Political Violence and Adolescent Out-group Attitudes and Prosocial Behaviors: Implications for Positive Inter-group Relations

Abstract

The negative impact of political violence on adolescent adjustment is well-established. Less is known about factors that affect adolescents’ positive outcomes in ethnically-divided societies, especially influences on prosocial behaviors toward the outgroup, which may promote constructive relations. For example, understanding how intergroup experiences and attitudes motivate outgroup helping may foster intergroup cooperation and help to consolidate peace. The current study investigated adolescents’ overall and outgroup prosocial behaviors across two time points in Belfast, Northern Ireland (N = 714 dyads; 49% male; Time 1: M = 14.7, SD = 2.0, years old). Controlling for Time 1 prosocial behaviors, age and gender, multivariate structural equation modeling showed that experience with intergroup sectarian threat predicted fewer outgroup prosocial behaviors at Time 2 at the trend level. On the other hand, greater experience of intragroup nonsectarian threat at Time 1 predicted more overall and outgroup prosocial behaviors at Time 2. Moreover, positive outgroup attitudes strengthened the link between intragroup threat and outgroup prosocial behaviors one year later. Finally, experience with intragroup nonsectarian threat and outgroup prosocial behaviors at Time 1 was related to more positive outgroup attitudes at Time 2. The implications for youth development and intergroup relations in post-accord societies are discussed.

Relations between political violence and adolescent adjustment problems, including aggression and hostility, are well-established (Merrilees et al., in press; Keresteš, 2006; Qouta, Punamaki, Miller, & El-Sarraj, 2008). However, relatively less is known about intergroup factors that affect adolescents’ positive behaviors in these contexts (Barber, 2009), including prosocial behavior toward outgroup members (Tropp & Mallet, 2011). In ethnically-divided societies, understanding how inter- and intra-group experiences and attitudes motivate prosocial behavior across group lines, such as outgroup helping (Stürmer & Snyder, 2009; Vollhardt, 2009), may provide bases for fostering intergroup cooperation and ultimately consolidating peace (Kelman, 2008). The aim of the present study is to better understand conflicting findings on the impact of political violence on youth prosocial behaviors. To do so it will tease apart adolescents’ overall prosocial behavior from that directed toward outgroup members. In addition, analyses will examine the reciprocal relations among these prosocial behaviors and different types of community threat and outgroup attitudes. By differentiating between risk factors and adolescent outcomes over time, these findings may contribute to resolving conflicts that have spanned generations. Specifically, understanding the factors that promote resilience and prosocial psychosocial behaviors among youth born after a peace agreement may offer avenues for consolidating peace processes (McEvoy-Levy, 2006).

Prosocial behaviors, or voluntary acts that benefit another without personal profits or external awards (Bar-Tal, 1976), are an example of positive youth adjustment that may have long-term implications for social justice and peacebuilding. Yet, the research on the development of prosocial behaviors, such as helping, altruism and social solidarity in contexts of political violence and intergroup tension is mixed (Stürmer & Snyder, 2009). Some studies suggest a positive link between exposure to political violence and constructive prosocial engagement. Former child soldiers in Uganda were more likely to take on political leadership roles compared to non-abducted peers (Blattmann, 2009), Intifada exposure was related to greater activism among Palestinian youth (Barber, 2009), and during the civil war in Lebanon, separation from parents and witnessing violent acts was linked to increased levels of youth prosocial behaviors (Macksoud & Aber, 1996). On the other hand, some research suggests a negative relation between experiencing intergroup conflict and youth prosocial acts. In Croatia, greater war stress related to fewer adolescent prosocial acts (Keresteš, 2006); in Sierra Leone, child soldiers who reported killing or injuring others showed a decrease in prosocial behaviors in the post-war context (Betancourt, Brennan, Rubin-Smith, Fitzmaurice & Gilman, 2010). In Northern Ireland, experience with sectarian, or intergroup, violence was directly related to fewer prosocial acts; however, nonsectarian antisocial acts, or intragroup threats from one’s own community, predicted more prosocial behaviors in general (Cummings et al., 2010b).

These conflicting findings may reflect that experiences with different types of violence and threats may be related to different forms of positive adjustment. For example, whether threats were in the past during overt conflict or war or are part of the daily life in a post-accord period, may relate to divergent findings. In the current study, youth were born after the height of the contemporary violence in Northern Ireland, known as the ‘Troubles.’ Therefore, this paper extends past work focused on conflict experiences during war by focusing on experience with current, on-going intergroup and intragroup conflict. In addition, this work extends studies that focus on outgroup, rather than ingroup, threat, and also complements experimental studies to examine relations among groups in the social ecological of adolescent’s lives (Nesdale, Durkin, Maass, & Griffiths, 2005; Nesdale, Maass, Durkin, & Griffiths, 2005). Finally, this study distinguishes between targets of positive youth outcomes. That is, rather than measuring only general prosocial behaviors that could be directed toward any target, the current study differentiates between overall prosocial behavior and positive acts directed at members of the ‘other’ group. Thus, the findings hold social significance, as outgroup helping may be a key factor in fostering intergroup cooperation and serve as a building block for more peaceful relations in divided societies (Kelman, 2008; Stürmer & Snyder, 2009).

