ENRIQUE CHAUX
JULIANA ARBOLEDA
CLAUDIA RINCÓN
Universidad de los Andes, Bogotá, Colombia

Correspondence concerning this article should be addressed to Enrique Chaux, e-mail: echaux@uniandes.edu.co.
Department of Psychology, Universidad de los Andes.

SCIENTIFIC RESEARCH ARTICLE
RECEIVED: 4 APRIL 2012 - ACCEPTED: 19 OCTOBER 2012

Abstract

Children exposed to higher levels of violence tend to be more aggressive. Specific mechanisms explaining this relationship are still being uncovered. This study sought to identify the relationship between exposure to community violence and reactive and proactive aggression, as well as cognitive and emotional variables mediating this relationship. Participants were 1,235 students (from fifth to ninth grade) from localities of Bogotá, Colombia, with varying levels of community violence. Analyses of self-reported measures confirmed significant associations between exposure to community violence and both reactive and proactive aggression. Normative beliefs supporting aggression, hostile attribution of intent, positive expectations for aggression, and lack of guilt after aggression, partially mediated these relationships, suggesting strategies for prevention.

Keywords: reactive aggression, proactive aggression, community violence, political violence, guilt, hostile attribution of intent, normative beliefs, Colombia.

Resumen

Los niños y niñas expuestos a elevados niveles de violencia tienden a ser más agresivos que los demás, pero los mecanismos específicos que explican esta relación todavía se están descubriendo. Este estudio buscó identificar la relación entre la exposición a la violencia comunitaria y la agresión reactiva y proactiva, así como las variables cognitivas y emocionales que median dicha relación. Con la participación de 1,235 estudiantes (de quinto a noveno grado) de diferentes localidades de Bogotá, Colombia, y de diversos niveles de violencia comunitaria, los análisis de las medidas de autoreporte confirmaron la existencia de asociaciones significativas entre la exposición a la violencia comunitaria y la agresión tanto reactiva como proactiva. Las creencias que legitiman la agresión, la atribución hostil de intenciones, las expectativas positivas sobre la agresión, y la falta de culpa al agredir, mediaron parcialmente estas relaciones, sugiriendo estrategias de prevención.

Palabras clave: agresión reactiva, agresión proactiva, violencia comunitaria, violencia política, culpa, atribución hostil de la intención, creencias normativas, Colombia. 

Resumo

As crianças expostas a elevados níveis de violência tendem a ser mais agressivas que as demais; no entanto, os mecanismos específicos que explicam esta relação ainda estão se descobrindo. Este estudo pretendeu identificar a relação entre a exposição à violência comunitária e a agressão reativa e proativa, assim como as variáveis cognitivas e emocionais que mediam tal relação.  Com a participação de 1.235 estudantes (de 8 a 18 anos) de diferentes localidades de Bogotá (Colômbia) e de diversos níveis de violência comunitária, as análises das medidas de autorrelatório confirmaram a existência de associações significativas entre a exposição à violência comunitária e a agressão tanto reativa quanto proativa. As crenças que legitimam a agressão, a atribuição hostil de intenções, as expectativas positivas sobre a agressão e a falta de culpa ao agredir, mediaram parcialmente estas relações e sugeriram estratégias de prevenção.

Palavras-chave: agressão reativa, agressão proativa, violência comunitária, violência política, culpa, atribuição hostil da intenção, crenças normativas, Colômbia.

GROWING UP in a violent environment has many negative long-lasting effects. One of these is a higher risk for the development of aggression. Several studies have confirmed this effect in the context of the family (e.g., Dodge, Bates, & Pettit, 1990; Huesmann, Eron, Lefkowitz, & Walder, 1984; Thornberry, Freeman-Gallant, Lizotte, Krohn, & Smith, 2003; Widom, 1989). Other studies have found that exposure to community violence also has an effect on the development of aggression (e.g., Allwood & Bell, 2008; Bradshaw, Rodgers, Ghandour, & Garbarino, 2009; Brookmeyer, Henrich, & Schwab-Stone, 2005; Calvete & Orue, 2011; Flannery, Wester, & Singer, 2004; Gorman-Smith, Henry, & Tolan, 2004; Guerra, Huesmann, & Spindler, 2003; Liddell, Kvalsvig, Qotyana, & Shabalala, 1994; Miller, Wasserman, Neugebauer, Gorman-Smith, & Kamboukos, 1999; Musher-Eizenman et al., 2004; Orue et al., 2011; Schwab-Stone et al., 1999; Schwartz & Proctor, 2000). These studies have confirmed that exposure to violent events in the community, either as a witness or as a victim, increase the chances of developing aggressive behaviors. Furthermore, some of these studies have identified cognitive and emotional mechanisms that may be mediating or moderating the effect of exposure to community violence on aggressive behavior (e.g., Allwood & Bell, 2008; Bradshaw et al., 2009; Calvete & Orue, 2011; Guerra et al., 2003; Musher-Eizenman et al., 2004; Orue et al., 2011; Schwartz & Proctor, 2000). In this way, the dynamics that lead to the development of aggression can be better understood and ways of minimizing the negative effects of community violence could be developed. The current study seeks to make a contribution to these theoretical and practical goals by identifying cognitive and emotional variables mediating the relationship between community violence and aggression, in urban contexts where violence may be associated with political violence or common crime.

