Abstract
This study examined cultural specificity in how interpretations about peer provocation are associated with revenge goals and aggression. The sample consisted of young adolescents from the United States (369 seventh graders; 54.7% male; 77.2% identified as White) and from Pakistan (358 seventh graders; 39.2% male). Participants rated their interpretations and revenge goals in response to six peer provocation vignettes and completed peer nominations of aggressive behavior. Multigroup structural equation models (SEMs) indicated cultural specificity in how interpretations were related to revenge goals. Interpretations that a friendship with the provocateur was unlikely were uniquely related to revenge goals for Pakistani adolescents. For U.S adolescents, positive interpretations were negatively related to revenge but self-blame interpretations were positively related to vengeance goals. Revenge goals were related to aggression similarly across groups.
Responding to peer provocation is a particularly challenging social task for children and adolescents (Asher et al., 2001, 2008). Although some youth may disengage or react assertively after peer provocation, others may seek revenge and retaliate for perceived harm (Dirks et al., 2017; McDonald & Asher, 2018). However, seeking revenge is not associated with social success; revenge goals and retaliation in response to unambiguous peer provocation have been associated with aggressive behavior, victimization, and peer dislike (Dirks et al., 2017; McDonald & Asher, 2018). Thus, it is important to understand what cognitions may contribute to desires for revenge after peer provocations.
There are many social, cognitive, and contextual factors that likely determine how young adolescents respond when provoked by peers. Although social information processing (SIP) models (Crick & Dodge, 1994; Lemerise & Arsenio, 2000) outline several important steps in social-cognitive and affective processing that are relevant for responding to peer provocations, few studies have examined components of this model cross-culturally. In two exceptions, hostile attributions were found to be similarly related to future aggressive behavior across nine countries and aggressive strategy generation was similarly related to later externalizing problems in eight countries (Dodge et al., 2015; Lansford et al., 2017). Expanding on this work, the current study examines how a wider variety of interpretations about peer aggression are related to revenge motivations in a cross-cultural comparison of youth from the United States and Pakistan. Interpretations of peer provocations and revenge motivations have not been examined in cross-cultural examinations of SIP, even though there is cultural variation in perceptions of aggressive acts and in norms surrounding revenge and retaliation (Leung & Cohen, 2011; Severance et al., 2013).
Interpretations and Revenge Goals
Social goals are the precursors and motivations behind specific behaviors, such as aggression. Revenge goals have been operationally defined as goals to “get even,” “get back,” or “hurt” a person who is perceived as responsible for an offense (e.g., McDonald et al., 2021). Revenge goals are particularly important to study as a response to provocation, not only because they are associated with poorer social functioning but also because retaliation in response to provocation may prolong the cycle of aggression (Kochenderfer-Ladd, 2004; Shelley & Craig, 2010).
There may be many interpretations, beyond attributions of negative intent, which are relevant in understanding revenge motivations. When college students responded to vignettes involving fairly minor conflicts of interest, interpretations of being disrespected or rejected, as well as interpretations of a relationship being unviable, were correlated with revenge goals (McDonald & Asher, 2013). On the contrary, interpreting that the particular conflict was resolvable was negatively correlated with revenge goals (McDonald & Asher, 2013). In a subsequent study with young adolescents, two groups of youth who made negative interpretations of the provocation (e.g., interpretations of rejection, disrespect, wrongdoing) at high levels but differed on whether they desired revenge were identified and compared (McDonald & Asher, 2018). Youth who refrained from revenge-seeking, who they termed “pacifists,” endorsed positive interpretations (e.g., interpretations of unintended emotional pain, resolvability, and relationship viability) more than revenge-seekers.
Altogether, evidence is accumulating that there are other interpretations relevant to predicting youth revenge goals in response to peer provocations. The current study sought to examine several interpretations that may promote or inhibit revenge goals after peer provocation. As revenge-seekers and pacifists were equally likely to make negative interpretations of a provocateur’s behavior (McDonald & Asher, 2018), we hypothesized that negative interpretations would not be uniquely related to revenge goals, but instead that positive interpretations of a provocation would be directly and negatively related to revenge.