Promoting Outgroup Prosocial Behaviors

Altruism born of suffering is a theory that describes how experiencing adversity may enhance “motivation to help other disadvantaged members of society, including outgroups” (Vollhardt, 2009, p. 53). The theory specifically identifies the risk of intentional harm caused by other humans, which can threaten individuals and groups. Altruism, or the motivation to improve the well-being of others (Batson, Klein, Highberger, & Shaw, 1995), can be enhanced because individual or collective pain increases perspective taking, heightens awareness of injustice more broadly, and promotes identification with victims of similar types of violence. Identifying with those who have suffered, even victims from the “other” group, can influence intergroup helping behaviors (Stürmer, Snyder, & Omoto, 2005).

Altruism born of suffering can also be related to cognitive shifts, such as perceived common fate, in settings of intergroup conflict (Staub & Vollhardt, 2008). Conditions of high stress may inspire individuals to help others they perceive to share a common fate (Dovidio & Morris, 1975; Vollhardt, 2009). Moreover, the “salience of victimization may create the perception of shared group membership among individuals who suffered, giving rise to increased motivations to help other victims” (Vollhardt, 2009, p. 70). For example, even in conflict settings where individuals would typically be categorized as an outgroup member, the experience of mass violence provided the motivation to re-conceptualize other victims as fellow sufferers (Penner et al., 2005). Thus, altruism born of suffering outlines how this cognitive shift toward shared membership based on perceived common fate between conflicting groups may underlie outgroup prosocial behaviors (Reicher, Cassidy, Wolpert, Hopkins & Levine, 2006). Shared identities do not replace group membership, however, but rather re-define shared affiliations (Gaertner & Dovidio, 2000), increasing motivations underlying prosocial behaviors, including toward outgroup members.

Constraining Outgroup Prosocial Behaviors

Another set of theories suggest that threat from the outgroup may dampen or discourage the will to help members of that group. Social identity theory and self-categorization theory (Tajfel & Turner, 1979; Turner, 1982) explain that external threats and intergroup violence may lead individuals to help only those they perceive to be similar to themselves (Reicher, Hopkins, & Condor, 1997). In emergency situations, individuals are more likely to help those with shared membership when ingroup identity is salient (Levine, Prosser, Evans, & Reicher, 2005). In the face of intergroup violence, the ingroup may pull together (Taylor et al., 2011), and stronger social identity may promote individuals to work together for social and political change (Hammack, 2010). Thus, when individuals in tight-knit groups experience intergroup or sectarian violence, they should be more likely to help their own group (Coser, 1956), which may decrease the likelihood of reaching out to help members of the other group.

In settings of intergroup conflict, it is also possible that helping peers from the other group may be perceived negatively by the ingroup. For example, children with stronger ingroup commitment were more likely to negatively judge their peers who deviated from group loyalty norms (Abrams, Rutland, Ferrell, & Pelletier, 2008). Following this “black sheep” phenomenon (Abrams, Palmer, Rutland, Cameron & Van de Vyver, 2013; Marques, Yzerbyt, & Leyens, 1988), deviant acts, such as acting prosocially toward outgroup members, may lead to social sanctions or punishment (Pinto, Marques, Levine, & Abrams, 2010). Cooperating with the other side may even put an individual at risk for greater intragroup threat or retaliatory punishment violence (Monaghan & McLaughlin, 2006). In this context, fear of facing future intragroup threat may constrain adolescents from engaging in constructive behaviors and relations with outgroup peers. However, not all antisocial behavior committed by the ingroup is politically motivated; common crime or nonsectarian threats, such as substance misuse or gang membership, that occur within the boundaries of homogenous neighborhoods constitute another form of intragroup threat (van der Merwe & Dawes, 2007). That is, youth may withdraw from community, decreasing the amount of prosocial acts, as they try to avoid antisocial behavior unrelated to historic conflict in Northern Ireland.

Moderating the Risk Factors

Moderation tests are particularly relevant to psychosocial processes that are not uniform across people (Cummings, Davies, & Campbell, 2000). Within a risk and resilience framework, moderators can help to identify for whom and under what conditions the link between risks and outcomes is stronger. In a conflict environment, intergroup attitudes may moderate how different types of threat affect outgroup prosocial behaviors in different scenarios (Dovidio, Gaertner, Shnabel, Saguy, & Johnson, 2010). For example, when individuals hold negative intergroup perceptions or attitudes, they may be more likely to avoid intergroup interactions or respond with hostility or aggression toward the other group (Van Zomeren, Fischer & Spears, 2007). On the other hand, individuals with more positive outgroup attitudes may be more likely to exhibit approach responses, such as prosocial behaviors, in intergroup settings (Butz & Plant, 2011). However, compared to negative attitudes and prejudice (Rutland, Abrams, & Levy, 2007), there is far less research on the role of positive intergroup evaluations (Tropp & Mallett, 2011; Stürmer & Snyder, 2009). Therefore, the role of intergroup attitudes and perceptions should be considered to better understand intergroup interactions and behaviors (Tropp & Pettigrew, 2004), which have practical implications for social reconstruction in divided societies.