In a study conducted in high crime neighborhoods of Chicago, Guerra et al. (2003) found that exposure to community violence, such as witnessing someone being beaten or having to hide because of shootings in their neighborhood, predicted increases in aggressive behavior one year later (interestingly, the contrary was not true, that is, aggressive behavior did not predict exposure to community violence). Furthermore, this effect of exposure to community violence on aggressive behavior was in part mediated by two kinds of social cognitions: normative beliefs supporting aggression and aggressive fantasies. That is, being exposed to violent events in the community leads to an increase in normative beliefs that support aggression and in aggressive fantasies which, in turn, lead to an increase in aggressive behaviors. At least three other studies have also found that normative beliefs about aggression mediate this relationship (Allwood & Bell, 2008; Bradshaw et al., 2009; Orue et al., 2011).

Musher-Eizenman et al. (2004) also found that social cognitions play a mediating role between exposure to violence and children's aggressive behavior. In particular, aggressive fantasies, beliefs legitimizing retaliation, and positive self-evaluations after behaving aggressively mediated the effect of exposure to violence (as witness) on aggression. In addition, they found that while social cognitions related to direct aggression (e.g., fantasizing about hitting someone) mediated the effect of exposure to violence on direct aggression, social cognitions related to indirect aggression (e.g., fantasizing about spreading rumors about someone) mediated the effect of exposure to violence on indirect aggression. This suggests that the pathways by which exposure to violence affects aggression might be different for different types of aggression. The study, however, concentrated on the form of aggression (direct vs. indirect), but not on the function of aggression (proactive vs. reactive) (Little, Jones, Henrich, &  Hawley, 2003).

Schwartz and Proctor (2000) conducted a similar study in high crime neighborhoods in Los Angeles. They found that while both witnessing and being victimized by community violence are associated with aggression, only being victimized was associated with being rejected and bullied by peers. Furthermore, the mediating psychological variables seem to be different, too. On the one hand, problems of emotional dysregulation were found to mediate the relation between being a victim of violence in the community and being rejected and bullied. On the other hand, cognitive biases such as positive expectations about the outcomes of aggression and efficacy beliefs for aggression (i.e., considering that it is rather easy for them to behave aggressively) mediated the relation between witnessing violence in the community and behaving aggressively.

The results of Schwartz and Proctor's (2000) study suggest that different types of experiences with community violence might lead to the development of different types of aggression. Exposure as a witness was found to be related to positive expectations about the outcome of aggression, a social cognitive bias which other studies (e.g., Crick & Dodge, 1996) found to be associated with proactive aggression (i.e., aggression used as an instrument to obtain a goal without any previous provocation). In contrast, exposure as a victim was found to be related to problems of emotion regulation, which according to Dodge (1991) are associated with reactive aggression (i.e., aggression used as a reaction against a real or perceived provocation). However, Schwartz and Proctor (2000) did not explicitly differentiate these two types of aggression in their study.

In a three wave longitudinal study conducted in the Basque Country, Spain, Calvete and Orue (2011) did differentiate reactive and proactive aggression. They found that hostile attribution of intent (i.e., interpreting others' intentions hostilely even when those intentions are not evident; Dodge et al., 1990; Orobio de Castro, Veerman, Koops, Bosch, & Monshouwer, 2012), anger reactions, and selection of aggressive responses mediate the relationship between exposure to violence and reactive aggression. However, they did not find this mediation for proactive aggression, suggesting that different mechanisms might be associated with the effect that exposure to violence has on reactive versus proactive aggression.

This difference seems crucial since several studies have suggested that reactive and proactive aggression are associated with different peer processes (e.g., Dodge, Lochman, Harnish, Bates, & Pettit, 1997; Poulin & Boivin, 2000), different cognitive and emotional processes (Crick & Dodge, 1996; Dodge et al., 1997; Orobio de Castro et al., 2005) and seem to follow different developmental trajectories (Brendgen et al., 2001; Dodge et al., 1997; Vitaro & Brendgen, 2005; Vitaro, Gendreau, Tremblay, & Oligny, 1998). It may also be that the effect of exposure to community violence on the development of reactive aggression could be mediated by variables different from those affecting proactive aggression. Thus, one of the goals of the current study was to investigate the relation between exposure to community violence and both types of aggression and to uncover possible mediating cognitive and emotional variables.

As mentioned above, several studies have found that normative beliefs supporting aggression mediated the relationship between exposure to community violence and aggressive behavior. However, just as Musher-Eizenman et al. (2004) found differences in the beliefs mediating the effect on direct versus indirect aggression, beliefs mediating the effect of exposure to violence on reactive aggression may be different from those mediating the effect on proactive aggression. For example, considering that "it is OK to hurt someone who hurt you first" may support the use of reactive aggression while considering that "sometimes you have to use threats to get what you want" may support the use of proactive aggression. A second goal of the current study was, therefore, to investigate the possible mediating roles of reactive versus proactive normative beliefs.

If violence is common in their social environment, children may learn that others are likely to use violence in their interactions and could therefore develop hostile attribution of intent. Furthermore, by having many opportunities to witness the use of violence among young and adult role models around them, they may learn to expect positive outcomes of using aggression. Both of these cognitive biases are likely to mediate the effect of exposure to community violence on the development of aggression, although differently for both types of aggression, since hostile attribution of intent has been related to reactive aggression while positive expectations about the outcome of aggression has been related to proactive aggression (Crick & Dodge, 1996). The mediating role of these two cognitive biases was also investigated in the current study.