Variability Across Cultures
Another primary goal of the study was to consider cultural variation in how interpretations and revenge goals are related. As cultural norms and values influence scripts for interpersonal interactions (e.g., French et al., 2005; Hofstede, 2001; Martínez-Lozano et al., 2011), individuals may interpret the meaning of provocations differently across cultures and these interpretations may be differentially predictive of revenge goals (Severance et al., 2013).
We examined the similarities and differences in how interpretations of provocation may be related to revenge goals in samples of youth from the United States and Pakistan. Pakistan is considered more of a collectivistic culture than the United States (Hofstede Insights, 2019). In Pakistan, relationships are prioritized over individual needs and it is considered one’s “duty” to maintain interconnectedness (Zaman, 2014). More than 95% of people living in Pakistan are Muslim (Pew-Templeton Global Religious Futures Project, 2020). Religious injunctions are used in child rearing practices to instill the importance of relationships and sacrifice to maintain familial and community ties. Even with rapid urbanization and changes in observed family structure, the value of maintaining ties to the extended family and community is still given great importance (Zaman, 2014). Furthermore, in Punjab, the biggest province of Pakistan, the “Biradari” (brotherhood) forms the strongest ties and often conflicts are resolved in community settings (“Panchayat”), where the elders decide the disputes between members of communities (Chaudhry & Ahmed, 2014). Even in urban settings, where individualistic values are often adhered to in the domain of agency and personal control, conflicts are still resolved keeping in mind the future outcomes of one’s response for relationships and the community.
Honor (or izzat) is also important in Pakistan and retaliation may be expected in response to degradations of honor. However, honor norms may be more prevalent in rural areas and younger people may not feel as tied to upholding these norms (Pier, n.d.). Thus, we did not have direct hypotheses about whether Pakistani adolescents would endorse revenge goals more than U.S. adolescents after peer provocation. However, Pakistani people also emphasize maintaining harmonious interactions with in-group members and, as such, we theorized that Pakistani adolescents would be less likely to respond negatively to provocations if they anticipated future interactions with the provocateur (Chen et al., 2009; Jung et al., 2014). To examine this question, we compared how interpretations about the viability of a friendship, which would imply future and consistent future interactions with a provocateur, would be a greater determinant of revenge-seeking for Pakistani and U.S. youth.
Method
Participants
In the United States, letters were sent home to 378 seventh-grade students to participate in a study about peer relationships. Of these students, 373 (98.67%) received parental consent and assented to participate in the study. Four other students were removed from analyses due to incomplete data and absences on data collection days. Thus, the U.S. sample consisted of 369 seventh graders (54.7% male; 77.2% White, non-Hispanic; 12.7% White Hispanic; 3.8% African American; 1.9% Asian American; and 4.3% more than one racial/ethnic group) from a suburban, mixed-sex, public middle school in the Midwest, recruited as part of a larger project about how youth respond to peer provocations (McDonald & Asher, 2018). School-level statistics indicated that 8.31% of the students at the school received free or reduced-price lunch. The percent of students in the state at the time who received free or reduced-price lunch was 39.8% to 46.1% (U.S. Department of Education, National Center for Education Statistics, Common Core of Data, 2012); thus youth in this school were primarily middle- or upper-middle class for the area in which they resided.
In Pakistan, letters were sent home to parents of 376 seventh graders asking for parental consent for their children to participate in a study of peer relationships. Students were recruited from eight private schools on the outskirts of a large city in Punjab. Of these eight schools, four schools were single-sex, two schools were completely co-educational, and two schools had some classrooms that were single-sex and some that were co-educational. From these schools, students came from 13 classrooms (three were co-educational classrooms, six were all female classrooms, and four were all male classrooms). Private school attendance is common in Pakistan; 42% of students in Pakistan attend private schools and 89% of private schools are considered extremely low in cost (Qureshi & Razzaque, 2021). Of those recruited, 358 (95.21%) received consent and assented to participate. Participants were 39.2% male and primarily Muslim. Participants indicated that 48.7% had a mother with a college degree or higher and 59.8% had a father with college degree or higher. These rates exceed what is typical, as only 9% of adults in Pakistan are enrolled in tertiary education programs (Hunter, 2020). Thus, Pakistani families were primarily middle to upper class for the area in which they resided.