Intergenerational Political Conflict

During a post-accord period in protracted conflicts, the threat of on-going intergroup violence is well-documented (Walter, 2004). Approximately 50% of nations that experience a civil war will experience renewed mass violence (Doyle & Sambanis, 2000). Post-accord periods are also susceptible to new forms of criminal violence. However, the emergence of new threats, such as intragroup violence, receives far less attention in the changing post-accord environment. Following the peace agreements in South Africa and Northern Ireland, levels of common crime have increased as forms of political, or sectarian violence, declined (PSNI, 2011; Shirlow, et al., 2012; van der Merwe & Dawes, 2007). The impact of nonsectarian crime, or intragroup antisocial behaviors, has also been shown to have distinct effects on child adjustment and family processes (Cummings et al., 2010a, 2010b; Taylor et al., in press). Other post-accord challenges, such as daily hardships and community acceptance, have also been found to have predictive power for youth outcomes independent of war experiences (Betancourt et al., 2010). Thus, multiple post-accord challenges should be considered to more fully understand adolescent social development.

Protracted or intractable conflict spans decades, and as such, requires peacebuilding plans on a similar scope. This “decades-thinking” approach (Lederach, 1997), or long-term visions for a shared future, highlights the important role of youth in these processes. Yet young people are often portrayed as helpless victims or violent perpetrators of conflict (Barber, 2009); their positive agency is often overlooked (McEvoy-Levy, 2006). If peacebuilding entails fostering the structures and processes that redefine violent relationships into constructive and cooperative patterns (Lederach, 1997), then in settings of protracted conflict, young people must be engaged in establishing new patterns of more collaborative intergroup relations. Youth contributions, such as intergroup prosocial behaviors, may become particularly important in post-accords periods, where their absence in decision making processes is particularly pronounced (McEvoy-Levy, 2006).

Post-Accord Northern Ireland

The conflict in Northern Ireland extends back for centuries, with the most recent outbreak of sectarian violence occurring between 1968 and 1998, known as the Troubles. The dispute is largely a constitutional one, with the Unionists/Protestants wishing to remain part of the United Kingdom while the Republicans/Catholics aim for Ireland’s re-unification. Despite the 1998 Belfast Agreement, sectarian violence persists (PSNI, 2011). For example, in 2009 in Belfast, there were over 1,200 sectarian incidents reported to the police, including acts of violence and damage to symbolic property such as churches. Yet, “‘minor’ forms of sectarianism, such as verbal abuse, harassment, visual displays, and graffiti” (Jarman, 2005, p. 21) are largely under-reported in police data. Other studies have found that one in four adolescents is the victim of sectarian violence and over 80% will experience sectarianism directly or indirectly (Byrne, Conway, & Ostermeyer, 2005; McAloney, McCrystal, Percy, & McCartan, 2009).

In addition to continued sectarianism, new forms of nonsectarian antisocial behavior and crime, or intragroup threats posed by ingroup members, have grown, in part, due to the lack of social control by the paramilitaries (Monaghan & McLaughlin, 2006). The increase in non-conflict related crime is not unique to Northern Ireland (Cox, 2000). In the context of post-accord Belfast, nonsectarian antisocial behavior is understood by local residents as being perpetrated by “teenagers within their own communities” (Goeke-Morey et al., 2009, p. 373). Because of the largely segregated nature of housing (Shirlow & Murtagh, 2006), this conventional type of antisocial behavior, such as drinking, drugs, and home robberies, occurs within the boundaries of homogenous neighborhoods (Taylor et al., 2011). Along with the rise in conventional nonsectarian crime, there are also continued paramilitary-style assaults and shootings (see “Man shot in legs,” 2013). Thus, youth born after the Belfast Agreement experience both inter- and intragroup threat; both sources of stress should be considered to more completely understand the possible contributions of youth to peacebuilding in Northern Ireland (Cummings, Goeke-Morey, Schermerhorn, Merrilees, & Cairns, 2009).

Current Study

The current paper examines how different types of post-accord threats, both intergroup and intragroup, and intergroup attitudes affect adolescent prosocial behaviors. While overall prosocial behaviors may have benefits for society more broadly, outgroup helping among youth may serve as a building block for future intergroup cooperation. Based on previous research, it is hypothesized that the source of threat will have opposite effects on adolescent outgroup helping; intergroup threat will decrease outgroup prosocial behavior, whereas threat from one’s own group will be linked to greater outgroup prosocial behavior. We also hypothesize that this positive relationship will be stronger for youth who have positive attitudes toward the outgroup.

Method

Participants

Mother/adolescent dyads (N = 714) were drawn from waves 4 and 5 of a longitudinal study on political violence, family processes, and youth adjustment in Belfast, Northern Ireland. A primary benefit of including mother and child reports includes decreasing shared method variance of a single-reporter in survey data. Children were evenly balanced in terms of gender (49% male) and were 14.66 (SD = 1.96) and 15.75 (SD = 1.97) years old on average at each of the two time points. Mothers were 39.73 (SD = 6.22) and 40.72 (SD = 6.32) years old on average, and 63% were single while 37% were in two-parent households. Representative of the demographics of Northern Ireland, all participants were White, and 59% and 41% were from Protestant and Catholic backgrounds, respectively.

