, M.,

& Neuner, F. (2006). The inﬂuence of organized violence and terror on brain and mind: A co- constructive perspective. In P. Baltes, P. Reuter- Lorenz, & F. Rösler (Eds.),Lifespan development and the brain: The perspective of biocultural co- constructivism (pp. 326 –349). Cambridge, Eng- land: Cambridge University Press. doi:10.1017/

CBO9780511499722.017

Elbert, T., Weierstall, R., & Schauer, M. (2010).

Fascination violence: On mind and brain of man hunters. European Archives of Psychiatry and Clinical Neuroscience, 260(Supp. 2), S100 –S105.

doi:10.1007/s00406-010-0144-8

Ellis, B. H., Lhewa, D., Charney, M., & Cabral, H.

(2006). Screening for PTSD among Somali adolescent refugees: Psychometric properties of the UCLA PTSD Index.Journal of Traumatic Stress, 19,547–551.doi:10.1002/jts.20139

Faul, F., Erdfelder, E., Lang, A. G., & Buchner, A.

(2007). G^ⴱPower 3: A ﬂexible statistical power analysis for the social, behavioral, and biomedical sciences. Behavior Research Methods, 39, 175–

191.doi:10.3758/BF03193146

Fehon, D. C., Grilo, C. M., & Lipschitz, D. S. (2001).

Correlates of community violence exposure in hos- pitalized adolescents.Comprehensive Psychiatry, 42,283–290.doi:10.1053/comp.2001.24580 Fincham, D.Grimsrud, A., Corrigall, J., Williams,

D. R., Seedat, S., Stein, D. J., & Myer, L. (2009).

Intermittent explosive disorder in South Africa:

Prevalence, correlates and the role of traumatic exposures. Psychopathology, 42, 92–98. doi:

10.1159/000203341

Foster, D. (2012). Gender, class, ‘race’ and violence.

In: C. Ward, A. van der Merwe, & A. Dawes (Eds.), Youth violence: Sources and solutions in South Africa(pp. 23–51). Cape Town, South Af- rica: UCT Press.

Frosh, S., Phoenix, A., Pattman, R. (2002). Young masculinities: Understanding boys in contempo- rary society. Basingstoke, Basingstoke, UK: Pal- grave.

Goodman, R. (2001). Psychometric properties of the strengths and difﬁculties questionnaire.Journal of the American Academy of Child & Adolescent Psy- chiatry, 40, 1337–1345. doi:10.1097/00004583- 200111000-00015

Goodman, R., & Scott, S. (1999). Comparing the strengths and difﬁculties questionnaire and the child behavior checklist: Is small beautiful?Jour- nal of Abnormal Child Psychology, 27, 17–24.

doi:10.1023/A:1022658222914

Hagedorn, J. M. (2008). A world of gangs: Armed young men and gangsta culture. Minneapolis, MN:

University of Minnesota Press.

Hermenau, K., Hecker, T., Ruf, M., Schauer, E., Elbert, T., & Schauer, M. (2011). Childhood ad- versity, mental ill-health and aggressive behavior in an African orphanage: Changes in response to trauma-focused therapy and the implementation of a new instructional system.Child and Adolescent Psychiatry and Mental Health, 5,1–9.

Hermenau, K., Hecker, T., Schaal, S., Maedl, A., &

Elbert, T. (in press). Addressing post-traumatic stress and aggression by means of narrative exposure–A randomized controlled trial with ex- combatants in the eastern DRC.Journal for Mal- treatment, Aggression and Trauma.

Hunt, K. L., Martens, P. M., & Belcher, M. E. (2011).

Risky business: Trauma exposure and rate of posttraumatic stress disorder in African American children and adolescents.Journal of Traumatic Stress, 24,365–369.doi:10.1002/jts.20648

Jakupcak, M., Conybeare, D., Phelps, L., Hunt, S., Holmes, H. A., Felker, B., . . . McFall, M. E.

(2007). Anger, hostility, and aggression among Iraq and Afghanistan war veterans reporting PTSD and subthreshold PTSD. Journal of Traumatic Stress, 20,945–954.doi:10.1002/jts.20258 Jones, D. (2008) Human behaviour: Killer instincts.

Nature, 451,512–515.doi:10.1038/451512a Kaminer, D., Grimsrud, A., Myer, L., Stein, D. J., &

Williams, D. R. (2008). Risk for post-traumatic stress disorder associated with different forms of interpersonal violence in South Africa.Social Sci- ence & Medicine, 67, 1589 –1595. doi:10.1016/j .socscimed.2008.07.023

Kennedy, A. C. (2008). An ecological approach to examining cumulative violence exposure among urban, African American adolescents.Child & Ad- olescent Social Work Journal, 25, 25– 41. doi:

10.1007/s10560-007-0110-0

Kinnes, I. (1995). Reclaiming the Cape Flats: A community challenge to crime and gangsterism.

Crime and Conﬂict, 2,5– 8.

Kolassa, I. T., Ertl, V., Eckart, C., Kolassa, S., Onyut, L. P., & Elbert, T. (2010). Spontaneous remission from PTSD depends on the number of traumatic event types experienced.Psychological Trauma:

Theory, Research, Practice, and Policy, 2, 169 – 174.doi:10.1037/a0019362

Kynoch, G. (1999). From the Ninevites to the Hard Livings gang: Township gangsters and urban violence in twentieth-century South Africa.