Procedure
The study was approved by the Duke University (IRB # 2100) and University of Alabama (IRB #11-002) institutional review boards in the United States and by a government and local school authority ethics review board in Pakistan. In the United States, consent letters were sent by mail to the parents of seventh graders. Following procedures used in previous research (e.g., Ebesutani et al., 2012; Rose & Asher, 1999) and consistent with Federal Regulation 45.46.116 and the Duke University Institutional Review Board, parents were informed about the study and were asked to contact the researchers or the school if they had any questions or did not want their child to participate. In Pakistan, letters were also sent home to parents with children asking for permission for their child to participate. Parents returned forms with their signatures indicating their permission for their children to participate. In both Pakistan and the United States, children were also asked for assent on each data collection day. In both countries, participants completed measures during two, 40-min classroom sessions.
Measures
Participants were presented with six hypothetical situations depicting unambiguous peer provocation that involved exclusion and criticism (see Table 1 in Supplementary Materials) and were instructed to imagine that each situation really happened to them. All provocation situations were designed to be ecologically valid, depicting realistic situations that youth could encounter in their daily school lives. Collaborators in Pakistan agreed that these situations were ecologically valid in Pakistan.
During group testing sessions, researchers read the vignettes aloud while participants followed along. In response to every vignette, participants rated the likelihood that they would experience certain emotions, make certain interpretations, enact certain behavioral strategies, and pursue a variety of goals on a scale from 1 (strongly disagree) to 10 (strongly agree; see McDonald & Asher, 2018, for more information on the measure). For this study, we only examined interpretations and revenge goals. To assess interpretations, participants were asked, “What would you be thinking in this situation?” Item wording and internal reliabilities for each interpretation averaged across the six vignettes are presented in Table 2 of the Supplementary Materials. For the U.S. sample, internal reliabilities for each interpretation across the vignettes were above .70. For the Pakistan sample, internal reliabilities ranged from .60 to .74.
To assess goals, participants were asked, “What would your goals be in this situation?” Specific to revenge goals, participants were asked to rate two relevant statements: “I would be trying to get back at this person” and “I would be trying to hurt this person like they hurt me.” The internal reliability for these statements across all six vignettes was high (U.S.: α = .91; Pakistan: α = .89).
Aggressive Behavior
Participants completed an unlimited peer nomination measure in which they circled the names of peers on a roster who fit particular behavioral descriptors. Participants were asked to nominate peers who fit descriptions for three physical/verbal aggression items (“calls other kids mean names”; “starts fights”; “hits and pushes”) and two relational/social aggression items (“gossips”; “tries to keep certain kids from being in their group”). In the United States, students were in teams ranging from 61 to 171 students. U.S. nomination lists were randomized so that every participant was on 61 other rosters, 61 names of other students were listed for each item, and lists of names were different for each item. Participants were given the option of crossing out names of peers they did not know. Crossed-off names were removed from the denominator when calculating proportion scores. For the Pakistan sample, since students stayed with the same children all day there was no need to provide children with the option of crossing off the names of students they did not know. Classes in Pakistan ranged from 19 to 41 students. Each child’s score for each item was computed as the proportion of nominations that the child received divided by the total number of children who provided nominations. In both countries, proportion scores then were averaged to create the final aggression score used in analyses (U.S.: α = .93; Pakistan: α = .94).
Results
Plan of Analysis
First, a multiple-group confirmatory factor analysis (MGCFA) was conducted to assure that the structure of the interpretations was equivalent for U.S. and Pakistani youth. Next, preliminary analyses were conducted to examine correlations among variables of interest and mean-level country differences. Finally, to address the main study hypotheses examining how interpretations were related to revenge goals and aggression across cultures, a multigroup structural equation model (SEM) analysis was conducted. We started with a model in which each interpretation factor predicted revenge goals and revenge goals predicted aggression for both groups.