In the current analyses, between Time 1 and Time 2 there was 81% retention. Compared to families who remained in the study, youth that did not return reported greater experience with nonsectarian antisocial behavior (t(519) = 2.52, p < .05; M = 5.22, SD = 6.63 attrition; M = 3.68, SD = 5.36 retention), more positive outgroup attitudes (t(628) = 2.48, p < .05; M = 54.67, SD = 28.49 attrition; M = 48.21, SD = 25.14 retention), and greater overall prosocial behavior (t(595) = 2.17, p < .05; M = 2.54, SD = .55 attrition; M = 2.41, SD = .61 retention). There were no significant differences among those who participated at Time 2 in child gender, age, experience with sectarian antisocial behavior, or outgroup prosocial behaviors.

Procedures

Participants were recruited through stratified random sampling that identified approximately 35 families in each of 22 interfaced neighborhoods selected to minimize socio-economic status differences and represent a range of experiences with sectarian conflict. Interfaced areas are ethnically homogenous neighborhoods (greater than 90% Catholic or Protestant) that border a community in which the other group is the majority. Oftentimes these interfaced areas are separated by “peace walls” or major roads which further enhance segregated living patterns that extend into personal, educational, and professional life (Shirlow & Murtagh, 2006). A common characteristic of interfaced areas is the lack of employment and livelihood opportunities. As such, all the study neighborhoods were working class and ranked in the bottom quarter of the most-socially deprived wards in Northern Ireland. Finally, there was a variety of experiences with political violence during the Troubles across the study areas. Neighborhoods were selected to include both higher and lower levels of historical political deaths, as well as variation in more recent levels of sectarian crime.

Children and mothers provided assent and consent prior to participating, and all data were collected through annual face-to-face surveys conducted in the participant’s home by professionals from an established market research firm in Northern Ireland. Each interview lasted approximately 60 minutes and families received £40 at Time 1 and £50 at Time 2 for their participation. The Institutional Review Board at all participating universities approved the study.

Measures

Intragroup threat was measured using the nonsectarian antisocial behavior scale. This set of questions was designed to measure ordinary crime that was not related to the intergroup conflict. In the Belfast focus groups, participants clearly distinguished between sectarian and nonsectarian acts. For example, nonsectarian antisocial behaviors largely emerged after the peace accord, were committed by one’s “own” against the ingroup, and occurred within rather than between homogenous communities (Taylor et al., 2011), and these qualitative data were confirmed by a two-wave pilot test (Goeke-Morey et al., 2009). These papers established the construct and predictive validity of this scale. In the current analyses, youth reported on how frequently in the last three months a series of items occurred using a 5-point scale from 0 (not in the last 3 months), 1 (once in the last 3 months), 2 (every month), 3 (every week), to 4 (every day). The 7-item scale included items such as drugs being used or sold, home break-ins, and robberies/muggings. Internal consistency measures the degree to which items aiming to measure the same general construct produce similar scores. Cronbach’s α represents how strongly correlated are different pairs of items are correlated. Across the two waves, internal consistency of the intragroup threat scale was .89 and .88, respectively, indicating good to excellent correlations among the items.

Intergroup threat was assessed with the sectarian antisocial behavior, which was developed to assess awareness of sectarian, or intergroup, antisocial behaviors in the context of Belfast. Through qualitative focus groups and a two-wave pilot study (Goeke-Morey et al., 2009; Taylor et al., 2011), sectarian acts were identified as those that were rooted in the legacy of intergroup violence, were committed by “them” against “us,” and frequently occurred along interface lines or boundaries. Using the same response scale as the intragroup, nonsectarian threat, the 12 items included events such as stones or objects thrown over walls, name calling by people from the other community, and deaths or serious injury caused by the other community (Cummings et al., 2010b). Cronbach’s α for this scale was .94 at both time points.

Outgroup attitudes were assessed by two separate questions that asked youth to indicate their “overall feeling toward the Protestant/Catholic community” (Cairns, Kenworthy, Campbell, & Hewstone, 2006). The scale represented a feeling thermometer, ranging from 0 (unfavorable) to 100 (favorable). This approach to distinguishing between ingroup and outgroup attitudes (Brewer, 2001) has been used previously in Northern Ireland and other settings of intergroup conflict (Hewstone, Rubin, & Willis, 2002). Only outgroup attitudes were in analyses; that is, if a child self-identified as Catholic, only the Protestant attitude score was used.