Finally, by observing interactions around them and by interacting with others, children learn the moral standards of their environments (Bandura, 1991). If violence is common in their communities, they are likely to learn that it is not against moral standards to use violence as a way to defend themselves or to obtain goals. Thus, they are not likely to feel guilt when behaving aggressively, since guilt is associated with acting (or considering to act) against internalized moral standards (Hoffman, 1998; Kochanska, Gross, Lin, & Nichols, 2002). Since feelings of guilt serve as internal constraints against particular behaviors, not feeling guilt when having aggressive thoughts or behaviors may increase the chances of actually behaving aggressively. In this way, lack of guilt after aggression may be another mediating variable in the relation between exposure to community violence and development of aggression. This relation may be particularly evident for proactive aggression, since this type of aggression has been related to a certain coldness and lack of empathy towards the victims of aggression (Arsenio & Lemerise, 2001; Dodge, 1991; Endresen & Olweus, 2001; Parra, 2005). Thus, another goal of this study was to consider the possible mediating role of lack of guilt after aggression. 

Most studies of the impact of community violence on children's aggression have been conducted in the context of inner city violence where crime rates related to illegal drug activities and gang violence are high. Few studies, however, have been conducted in urban contexts where political violence is also present (Liddell et al., 1994, and Punamäki, Muhammed, & Abdulrahman, 2004, are some of the exceptions). Some large cities in Colombia, such as the capital, Bogotá, have neighborhoods where community violence is related to political violence (caused by the five-decade-old, low-intensity internal conflict between left-wing guerrillas, right-wing paramilitaries, and the Colombian army) and crime-based violence intermingle, while in other neighborhoods, community violence is only related to common crime. This allows us to make comparisons between the effects of exposure to different kinds of community violence. The final goal of this study was thus to consider the possible relations between aggression and exposure to different types of community violence.

To summarize, the current study sought to answer the following research questions and, based on previous studies, tested the following hypotheses:

1. What is the relation between exposure to community violence and reactive and proactive aggression? We were expecting to find significant relations between exposure to community violence and both types of aggression.

2. What cognitive and emotional variables may be mediating that relation? We expected that beliefs supporting reactive aggression and hostile attribution of intent would mediate the relation between exposure to community violence and reactive aggression. In contrast, we expected that beliefs supporting proactive aggression, lack of guilt and positive expectations for aggression would mediate the relation between exposure to community violence and proactive aggression.

3. How does aggression and its related cognitive and emotional variables compare between children and adolescents who are exposed to crime-based community violence and those who, additionally, live in communities where there is presence of political violence? Because of the dearth of studies comparing the effect of exposure to different types of community violence, we did not have clear hypotheses related to this question.

Methodology

Participants

Participants in this study were 1,235 children and adolescents (49.7% female) from seven public schools in Bogotá, Colombia. Their ages ranged from 8 to 18 years, although 87% were between 10 and 15 years old. Two fifth, seventh, and ninth grade classes were randomly selected from each of the schools (except for one school which was much larger than the others and where four classes from each of the three grades were randomly selected).

Schools were selected seeking diversity in levels and types of community violence (crime-based or political) in their neighborhoods. These schools are located in four of the 20 localities (localidades) of this city with a population of 6.8 million. Two schools are located in Santafé, a downtown area with the city's highest crime and homicide rates, but with very little presence of urban militias from guerrilla and paramilitary groups. Socio-economic conditions in these neighborhoods are low, but not the lowest in Bogotá (see Table 1).

Two other schools are from Ciudad Bolívar, which is a peripheral area with very low socio-economic conditions. Ciudad Bolívar is also the area of the city with the greatest incidence of political violence at the time of data collection. Although direct combats are very rare, urban militias from guerrilla and paramilitary groups compete for territorial control, extortion of commercial activities and recruiting of members, especially among the locality's youth (Alcaldía Mayor de Bogotá, 2005). It is also the area receiving most internally displaced families who arrive in Bogotá from rural areas because they received threats or were direct victims of Colombia's internal armed conflict  (see Table 1). Youth gangs are also more prevalent in Ciudad Bolívar than in any other area of the city (Alcaldía Mayor de Bogotá, 2005).

Two schools are located in Usme, which is another peripheral and very poor area of the city, but with moderate crime and homicide rates. Armed groups are also present in Usme, but not to the extent of Ciudad Bolívar. Finally, one school is located in Tunjuelito, which has relatively low homicide rates and low presence of political violence. Contrary to all the other schools, this one draws its students from many areas of the city and not only from nearby neighborhoods. For this reason, students from this school were not considered in the ANOVAs comparing students from different localities of the city.

Instruments

Measures were created for this study or translated into Spanish and adapted from existing ones. In order to verify whether students understood the questions the way we intended and responded to them in consistent ways, three phases of pilot tests and focus groups were conducted with 421 students in total from schools in similar neighborhoods in Bogotá. Several changes in the language and in the kind of questions were introduced to each of the instruments following these pilot tests and focus groups¹.