For both the MGCFA and the multigroup SEM, we used Mplus v. 8 (Muthén & Muthén, 1998–2012) and employed a robust maximum likelihood estimation (MLR) that accounts for non-multivariate normality with continuous variables (Muthén & Muthén, 1998–2012). We tested equality constraints for each path/loading one-by-one and we used ΔCFI to evaluate measurement invariance. As suggested by Cheung and Rensvold (2002), when ΔCFI < .01 the more restrictive model is justified. The goodness-of-fit indices such as the comparative fit index (CFI), the root mean square error of approximation (RMSEA), and standardized root mean squared residual (SRMR) (Hu & Bentler, 1999) were also used as additional indicators of fit for the current study. When CFI ⩾ .95, SRMR ⩽ .08, and RMSEA ⩽ .06 (Hu & Bentler, 1999), the model was deemed to have acceptable fit.
Structure of Interpretations
An MGCFA was conducted to examine how the structure of interpretations was similar or different by gender and cultural group (see Table 4 and Figures 1 and 2 of the Supplementary Materials). Negative interpretations (e.g., disrespect, rejection, wrongdoing, and disassociation) were grouped together because they included negative and rejecting interpretations of the provocateur’s behavior, whereas positive interpretations (e.g., resolvable, unintended emotional pain, minimizing) were grouped together because they pertained to more benign attributions of the provocateur’s behavior or more optimistic interpretations about the conflict. Self-blame interpretations were separated from negative and positive interpretations because this attribution placed the blame for the provocateur’s action on the self. Relationship unviable interpretations were kept separate to increase model fit and because these attributions were about the future of a friendship and did not pertain the singular actions of either provocateur or target. Analyses indicated that the model comparing boys and girls and the model comparing Pakistani and U.S. adolescents met criteria for partial scalar invariance.
Preliminary Analyses
Descriptive statistics, correlations among variables, and gender differences can be found in Table 3 of the Supplementary Materials. Initial t tests comparing interpretations, revenge goals, and aggression indicated several differences. Pakistani adolescents endorsed wrongdoing, relationship unviability, resolvability, unintended emotional pain, and self-blame interpretations more than U.S. youth, whereas U.S. youth made more minimizing interpretations more than Pakistani youth. Pakistani adolescents also endorsed revenge goals more and reported peers as more aggressive than U.S. youth.
Interpretations Predicting Revenge Goals and Aggression
Next, we conducted a multigroup SEM model to examine how negative, positive, self-blame, and relationship unviability inter-pretations were related to revenge goals and aggression (see Figure 1). Preliminary analyses also explored if interpretations were directly related to aggression but did not find these paths to be significant.

Interpretations About Provocation Predict Revenge Goals and Revenge Goals Predict Aggression. Standardized coefficients are reported with those for the U.S. sample presented first and those for Pakistan presented second. Coefficients in bold indicate that the coefficients significantly differed across countries.
The initial model with all interpretations predicting revenge goals was acceptable; however, for both groups negative interpretations were not significantly predictive of revenge goals, so this path was removed from our model. Next, we constrained each path across countries to see whether there was significant change in model fit. The only path that could be constrained across groups without sacrificing model fit was that between revenge goals and aggression. The final model allowed all the other paths from interpretations and revenge goals to vary by cultural group. The final model fit well, χ2(82) = 201.15, CFI = .96, RMSEA = .06, SRMR = .06.
As can be seen in Figure 1, self-blame interpretations were positively related to revenge goals for U.S. adolescents but were not for Pakistani adolescents. Relationship unviability interpretations were significantly predictive of revenge goals for Pakistani youth but not for U.S. youth. Finally, positive interpretations were negatively predictive of revenge goals for U.S. adolescents but were not for Pakistani adolescents. For both groups of adolescents, revenge goal endorsement was similarly related to peer-nominated aggression.