Prosocial behaviors were assessed by both overall and outgroup-specific helping acts. Mothers reported on child overall prosocial behaviors using the prosocial with peers subscale of the Child Behavior Scale (CBS; Ladd & Profilet, 1996). On a scale from 1 (doesn’t apply), 1 (applies sometimes), to 3 (certainly applies), mothers responded to a 7-item scale to the degree their child shows a recognition of the feelings of others, cooperative with other children, and offers help or comfort when other children are upset. This subscale has been used to measure prosocial behaviors internationally (Meyer et al., 2011), and its construct and criterion validity have been established longitudinally through early adolescence (Ladd, Herald-Brown, & Andrews, 2009). Cronbach’s α were .96 and .95 across the two time points for overall prosocial behaviors.

Outgroup prosocial behaviors were assessed following overall demonstration of empathy, cooperation, and concern, with a single item question that asked, “Thinking about all these things, how often did your child do these toward people from the OTHER community?” Mothers responded on a 5-point scale from 0 (never) to 4 (very often).

Results

Table 1 includes the means, standard deviations, and bivariate correlations for all manifest variables, controls, and a composite scale of overall prosocial behavior. AMOS Graphics 18 (Arbuckle, 2009) was used to model the following paths of interest using structural equation modeling (SEM). For the SEM model of direct and moderation effects, coefficients were derived using maximum likelihood which accurately estimates parameters under the assumption that data are missing at random. Overall model fit was assessed a χ²/df index ≤ 3, a Tucker Lewis Index (TLI) and comparative fit index (CFI) ≥ .90, and a root mean square residual (RMSEA) ≤ .08 (Hu & Bentler, 1999).

Table 1. Means, Standard Deviations, and Correlations for All Study Variables (N=714).

Variables	M	SD	1	2	3	4	5	6	7	8
1. Gender			-
2. Age (T1)	14.	1.9	−0.04	-
	66	6
3. Intergroup Threat (T1)c	2.9	6.8	-	-	-
	1	7	0.15*	0.01
			**
4. Intragroup Threat (T1)c	4.0	5.6	-	0.15	0.53*	-
	0	7	0.13*	**	**
			*
5. Outgroup Attitudes (T1)c	49.	25.	0.15*	0.01	-	-	-
	50	93	**		0.20*	0.26*
					**	**
6. Overall Prosocial Behaviors (T1)m	2.4	0.6	0.27*	-	0.13*	0.23*	−0.07	-
	4	0	**	0.07	*	**
7. Overall Prosocial Behaviors (T2)m	2.4	0.6	0.22*	-	0.10*	0.27*	-	0.71*	-
	1	3	**	0.07		**	0.11*	**
8. Outgroup Prosocial Behaviors (T1)m	1.0	0.9	0.04	-	0.07	0.06	0.32*	0.17*	0.14*	-
	7	7		0.09			**	**	*
				*
9. Outgroup Prosocial Behaviors (T2)m	1.4	1.3	0.11*	-	0.01	0.22	0.15*	0.47*	0.55*	0.32*
	0	5		0.06			*	**	**	**

Note:

^*p < .05.

^**p < .01.

^***p < .001.

^c= child report,

^m= mother report.

Significant correlations among indicator variables for overall prosocial behavior supported the creation of a latent variable. Little and colleagues (2002) review how parcels represent an aggregate-level indicator of two or more items, which increases the reliability and representativeness of the scale, as well as the stability of the factor solution (Little, Cunningham, Shahar, & Widaman, 2002). Compared to individual indicators, analyses based on parcels are also more efficient, more normally distributed and have smaller intervals between points on the scale. Using Little and colleagues’ (2002) technique for building parcels, a just-identified latent variable (i.e., with three parcels) model was selected because it represents a unique solution that optimally captures the relations among the items. The items-to-construct balance approach used the loadings of the seven item indicators as a guide to form parcels. The factor loadings were ranked from highest to lowest; the three strongest items were assigned to “anchor” each of the three parcels. The next highest three were added in inverted order, and the final of the seven items was added to the first parcel. A different number of items for each parcel is acceptable to achieve a reasonable balance (Little et al., 2002). The measurement model for the latent constructs was tested by running a confirmatory factor analysis for prosocial behaviors at Time 1 and 2; parcels were forced to load on their given factor and the prosocial latent variables at Time 1 and 2 were allowed to correlate yielding a model with acceptable fit (χ²(8) = 32.31, p < .05, N = 714; χ²/df = 4.04; TLI = .98; CFI = .99; RMSEA = .07 (CI : .04, .09)).

The main effects of the two risk factors, intergroup and intragroup threat at Time 1, on overall and outgroup prosocial behaviors at Time 2 were examined, controlling for child gender and age at Time 1, as well as the auto-regressive controls for each of the outcomes of interest and the main effect of outgroup attitudes at Time 1. The interaction terms were created with the centered values of each type of threat and outgroup attitudes, then added to the model as exogenous predictors. To examine the possibility of reciprocal effects, a full cross-lagged model was tested; the relations among overall and outgroup prosocial behaviors at Time 1 and intergroup and intragroup threat and outgroup attitudes at Time 2 were also included. All exogenous variables and control variables were allowed to correlate with each other; error terms for the endogenous outcomes of interest were also allowed to correlate.