Reactive (α=.560) and proactive aggression (α=.747). Reactive aggression was evaluated with the following questions (translated from Spanish): "When they treat you badly, do you retaliate immediately?" and "When they treat you badly, do you wait a while before retaliating?". Proactive aggression was measured with the following questions: "Do you threaten others to get what you want?", "Do you bully and make others feel bad?" and "Do you enjoy treating others badly?" Possible answers were never (coded as 0), almost never (1), almost always (2), and always (3). Correlation between reactive aggression items was .390 and significant at p<.001. A composite of all five items was used as a general measure of aggression (α=.759).

Exposure to community violence (α=.573). Exposure to community violence was measured with the following three questions: "Within the last month, how many times have you...": "seen or heard gun shots in your neighborhood?", "seen fights in the street?", and "heard about or seen someone being robbed with violence in your neighborhood?" Possible answers were never (coded as 0), once (1), two to four times (2), and five or more times (3).

Exposure to gangs (α=.672). Exposure to youth gangs was measured with the following three questions: "Are there any gangs in your neighborhood?", "Do you know anyone who belongs or has belonged to a gang?", and "Do you have friends who belong to gangs?" Possible answers were no (coded as 0), yes (1), or in the case of the first question don't know (0).

Parental supervision (α=.732). Parental supervision was measured with the following two questions: "Do they know at home where you spend your free time?" and "Do they know at home with whom you spend your free time?". Possible answers were never (coded as 0), almost never (1), almost always (2), and always (3).

Hostile attribution of intent (α=.415). Hostile attribution of intent was measured by asking participants to imagine themselves in two different situations in which someone ambiguously hurt them. For example: "Imagine that in the next class break you are talking to your best friend and suddenly a student passes by and bumps into you. You fall down and hurt yourself". Each of the stories is followed by questions asking for the intention of the others. In this example, the question was: "Why did he/she bump into you?" The possible answers are: he/she wanted to hurt you (coded as 1), it was an accident (0), and you don't know (0). One of the stories was adapted from Milich and Dodge (1984), and the other was drawn from previous research about actual interpersonal conflicts among Colombian children and adolescents (Chaux, 2001). Alpha's Cronbach was particularly low in part because the measure included only two (dichotomous) items. Correlation between these items was .263 and significant at p<.001.

Lack of guilt (α=.641) and positive expectations about the outcome of aggression (α=.713). Four different kinds of expectations about the outcome of aggression were evaluated using short hypothetical stories: (a) that the outcome will be beneficial for you; (b) that you will be respected; (c) that you won't feel guilt after behaving aggressively; and (d) that aggression won't get you into trouble. Each of the stories asks participants to imagine themselves in a situation where they would behave aggressively. For example, "Imagine that someone in your class has been bothering you and then you hit that person". Then four questions were asked about each case. For example: (a) "Will the student stop bothering you?" (benefits), (b) "Others will respect you?" (respect), (c) "Would you feel fine about what you did afterwards? (guilt)", and (d) "Would you get into trouble?" (problems). Possible answers were yes or no (benefits, respect, lack of guilt and lack of problems were coded as 1 when answering yes). This measure was adapted from Zelli, Dodge, Lochman, Laird, & Conduct Problems Prevention Research Group (1999).

Normative beliefs supporting reactive (α=.624) and proactive aggression (α= .661 for reaching goals; α=.619 for protecting image). Normative beliefs supporting aggression were measured with a list of 33 statements to which participants responded true or false. Some of these statements were adapted from Slaby and Guerra (1988), and Stuckless and Goranson (1992), while others were created for this study. A series of factor analyses showed that a subset of these items could be categorized into beliefs legitimizing reactive and proactive aggression. Specifically, an initial factor analysis showed a six factor structure, three of which were not directly related to reactive or proactive aggression ("victims deserve the aggression they get", "aggression is OK to defend friends", and "aggression is OK if everyone uses it"). Items that loaded equally high on two or more factors were eliminated (e.g., "if someone hurts you, you have to respond in the same way", loaded .37 on Factor 1 and .32 on Factor 2). The remaining 15 items grouped into two factors of beliefs legitimizing proactive aggression (Aggression is Useful to Reach Goals and Aggression Protects Image) and a factor of beliefs legitimizing reactive aggression (Reactive Aggression is OK; see Table 2).

Procedure

Students were invited to participate and were given letters to their parents informing them about the purpose of the study. Letters also explained that they were free to request their children's withdrawal from the study by filling and returning an attached slip. A few parents requested more information about the study by phone (some asked for advice about how to manage aggression at home) but only five (0.4%) returned the slip denying their children's participation.

Two research assistants administered the questionnaire in each classroom. They did not read the questions out loud to the students, but were available to answer questions by individual students about how to fill out the questionnaire. All responses were anonymous. Data collection was conducted in all schools during the second month of the school year.

Data Analyses

Analyses of variance were performed to identify significant gender differences and differences between the localities in the means of all the variables. Cluster analyses were conducted to identify profiles of students in terms of their reactive and proactive aggression. Analyses of variance were carried out to check for differences between the clusters identified. Bivariate Pearson's correlations were performed to identify significant bivariate relationships between the variables. Multiple regressions were conducted using SPSS to identify the variables that best predict reactive and proactive aggression. The forward method of introduction of variables was used. Structural equation models were constructed using the program EQS to identify how much the relationships between exposure to violence and reactive or proactive aggression are mediated by the measured psychological variables.