Discussion
The aims of the current study were to examine how adolescents’ interpretations after peer provocation were predictive of revenge goals and whether these relations were similar or different for adolescents from the United States and Pakistan. It was hypothesized that positive interpretations would be more predictive of revenge goals than negative interpretations for both groups of adolescents. As prior research suggests that individuals from countries whose norms focus on in-group harmonious interaction are less likely to retaliate if they expect that they will interact with the provocateur in the future (Chen et al., 2009; Jung et al., 2014), we also hypothesized that relationship unviability interpretations would be more related to revenge goals for Pakistani youth than for U.S. adolescents. In addition, because there may be greater expectations for retaliation under conditions of honor threat in cultures that prioritize honor (Leung & Cohen, 2011), we also explored whether Pakistani adolescents would endorse revenge more for peer provocations than U.S. youth.
Results partially supported our hypotheses. As expected, positive interpretations of the provocation were negatively related to revenge goals for U.S. adolescents. As others have found (Dodge & Somberg, 1987; Williams et al., 2003), interpretations that a conflict is resolvable, not intended, or not important are negatively associated with retaliatory motives. In addition, negative interpretations were not uniquely related to revenge goals for either U.S. or Pakistani adolescents suggesting other intervening cognitive or emotional processes are important when considering how negative interpretations of peer provocation predict revenge and later aggressive responses.
For Pakistani adolescents, relationship unviability interpretations were the only interpretation type that was uniquely related to revenge desires when controlling for the other interpretations. Thus, and similarly to what was found by others (e.g., Jung et al., 2014), the anticipation of a future relationship with the provocateur inhibited revenge responding for Pakistani youth. The idea of being “friends” with the provocateur was more inhibiting of revenge responding in Pakistan than in the United States.
Finally, comparisons indicated that Pakistani adolescents’ rated revenge goals more highly than U.S. youth. This could align with idea that retaliation is more acceptable to threats to honor in Pakistan (Leung & Cohen, 2011). However, there were no group differences on how disrespected U.S. and Pakistani adolescents reported they felt, indicating that there may not be differences in how much they perceived their honor to be threatened. Also, there is heterogeneity in the endorsement of honor norms within Pakistan (Pier, n.d.); therefore, future research should consider measuring individual endorsement of honor norms. Another reason for group differences in revenge goals could be that Pakistani youth perceived these provocations to be more threatening to a potential friendship, as Pakistani youth endorsed interpretations of relationship unviability more than U.S. youth. Differences in this interpretation could be due to the relative importance of friendships in these cultures (Lu et al., 2021) or it could be that the emphasis on maintaining positive in-group relationships may also make these types of provocation more threatening to relationships in Pakistan. Pakistani youth may be more likely to question their friendship when they are excluded and insulted by a peer.
Of note, self-blame was differentially related to interpretations and revenge goals across our samples, suggesting that self-blame may a focus for future cross-cultural study (cf. Yang et al., 2022). For instance, self-blame was directly and positively related to revenge goals for U.S. adolescents. In past research, self-blame for victimization has been linked to internalizing problems (Perren et al., 2013) but seems to be less strongly linked or unrelated to externalizing problems (Mumtaz et al., 2017; Schacter et al., 2015). Past studies typically differentiate between characterological self-blame (i.e., self-blame based on characteristics that are stable and uncontrollable) and behavioral self-blame (i.e., self-blame attributed to controllable unstable behaviors). Although we did not differentiate between these forms, it may be that these different types of self-blame would be differentially associated with revenge goals. Self-blame, especially characterological self-blame (Tilghman-Osborne et al., 2008), may manifest as feelings of shame, a self-conscious emotion associated with feeling defective (Tangney, 1998) and shame may threaten one’s sense of self and lead to aggression (Baumeister & Bushman, 2003; Tangney et al., 1996). Thus, it could be that self-blame that manifests as shame could lead to revenge desires for the American adolescents.
Self-blame was not related to revenge goals for Pakistani adolescents, suggesting that the associations of self-blame, shame, and anger may differ across cultures. Shame has been linked with anger in American youth, but was unrelated to anger in Japanese youth (Bear et al., 2009). However, others find that although there are differences across cultural groups in the tendency to feel shame, shame is similarly related to aggression-related constructs for children in Japan, Korea, and the United States (Furukawa et al., 2012). Thus, the cultural similarities and differences in the links between shame, self-blame, and aggression remain unclear. Future studies should consider how shame, guilt, anger, and self-blame for peer provocation are interrelated in youth from a variety of cultures.