Figure 1 depicts the full cross-lag model, including both direct and moderation effects, that was tested; the fit for the overall model was good (χ²(75) = 244.95, p < .05, N = 714; χ²/df = 3.266; TLI = .93; CFI = .97; RMSEA = .056 (CI: .049, .064)). Structural path estimates for the full model are in Table 2; for clarity, Figure 2 omits non-significant paths and stability estimates to present the significant paths of interest along with the standardized parameters. The control variable of child gender was significant; compared to males, female youth had higher overall (β = .11, p < .01) and outgroup (β = .11, p < .01) prosocial behaviors. Age at Time 1 did not predict either type of prosocial behaviors. The autoregressive stability paths from Time 1 to Time 2 were also significant for each outcome of interest including overall prosocial (β = .58, p < .001), outgroup prosocial behaviors (β = .25, p < .001), intragroup threat (β = .44, p < .001), intergroup threat (β = .61, p < .001), and outgroup attitudes (β = .57, p < .001). As a control, the main effect of outgroup attitudes at Time 1 did predict outgroup helping at Time 2 (β = .14, p < .01), but was not related to overall prosocial acts a year later (β = −.01, ns).

Figure 1.

Complete Cross-lag Model, Including Both Direct and Moderation Effects, to Predict Overall and Out-group Prosocial Behaviors. All Variables are Child Report, Unless Denoted with _m = Mother Report. Grey Dotted Lines Represent Non-significant Paths, Grey Full Lines Represent Trends, and Black Full Lines Represent Significant Paths. Exogenous Variables and Endogenous Outcomes Were Correlated, Respectively; Correlation Coefficients and Error Variances Were Omitted from the Model for Readability._tp < .10, *p < .05, **p < .01, ***p < .001. Model Fit: χ²(75) = 244.95, p < .05, N = 714; TLI = .93; CFI = .97; RMSEA = .056 (CI: .049, .064).

Table 2. Unstandardized, Standardized, and Significance Levels for Model in Figure 1 (Standard Errors in Parentheses; N = 714).

Parameter Estimate	Unstandardized	Standardized	p
Structural model estimate
Overall Prosocial (T1) → Overall Prosocial (T2)	.54 (.04)	0.59	<.001
Intergroup Threat (T1) → Overall Prosocial (T2)	.001 (.004)	0.007	0.88
Intragroup Threat (T1) → Overall Prosocial (T2)	.01 (.004)	0.13	0.004
Outgroup Attitudes (T1) → Overall Prosocial (T2)	.01 (.004)	0.13	0.004
Female → Overall Prosocial (T2)	.11 (.04)	0.11	0.004
Age (T1) → Overall Prosocial (T2)	−.01 (.01)	−.03	0.42
Outgroup Prosocial (T1) → Outgroup Prosocial (T2)	.36 (.06)	0.26	<.001
Intergroup Threat (T1) → Outgroup Prosocial (T2)	−.02 (.01)	−.09	0.09
Intragroup Threat (T1) → Outgroup Prosocial (T2)	.04 (.01)	0.3	<.001
Outgroup Attitudes (T1) → Outgroup Prosocial (T2)	.01 (.002)	0.17	<.001
Female → Outgroup Prosocial (T2)	.31 (.11)	0.11	0.006
Age (T1) → Outgroup Prosocial (T2)	−.03 (.03)	−.05	0.29
Ougroup Attd X Intragroup Threat   →Outgroup Prosocial (T2)	.001 (.001)	0.11	0.01
Ougroup Attd X Intergroup Threat   → Outgroup Prosocial (T2)	.001 (.001)	0.04	0.42
Intergroup Threat (T1) → Intergroup Threat (T2)	.61 (.04)	0.61	<.001
Intragroup Threat (T1) → Intergroup Threat (T2)	−.05 (.06)	−.04	0.36
Outgroup Attitudes (T1) →Intergroup Threat (T2)	−.001 (.01)	−.01	0.91
Female → Intergroup Threat (T2)	−.27 (.51)	−.02	0.60
Age (T1) → Intergroup Threat (T2)	−.21 (.13)	−.06	0.09
Intergroup Threat (T1) → Intragroup Threat (T2)	.07 (.03)	0.11	0.014
Intragroup Threat (T1) → Intragroup Threat (T2)	.44 (.04)	0.54	<.001
Outgroup Attitudes (T1) → Intragroup Threat (T2)	.004 (.008)	0.02	0.61
Female → Intragroup Threat (T2)	−.57 (.37)	−.06	0.12
Age (T1) → Intragroup Threat (T2)	.35 (.09)	0.15	<.001
Overall Prosocial (T1) → Outgroup Attitudes (T2)	2.28 (1.76)	−.06	0.11
Outgroup Prosocial (T1) → Outgroup Attitudes (T2)	3.19 (1.00)	0.12	0.001
Intergroup Threat (T1) → Outgroup Attitudes (T2)	−.13 (.16)	−.03	0.42
Intragroup Threat (T1) → Outgroup Attitudes (T2)	.39 (.20)	0.08	0.06
Outgroup Attitudes (T1) → Outgroup Attitudes (T2)	.57 (.04)	0.56	<.001
Female → Outgroup Attitudes (T2)	5.37 (1.86)	0.10	0.004
Age (T1) → Outgroup Attitudes (T2)	.84 (.47)	0.06	0.07

Note: χ²(75) = 244.95, p < .05, N = 714; TLI = .93; CFI = .97; RMSEA = .056 (CI: .049, .064); Measurement model parameters available upon request from first author.