Results

Differences between Localities

Reactive aggression was significantly higher in Santafé (high-crime downtown) compared to Usme (peripheral, moderate levels of political violence) and Ciudad Bolívar (high levels of political violence) (see Table 3). Students in Santafé also had significantly higher levels of proactive aggression when compared to Usme, but did not differ significantly from Ciudad Bolívar. Exposure to community violence was significantly lower in Usme when compared to all other groups. Exposure to youth gangs was highest among students in Ciudad Bolívar. Beliefs legitimizing aggression were significantly higher in Santafé when compared to Usme, with Ciudad Bolívar always in the middle. No significant differences between these localities were found in lack of guilt, hostile attribution of intent, and positive expectations for aggression (see Table 3).

Age and Sex Differences in Reactive and Proactive Aggression

Significant gender differences were found in almost all variables, with boys scoring higher in reactive and proactive aggression, exposure to community violence and gangs, lack of guilt, hostile attribution of intent, positive expectations for aggression and beliefs legitimizing aggression (see Table 4). In addition, participants were classified according to whether their scores in reactive or proactive aggression were higher or lower than the median score. In this way, 47.5% of the participants were classified as non-aggressive (NA), 14.5% were only reactive aggressive (RA), 14.4% were only proactive aggressive (PA), and 23.6% were both reactive and proactive aggressive (RPA). Girls were overrepresented in the NA group (57.1%) and underrepresented in all the other groups (47.2% in RA, 42.3% in PA, and 41.8% in RPA; χ²(3)=24.1; p<.000). Groups also differed in age, with older participants in the PA and RPA groups compared to the NA group (see Table 5).

Psychological Factors and Reactive and Proactive Aggression

Lack of guilt after behaving aggressively, hostile attribution of intent, positive expectations for aggression, and beliefs legitimizing reactive and proactive aggression were significantly correlated to reactive and proactive aggression scores (see Table 6) and were significantly higher in RPA in comparison to all the other groups (see Table 5). The RA and PA groups did not differ significantly in any of these psychological variables, but RAs were significantly higher than NAs in hostile attribution of intent and beliefs legitimizing aggression, while lack of guilt and beliefs legitimizing aggression were significantly higher in PAs than in NAs (see Table 5). In regression analyses, lack of guilt after aggression predicted proactive aggression; hostile attribution of intent, beliefs legitimizing reactive aggression, and beliefs legitimizing proactive aggression predicted both reactive and proactive aggression (see Table 7).

Community Violence and Psychological Factors

Exposure to community violence and to youth gangs were both significantly correlated to reactive and proactive aggression (see Table 6). However, in the regression analyses, exposure to community violence significantly predicted reactive and proactive aggression while exposure to youth gangs predicted only proactive aggression (see Table 7). Neither parental supervision nor sex moderated these relations.

Exposure to community violence was related to several psychological variables. As shown in Table 6, hostile attribution of intent, positive expectations for aggression, lack of guilt after behaving aggressively, and several normative beliefs were all significantly correlated with exposure to community violence.

Psychological Factors Mediating the Effect of Exposure to Community Violence on Reactive and Proactive Aggression

Structural equation models indicated that the aggregate of psychological factors measured (lack of guilt with aggression, hostile attribution of intent, positive expectations with aggression, and beliefs legitimizing aggression) mediated almost 60% of the effect of exposure to community violence into reactive and proactive aggression. More specific analyses indicated that beliefs legitimizing aggression had stronger mediation effects on both reactive and proactive aggression than all the other factors (see Tables 8 and 9, and Figures 1 and 2).

Discussion

Several studies have shown that exposure to community violence increases the chances of aggressive behaviors (Brookmeyer et al., 2005; Flannery et al., 2004; Gorman-Smith et al., 2004; Guerra et al., 2003; Liddell et al., 1994; Miller et al., 1999; Schwab-Stone et al., 1999; Schwartz & Proctor, 2000). This study adds evidence to this general finding by showing that children and adolescents living in neighborhoods where community violence is high due to political violence or to common crime, report higher levels of reactive and proactive aggression among peers than those who live in localities of the same city with lower levels of community violence. In addition, this study highlights some cognitive and emotional processes that may be mediating this relationship. In particular, children and adolescents exposed to higher levels of community violence feel less guilt after using aggression, expect more positive results when using aggression, have more beliefs legitimizing the use of aggression, and attribute more negative intentions to others. Furthermore, these cognitive and emotional biases seem to explain why they use more aggression than those who are not exposed to such high levels of community violence.

Beliefs legitimizing aggression were found to mediate the relationship between community violence and aggression better than any of the other psychological variables. This finding suggests that one of the greatest psychological effects of growing up in a violent environment might be the development of beliefs about the legitimacy of aggression. These beliefs, in turn, might have a great impact on behavior since they help remove moral restraints to using aggression (such as feeling guilt) and may even make using aggression socially and morally desirable (Bandura, 1999; Liau, Barriga, & Gibbs, 1998).