In addition, it may have been logical to expect that the interpretation that a friendship is unviable (i.e., “I could never be friends with this person”) would be highly (and negatively) related to how easily resolvable the issue was perceived to be (i.e., “This person and I can easily make-up”); however, we observed that unviable and resolvable interpretations were only moderately correlated and a measurement model that kept unviable interpretations separate from resolvable interpretations better fit the data (see Table 3 in Supplementary Materials). Although friends are more likely to report conflict resolution compared to non-friends (Newcomb & Bagwell, 1995), youth often resolve conflicts with acquaintances or non-friends via negotiation rather than through coercion and disengagement (Laursen et al., 2001), suggesting that positive conflict resolution strategies do not only occur in the friendship context. Furthermore, a small portion of youth seek revenge in disagreements, even with friends (Rose & Asher, 1999). Thus, interpretations about friendship viability and easy conflict resolution may not consistently co-occur. However, as forgiveness and conflict resolution have been suggested to play an integral role in friendship formation and maintenance (Asher et al., 1996), future investigations should continue to consider how interpretations of conflict resolution are related to interpretations about relationship viability under different situational constraints and in a variety of relationships and cultures.
There are some limitations of the current study that should be considered. First, the reliance on cross-sectional data limits conclusions about the direction of effects. Although the relations of interpretations to revenge goals and revenge goals with aggression support SIP theories, it may also be that aggressive youth have scripts that prioritize revenge responding and these revenge goals affect how youth interpret social situations. Future experimental research could attempt to manipulate interpretations to see whether revenge goals are subsequently affected. Longitudinal research examining within and between person change in interpretations, revenge goals, and aggression could also better address the direction of effects. Second, some of the interpretation subscales were lower on internal consistency (α = .60), especially within the Pakistan sample. This may be because the study was initially designed for a Western sample. It should also be noted that when using vignette-based measurement, there may be more variation in ratings due to the situation specificity and this may lower the internal reliabilities of the measure. These lowered reliabilities may have limited our ability to find significant relations in the Pakistan sample. Third, future research should expand comparisons of these processes to a greater variety of cultures as well as consider individual differences in cultural beliefs or norm endorsement as a predictor of these social-cognitive and interpersonal processes, instead of relying on country of origin alone as a predictor of differences (Leung & Cohen, 2011).
Finally, the adolescents recruited for this study are unlikely to be representative of all adolescents from their countries of origin, nor can we say that the samples are equivalent on all variables except country of origin. For instance, the Midwestern U.S. sample may not be representative of youth from other areas of the country or of urban adolescents in the United States, nor may the Pakistani sample represent youth from rural areas of Pakistan. Thus, we urge caution when making generalizations of our findings to the larger cultural contexts.
For far too long, the study of social information processing has neglected to consider how cultural norms may affect how youth consider and respond to peer provocation. This study demonstrates that there are both similarities and differences in how youth from Pakistan and the United States think about peer provocations and how these interpretations are associated with motivations for revenge. An understanding of cultural variation in these processes is important and relevant for improving the social interactions of youth around the world.
Supplemental Material
sj-docx-1-jbd-10.1177_01650254221121840 – Supplemental material for Interpretations and revenge goals in response to peer provocations: Comparing adolescents in the United States and Pakistan
Supplemental material, sj-docx-1-jbd-10.1177_01650254221121840 for Interpretations and revenge goals in response to peer provocations: Comparing adolescents in the United States and Pakistan by Kristina L. McDonald, Salma Siddiqui, Sunmi Seo and Carolyn E. Gibson in International Journal of Behavioral Development
Footnotes
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The research reported in this manuscript was supported in part by: 1) a dissertation fellowship to the first author funded by a National Institute of Child Health and Human Development Training Grant (T32 HD007376 19) to the Center for Developmental Science at the University of North Carolina; and 2) a dissertation fellowship to the first author from the Kenan Institute for Ethics at Duke University.
Supplemental Material
Supplemental material for this article is available online.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