Figure 2.

Simplified Cross-lag Model, Including All Significant Direct and Moderation Effects of Interest, Omitting Stability Parameters. As with Figure 1, All Variables Are Child Report, Unless Denoted with _m = Mother Report. Grey Full Lines Represent Trends, and Black Full Lines Represent Significant Paths. Exogenous Variables and Endogenous Outcomes Were Correlated, Respectively; Correlation Coefficients and Error Variances Were Omitted from the Model for Readability._tp < .10, *p < .05, **p < .01, ***p < .001. Model Fit: χ²(75) = 244.95, p < .05, N = 714; TLI = .93; CFI = .97; RMSEA = .056 (CI: .049, .064).

Experience with sectarian, or intergroup, antisocial behavior was related to less outgroup prosocial behavior at the trend level (β = −.09, p < .10), and was not significantly related to overall prosocial acts a year later (β = −.01, ns). Outgroup attitudes did not moderate the pathway between sectarian threat and outgroup helping. On the other hand, experience with nonsectarian, intragroup antisocial behavior was positively and significantly related to both overall prosocial behaviors (β = .16, p < .001) and outgroup helping (β = .36, p < .001). Moreover, more positive outgroup attitudes strengthened the link between intragroup threat and outgroup prosocial behaviors (β = .14, p < .01). This moderation effect is displayed in Figure 3; high and low outgroup attitudes are graphed at one standard deviation above and below the mean.

Figure 3.

Interaction of outgroup attitudes and intragroup threat, or nonsectarian antisocial behavior, on outgroup prosocial behaviors. The positive relationship between intragroup threat and outgroup prosocial acts is strengthened for those who have more positive attitudes toward the outgroup. Higher/lower outgroup attitudes depicted at ± 1 SD.

Finally, the cross-lagged relations between overall and outgroup prosocial behaviors at Time 1 and the other variables of interest at Time 2 were estimated. Greater mother report of outgroup prosocial acts related to higher youth report of outgroup attitudes one year later (β = .12, p < .001), but did not significantly relate to greater intragroup threat over time. Taken together, these findings seem to suggest a positive feedback loop between outgroup attitudes and behaviors. Moreover, the experience of intragroup threat appears to be a motivating factor to help outgroup members, without carrying the risk of retaliation as measured in the current study. In addition, the experience of intragroup threat at Time 1 is related to more positive outgroup attitudes at Time 2 at the trend level (β = .36, p < .10). Modeling these alternative directional pathways helps to understand the order of effects of intragroup threat on outgroup attitudes and prosocial behaviors for youth in a post-accord setting.

Discussion

The goal of the current study was to highlight positive youth outcomes that may have implications for intergroup relations. Prosocial behaviors were the outcome of interest because of the potential to foster intergroup cooperation if helping acts are extended to outgroup members. Outgroup attitudes were also considered as a possible moderator of the link between risk and behaviors one year later for adolescents living in Belfast, a setting of protracted conflict. Using two time points, and controlling for other relevant demographic variables, the analyses found that experience of intergroup antisocial behavior showed a trend toward decreasing outgroup helping, whereas intragroup antisocial behavior was related to greater overall and outgroup prosocial acts. The positive link between intragroup threat and outgroup helping was even stronger for those youth who had more positive outgroup attitudes. Moreover, both the experience of intragroup threat and outgroup helping were linked to more positive outgroup attitudes a year later. This set of findings makes an important contribution to the literature by differentiating between types of risk (intergroup vs. intragroup), as well as specifying the target of prosocial acts (overall vs. outgroup). Examining these reciprocal effects over time for adolescents in a real-world setting of protracted conflict, the results suggest that programs which foster more positive outgroup attitudes and provide opportunity for outgroup helping may have promising implications for more constructive intergroup relations in Belfast.

Focusing on the type of threat and timing of measurement, this paper helps to explain conflicting results surrounding prosocial outcomes in previous research on youth risk and resilience processes in settings of political violence. Altruism born of suffering may explain the positive link between intragroup threat and outgroup helping; although future research is needed to test this variable as a mediating process. According to this theory, pain caused by intentional human harm may facilitate individuals to see new perspectives, including greater attention to overall injustice and perceiving a common fate with other victims. Re-conceptualizing past rivals as fellow sufferers, a shared “victim” identity links individuals across group memberships. With this shift, empathy and compassion for others may evoke greater helping behaviors.

Inclusive victim consciousness, or recognizing the shared aspects of suffering, may also affect outgroup attitudes. In this study, experience of intragroup threat led to more positive outgroup attitudes a year later. Moreover, positive intergroup attitudes may increase the possibility of approach responses which can facilitate prosocial behaviors (Butz & Plant, 2011). The moderating role of positive outgroup attitudes strengthened the relation between intragroup threat and greater outgroup prosocial acts among adolescents in Belfast. At the same time the reciprocal effect between these two variables was not significant; that is, helping the outgroup at Time 1 did not significantly increase the experience of intragroup threat at Time 2. This set of findings suggests that a positive feedback loop may be in effect; more positive behaviors predict more positive attitudes, while at the same more positive attitudes lead to more positive behaviors. These relations appear to hold without being linked to greater experience with intragroup antisocial behavior for youth in Belfast.