This study was not designed to test for the direction of causal relationships. It actually seems just as likely that aggressive beliefs lead to aggressive behavior as it is that aggressive behavior leads to aggressive beliefs. Beliefs might appear after behaving aggressively as a way to calm conscience. Similarly, it seems plausible that aggressive children and adolescents are more prone to seek violent environments and therefore will be more exposed to community violence than non-aggressive children. Although it seems reasonable that all these relations might be bidirectional, Guerra et al. (2003) found in a longitudinal study that exposure to violence predicted normative beliefs and aggressive behaviors better than how normative beliefs and aggression predicted exposure to violence. In addition, Guerra et al. found that normative beliefs partially mediated the lagged effect of exposure to community violence on aggressive behavior. In any case, the consistent relationships found in this study, as well as in Guerra et al. (2003) and Musher-Eizenman et al. (2004), between beliefs and both exposure to violence and aggression suggest that more attention should be paid to this variable. Educational interventions able to change beliefs legitimizing aggression (e.g., Guerra & Slaby, 1990) might have an important potential to prevent violence, especially in violent contexts.

Contrary to our hypotheses, the results found in this study for reactive and proactive aggression were very similar. Although the correlation between the two (.50) was not as high as that found in other studies (e.g., .76 in Dodge & Coie, 1987), reactive and proactive aggression were related in very similar ways to all other variables. For example, there were no statistical significant differences in any of the community or psychological variables measured between the participants who scored high only on reactive aggression and those who scored high only on proactive aggression. Additionally, except for the different types of beliefs legitimizing aggression, the mediating variables between exposure to violence and reactive aggression were very similar to those between exposure to violence and proactive aggression. One possibility is that theorized differences between the two (e.g., Dodge, 1991) might only be evident in very extreme cases. In fact, in an examination of the results of a national test that more than a million fifth- and ninth-grade students in Colombia have to take, significant differences between reactive and proactive aggressive children in their levels of empathy and in their capacity to control anger were only evident when those who reported very high levels of reactive aggression but no proactive aggression were compared to those who reported very high levels of proactive aggression but no reactive aggression (Chaux, Arboleda, Kanayet, & Torrente, 2005). This suggests that, although conceptual differences between reactive and proactive aggression are very useful in practical terms since they suggest different prevention strategies, it continues to be a challenge to disentangle them in research (see also Bushman & Anderson, 2001).

Finally, students living in communities with moderate levels of crime and political violence (Ciudad Bolívar) had lower levels of reactive aggression, but similar levels of proactive aggression, when compared to students living in high-crime areas with no political violence (Santafé). This suggests that exposure to political violence in their communities might be more related to the development of proactive rather than reactive aggression. Children exposed to political violence might be learning early in life that violence could be used as an instrument to reach power and other goals. However, evidence here is preliminary and more studies should be conducted before a clear conclusion could be reached about the different effects of crime-based versus politically-based community violence.

There are several limitations of the study that need to be acknowledged. First, this was not a longitudinal study and therefore it was not possible to analyze the evolution of the variables or the direction of the relationships among them. Second, all the information is based on the same reporter and therefore it is possible that the relationship between some of the variables could have been overestimated. Third, all measures were self-reports, which creates risks of social desirability. Fourth, some of the measures created for this study did not reach high internal consistency (e.g., hostile attribution of intent) and therefore all results related to these variables should be interpreted with caution.

In addition, the study did not consider other variables which might help explain the complexity of the relationship between exposure to violence and development of aggression. For example, exposure to community violence was limited to witnessing events. Witnessing community violence seems to have different effects than being a victim of community violence (Schwartz & Proctor, 2000). Emotional variables were limited to lack of guilt after aggression, leaving out emotional competencies such as regulation of own anger, which are known to be related to aggression and especially to reactive aggression (Chaux et al., 2005; Dodge, 1991; Schwartz & Proctor, 2000). Exposure to violence was limited to community violence, excluding the possible interaction with other contexts where children and adolescents might be exposed to violence such as the family and the media. Finally, although the particular context of this study made it possible to contrast community violence related to common crime with that related to political violence, for security reasons, exposure to political violence was only measured indirectly with police records about the neighborhoods where the schools are located (i.e., there were no specific questions about participants' contact with urban guerrilla or paramilitary militias) limiting the possible statistical analyses that could be conducted to explore this difference in detail. All these identified variables suggest directions for future research and are in fact being considered in ongoing studies in the same context.

In spite of its limitations, this study contributes to our understanding of the relationships between community violence and children and adolescent's aggression and suggests points of intervention for prevention programs. In particular, the study highlights the potential that changing beliefs might have on reducing youth aggression. These interventions might be needed to contribute to the reduction of reactive and/or
proactive aggression, and in community contexts where common crime and/or political violence are common.

^* This study was made possible thanks to the financial contribution of COLCIENCIAS (project number 0026-2001). We also thank Amparo Ardila of the Department of Education in Bogotá and the principals, teachers, and students of the Manuela Beltrán, Compartir, El Recuerdo, Miguel de Cervantes Saavedra, Aulas Colombianas El Consuelo, Antonio José Uribe, INEM Santiago Pérez, Agustín Codazzi, Santiago de las Atalayas and Miravalle schools. We are very grateful to Ana María Velásquez for her contributions to the statistical analyses, María Angélica Gómez, Nancy González, Julio Cesar Pineda, and Bertha Sánchez for their participation in the data collection, and Álvaro Camacho, Angelika Rettberg, Frank Vitaro, and the anonymous reviewers for their very helpful feedback.

¹ The complete questionnaire in Spanish as well as its English translation is available upon request from the first author.