(12)

African Studies, 58, 55– 85. doi:10.1080/

00020189908707905

Lang, P. J. (1979). A bio-informational theory of emotional imagery. Psychophysiology, 16, 495–

512.doi:10.1111/j.1469-8986.1979.tb01511.x Lang, P. J. (1984). Dead souls: Or why the neurobe-

havioral science of emotion should pay attention to cognitive science. In E. B. Rockstroh, W. Lutzen- berger, & N. Birbaumer (Eds.),Self-regulation of the brain and behavior(pp. 255–272). New York, NY: Springer.doi:10.1007/978-3-642-69379-3_18 Layne, C. M., Stuvland, R., Saltzman, W., Djapo, N.,

& Pynoos, R. S. (1999). War trauma screening scale. Unpublished manuscript.

Mullick, M. S. L., & Goodman, R. (2001). Question- naire screening for mental health problems in Ban- gladeshi children: A preliminary study.Social Psy- chiatry and Psychiatric Epidemiology, 36,94 –99.

doi:10.1007/s001270050295

Nell, V. (2006). Cruelty’s rewards: The gratiﬁcation of perpetrators and spectators. Behavioral and Brain Sciences, 29,211–224; discussion 224-57.

doi:10.1017/S0140525X06009058

Neuner, F., Schauer, M., Karunakara, U., Klaschik, C., Robert, C., & Elbert, T. (2004). Psychological trauma and evidence for enhanced vulnerability for posttraumatic stress disorder through previous trauma among West Nile refugees.BMC Psychia- try,4 34.doi:10.1186/1471-244X-4-34

Raine, A., Dodge, K., Loeber, R., Gatzke-Kopp, L., Lynam, D., Reynolds, C., . . . Liu, J. (2006). The reactive-proactive aggression questionnaire: Dif- ferential correlates of reactive and proactive aggression in adolescent boys.Aggressive Behavior, 32,159 –171.doi:10.1002/ab.20115

Robjant, K., & Fazel, M. (2010). The emerging evidence for Narrative Exposure Therapy: A review.

Clinical Psychology Review, 30,1030 –1039.doi:

10.1016/j.cpr.2010.07.004

Rodriguez, N., Steinberg, A. M., & Pynoos, R. S.

(1999).UCLA PTSD Index for DSM-IV. Los An- geles, CA: Trauma Psychiatry Service, University of California at Los Angeles.

Rodriguez, N., Steinberg, A. M., Satzman, W., &

Pynoos, R. S. (2001, December). PTSD Index:

Psychometric analyses of the adolescent version.

Paper presented at the International Society for Traumatic Stress Studies, New Orleans, LA.

Sathiparsad, R., Taylor, M., & Dlamini, S. (2008).

Patriarchy and family life: Alternative views of male youth in rural South Africa. Agenda, 22, 4 –16.

Schauer, M., Neuner, F., & Elbert, T. (2011). Nar- rative Exposure Therapy (2nd ed.). Göttingen, Germany: Hogrefe.

Schauer, M., Neuner, F., Karunakara, U., Klaschik, C., Robert, C., & Elbert, T. (2003). PTSD and the

“building block” effect of psychological trauma among West Nile Africans.ESTSS (European So- ciety for Traumatic Stress Studies) Bulletin, 10, 5– 6.

Seedat, M., Van Niekerk, A., Jewkes, R., Sufﬂa, S.,

& Ratele, K. (2009). Violence and injuries in South Africa: Prioritising an agenda for preven- tion.The Lancet, 374,1011–1022.

Steel, Z., Chey, T., Silove, D., Marnane, C., Bryant, R. A., & van Ommeren, M. (2009). Association of torture and other potentially traumatic events with mental health outcomes among populations exposed to mass conﬂict and displacement: A sys- tematic review and meta-analysis.JAMA: Journal of the American Medical Association, 302,537–

549.doi:10.1001/jama.2009.1132

Turok, I., & Watson, V. (2001). Divergent development in South African cities: Strategic challenges facing Cape Town. InUrban Forum(Vol. 12, No.

2, pp. 119 –138). Netherlands: Springer.

Walker, L. (2005). Men behaving differently: South African men since 1994.Culture, Health & Sexu- ality, 7,225–238.

Weierstall, R., & Elbert, T. (2011). The Appetitive Aggression Scale⫺Development of an instrument for the assessment of human’s attraction to violence.European Journal of Psychotraumatology, 2,8430.doi:10.3402/ejpt.v2i0.8430

Weierstall, R., & Elbert, T. (2012). Formen und Klassiﬁkation menschlicher Aggression. In J. En- drass, A. Rossegger, F. Urbaniok, & B. Borchard (Eds.), Interventionen bei Gewalt- und Sexual- straftätern: Risk-Management, Methoden und Konzepte der forensischen Therapie (pp. 3–14).

Berlin, Germany: Medizinisch Wissenschaftliche Verlagsgesellschaft.

Weierstall, R., Schaal, S., Schalinski, I., Dusingize- mungu, J. P., & Elbert, T. (2011). The thrill of being violent as an antidote to posttraumatic stress disorder in Rwandese genocide perpetrators. Eu- ropean Journal of Psychotraumatology, 2,6345.

doi:10.3402/ejpt.v2i0.6345

Weierstall, R., Schalinski, I., Crombach, A., Hecker, T., & Elbert, T. (2012). When combat prevents PTSD symptoms—Results from a survey with former child soldiers in Northern Uganda.BMC Psy- chiatry, 12,41.doi:10.1186/1471-244X-12-41