Regarding the timing of measurement, the current study was conducted in a post-accord context. In this setting, the increase in nonsectarian or intragroup threat may motivate individuals to consider new possibilities not imagined during the Troubles. Bar-Tal and Halperin (2009) label this an “unfreezing process” in which groups are presented with inconsistent or conflicting information that “creates some kind of tension, dissonance, dilemma, or internal conflict, thereby stimulating people to move away from their previous stance and search for new information” (p. 434; emphasis added). The emergence of intragroup antisocial behavior may function as this type of motivating factor, eroding perceived interdependence among ingroup members (Flippen et al., 1995), and encouraging individuals to consider alternative attitudes or behaviors. Thus, “collective prosocial behavior benefitting outgroup members may be particularly likely after suffering caused by members of the social ingroup” (Vollhardt, 2009; p. 79). With an eye toward the future, youth may be more open to the possibility of unfreezing the intergroup patterns that defined past sectarian conflict (Wallach, 2000).

By distinguishing between overall and outgroup prosocial acts, this paper attempts to challenge the portrayal of youth as perpetrators or helpless victims in times of conflict, and to recognize their agency in rebuilding relationships across group lines. Even in the face of continued intergroup violence, the findings provide insight on ways to promote intergroup cooperation through outgroup helping among adolescents exposed to intragroup violence, particularly by promoting more positive outgroup attitudes and providing opportunities for intergroup cooperation. Promoting dual identity (i.e., attachment to a subgroup and a shared common group) through extended contact has been shown to improve children’s attitudes toward outgroup members (Cameron, Rutland, Brown, & Douch, 2006). In Northern Ireland, the majority of young people who had participated in cross-community projects reported more favorable outgroup attitudes (Schubotz & Robinson, 2006). Future translational research should develop empirically-informed programs that foster intergroup cooperation, promote prosocial behaviors, and foster positive outgroup attitudes, which in turn may help to consolidate peace across group lines and in local communities.

Future research should also consider other constructive intergroup attitudes and behaviors, such as trust and forgiveness, and how these relate to attitudes toward peace for youth in Northern Ireland. Promoting positive outgroup attitudes may be one avenue to promote outgroup helping, but deepening trust may also improve intergroup behaviors over and above liking members of the other group (Tam, Hewstone, Kenworthy, & Cairns, 2009). Intergroup forgiveness, a key component of long-term peacebuilding, has also been related to empathy following intergroup contact (Tam et al., 2008). Trust and forgiveness among youth may further intergroup cooperation over time in situations of protracted conflict.

Two of the strengths of the current study include the inclusion of multiple reports to reduce method variance and the use of two time-points of data which allow for examination of change over time. Both of these elements increase the validity and reliability of the findings. Of course, additional changes could improve the study. First, peer or teacher reports of youth prosocial behaviors could reduce social desirability that may have influenced mothers’ reports of the prosocial behaviors of their children. Peer reports may also reveal important social norms that could promote overall and/or outgroup prosocial behsavior (Chung-Hall & Chen, 2010). Second, future research could investigate a more multidimensional construct of positive attitudes, such as affection, social ease, social closeness, and learning inspiration (Pittinsky, Rosenthal & Montoya, 2011). Third, outgroup helping may not always be prosocial; some individuals may help in order to maintain social dominance by reaffirming a high status position (Nadler, Harpaz-Gorodeisky & Ben-David, 2009). The scales used in the current study, however, do not permit parsing out specific intentions from the prosocial behaviors themselves. Fourth, the intragroup threat scale, designed in response to mother’s preoccupations about the safety and security of their families in Belfast (Taylor et al., 2011), largely included conventional forms of crime and antisocial behavior committed by ingroup members. Future research should extend this more common or daily threat to include threats related to paramilitary punishment attacks or threats as well. Finally, although to more fully support the role of altruism born of suffering as an explanatory theory, longitudinal meditational tests of this construct are needed.

Addressing protracted conflicts necessitates a generational approach to peacebuilding (Lederach, 1997). This paper adopts a “decades thinking” approach, which emphasizes prosocial psychosocial outcomes among youth in post-accord societies. Youth may be mobilized in the escalation of violent conflict, but their constructive contributions are often overlooked (McEvoy-Levy, 2006). Complementing previous research which aims to identify factors that constrain violent intergroup conflict, the goal of the current paper was to focus on aspects of the social ecology that can contribute to positive social reconstruction in conflict-affected societies. The peacebuilding potential of youth may generalize to other post-accord contexts in which levels of intragroup violence and common crime are on the rise (van der Merwe & Dawes, 2007). This paper shifts the focus from youth as troublemakers to youth as peacemakers (McEvoy-Levy, 2006).