References

Alcaldía Mayor de Bogotá. (2005). Ficha técnica situación humanitaria localidad Ciudad Bolívar Bogotá D.C. Secretaría de Hacienda. Departamento Administrativo de Planeación. Retrieved from galeon. com/alianzakennedy/situacion.pdf         [ Links ]

Allwood, M. A. & Bell, D. J. (2008). A preliminary examination of emotional and cognitive mediators in the relations between violence exposure and violent behaviors in youth. Journal of Community Psychology, 36, 989-1007.         [ Links ]

Arsenio, W. F. & Lemerise, E. A. (2001). Varieties of childhood bullying: Values, emotion processes, and social competence. Social Development, 10, 59-73.         [ Links ]

Bandura, A. (1991). Social cognitive theory of moral thought and action. In W. M. Kurtines & J. L. Gewirtz (Eds.), Handbook of moral behavior and development: Theory, research and applications (pp. 71-129). Hillsdale, NJ: Lawrence Erlbaum Associates.         [ Links ]

Bandura, A. (1999). Moral disengagement in the perpetration of inhumanities. Personality and Social Psychology Review, 3, 193-209.         [ Links ]

Bradshaw, C. P., Rodgers, C. R. R., Ghandour, L. A., & Garbarino, J. (2009). Social-cognitive mediators of the association between community violence exposure and aggressive behavior. School Psychology Quarterly, 24, 199-210.         [ Links ]

Brendgen, M., Vitaro, F., Tremblay, R., & Lavoie, F. (2001). Reactive and proactive aggression: Predictions to physical violence in different contexts and moderating effects of parental monitoring and caregiving behavior. Journal of Abnormal Child Psychology, 29, 293-304.         [ Links ]

Brookmeyer, K. A., Henrich, C. C., & Schwab-Stone, M. (2005). Adolescents who witness community violence: Can parent support and prosocial cognitions protect them from committing violence? Child Development, 76, 917-929.         [ Links ]

Bushman, B. J. & Anderson, C. A. (2001). Is it time to pull the plug on the hostile versus instrumental aggression dichotomy? Psychological Review, 108, 273-279.         [ Links ]

Calvete, E. & Orue, I. (2011). The impact of violence exposure on aggressive behavior through social information processing in adolescents. American Journal of Orthopsychiatry, 81, 38-50.         [ Links ]

Chaux, E. (2001). Peer conflicts in a violent environment: Strategies, emotions, reconciliations, and third parties in conflicts among Colombian children and early adolescents (Unpublished doctoral dissertation). Harvard University. Cambridge, Massachusetts.         [ Links ]

Chaux, E., Arboleda, J., Kanayet F., & Torrente C. (2005, November). Citizenship competencies: Results from a large-scale survey. Meeting of the Association for Moral Education. Cambridge, Massachusetts.         [ Links ]

Crick, N. R. & Dodge, K. A. (1996). Social-informationprocessing mechanisms in reactive and proactive aggression. Child Development, 67, 993-1002.         [ Links ]

Dodge, K. A. (1991). The structure and function of reactive and proactive aggression. In D. J. Pepler & K. H. Rubin (Eds.), The development and treatment of childhood aggression (pp. 201-218). Hillsdale, NJ: Erlbaum.         [ Links ]

Dodge, K. A., Bates, J. E., & Pettit, G. S. (1990). Mechanisms in the cycle of violence. Science, 250, 1678-1683.         [ Links ]

Dodge, K. A. & Coie, J. D. (1987). Social-informationprocessing factors in reactive and proactive aggression in children's peer groups. Journal of Personality and Social Psychology, 53, 1146-1158.         [ Links ]

Dodge, K. A., Lochman, J. E., Harnish, J. D., Bates, J. E., & Pettit, G. S. (1997). Reactive and proactive aggression in school children and psychiatrically impaired chronically assaultive youth. Journal of Abnormal Psychology, 106, 37-51.         [ Links ]

Endresen, I. M. & Olweus, D. (2001). Self-reported empathy in Norwegian adolescents: Sex differences, age trends, and relationship to bullying. In A. C. Bohart & D. J. Stipek (Eds.), Constructive & destructive society (pp. 147-165). Washington, DC, US: APA.         [ Links ]

Flannery, D. J., Wester, K. L., & Singer, M. I. (2004). Impact of exposure to violence in school on child and adolescent mental health and behavior. Journal of Community Psychology, 32, 559-573.         [ Links ]

Gorman-Smith, D., Henry, D. B., & Tolan, P. H. (2004). Exposure to community violence and violence perpetration: The protective effects of family functioning. Journal of Clinical Child and Adolescent Psychology, 33, 439-449.         [ Links ]

Guerra, N., Huesmann, R., & Spindler, A. (2003). Community violence exposure, social cognition and aggression among urban elementary school children. Child Development, 74, 1561-1576.         [ Links ]

Guerra, N. G. & Slaby, R. G. (1990). Cognitive mediators of aggression in adolescent offenders: II. Intervention. Developmental Psychology, 26, 269-277.         [ Links ]

Hoffman, M. L. (1998). Varieties of empathy-based guilt. In J. Bybee (Ed.), Guilt and children (pp. 91112.). San Diego, California: Academic Press.         [ Links ]

Huesmann, L. R., Eron, L. D., Lefkowitz, M. M., & Walder, L. O. (1984). Stability of aggression over time and generations. Developmental Psychology, 20, 1120-1134.         [ Links ]

Kochanska, G., Gross, J. N., Lin, M., & Nichols, K. E. (2002). Guilt in young children: Development, determinants, and relations with a broader system of standards. Child Development, 73, 461-482.         [ Links ]

Liau, A. K., Barriga, A. Q., & Gibbs, J. C. (1998). Relations between self-serving cognitive distortions and overt vs. covert antisocial behavior in adolescents. Aggressive Behavior, 24, 335-346.         [ Links ]

Liddell, C., Kvalsvig, J., Qotyana, P., & Shabalala, A. (1994). Community violence and young South African children's involvement in aggression. International Journal of Behavioral Development, 17, 613-628.         [ Links ]

Little, T. D., Jones, S. M., Henrich, C. C., & Hawley, P. H. (2003). Disentangling the 'whys' from the 'whats' of aggressive behaviour. International Journal of Behavioral Development, 27, 122-133.         [ Links ]

Milich, R. & Dodge, K. A. (1984). Social information processing in child psychiatric populations. Journal of Abnormal Child Psychology, 12, 471-489.         [ Links ]

Miller, L. S., Wasserman, G. A., Neugebauer, R., Gorman-Smith, D., & Kamboukos, D. (1999). Witnessed community violence and antisocial behavior in high-risk, urban boys. Journal of Clinical Child Psychology, 28, 2-11.         [ Links ]

Musher-Eizenman, D. R., Boxer, P., Danner, S., Dubow, E. F., Goldstein, S. E., & Heretick, D. M. L. (2004). Social-cognitive mediators of the relation of environmental and emotion regulation factors to children's aggression. Aggressive Behavior, 30, 389-408.         [ Links ]

Orobio de Castro, B., Merk, W., Koops, W., Veerman, J. W., & Bosch, J. D. (2005). Emotions in social information processing and their relations with reactive and proactive aggression in referred aggressive boys. Journal of Clinical Child and Adolescent Psychology, 34, 105-116.         [ Links ]

Orobio de Castro, B., Veerman, J. W., Koops, W., Bosch, J. D., & Monshouwer, H. J. (2012). Hostile attribution of intent and aggressive behavior: A meta-analysis. Child Development, 73, 916-934.         [ Links ]

Orue, I., Bushman, B. J., Calvete, E., Thomaes, S., Orobio de Castro, B., & Hutteman R. (2011). Monkey see, monkey do, monkey hurt: Longitudinal effects of exposure to violence on children's aggressive behavior. Social Psychological and Personality Science, 2, 432-437.         [ Links ]

Parra, A. (2005). La relación entre la exposición a violencia comunitaria y la violencia severa y roles en intimidación escolar: variables mediadoras y moderadoras [The relationship between exposure to community violence and severe violence and participant roles in school bullying: Mediator and moderator variables] (Unpublished master's Thesis). Universidad de los Andes, Bogotá, Colombia.         [ Links ]

Poulin, F. & Boivin, M. (2000). The role of proactive and reactive aggression in the formation and development of friendships in boys. Developmental Psychology, 36, 1-8.         [ Links ]

Punamäki, R. L., Muhammed, A. H., & Abdulrahman, H. A. (2004). Impact of traumatic events on coping strategies and their effectiveness among Kurdish children. International Journal of Behavioral Development, 28, 59-70.         [ Links ]

Schwab-Stone, M., Chen, C., Greenberger, E., Silver, D., Lichtman, J., & Voyce, C. (1999). No safe haven II: The effects of violence exposure on urban youth. Journal of the American Academy of Child & Adolescent Psychiatry, 38, 359-367.         [ Links ]

Schwartz, D. & Proctor, L. J. (2000). Community violence exposure and children's social adjustment in the school peer group: The mediating roles of emotion regulation and social cognition. Journal of Consulting and Clinical Psychology, 68, 670-683.         [ Links ]

Slaby, R. G. & Guerra, N. G. (1988). Cognitive mediators of aggression in adolescent offenders: 1. Assessment. Developmental Psychology, 24, 580-588.         [ Links ]

Stuckless, N. & Goranson, R. (1992). The Vengeance Scale: Development of a measure of attitudes toward revenge. Journal of Social Behavior & Personality, 7, 25-42.         [ Links ]

Thornberry, T. P., Freeman-Gallant, A., Lizotte, A. J., Krohn, M. D., & Smith, C. A. (2003). Linked lives: The intergenerational transmission of antisocial behavior. Journal of Abnormal Child Psychology, 31, 171-184.         [ Links ]

Vitaro, F. & Brendgen, M. (2005). Proactive and reactive aggression: A developmental perspective. In R. E. Tremblay, W. W. Hartup, & J. Archer (Eds.), Developmental origins of aggression (pp. 178-201). New York, US: Guilford Press.         [ Links ]

Vitaro, F., Gendreau, P. L., Tremblay, R. E., & Oligny, P. (1998). Reactive and proactive aggression differentially predict later conduct problems. Journal of Child Psychology and Psychiatry, 39, 377-385.         [ Links ]

Widom, C. S. (1989). The cycle of violence. Science, 244, 160-166.         [ Links ]

Zelli, A., Dodge, K. A., Lochman, J. E., Laird, R. D., & Conduct Problems Prevention Research Group. (1999). The distinction between beliefs legitimizing aggression and deviant processing of social cues: Testing measurement validity and the hypothesis that biased processing mediates the effects of beliefs on aggression. Journal of Personality and Social Psychology, 77, 150-166.         [ Links ]