Introduction

Intergroup relations are experiencing a rapid and troubling decline on a global scale1,2. This deterioration encompasses a variety of conflicts, including longstanding tensions and violence in the Middle East that have rapidly escalated in the past year, the ongoing war in Ukraine, enduring political strife in the western Balkan states, and increasing tensions between European Union citizens and immigrants3,4,5. This persistent erosion and polarization of intergroup relations has far-reaching implications: it erodes trust and cooperation at both the individual and societal levels6,7,8, hinders the effective functioning and flourishing of societies and poses tangible threats to the well-being of individuals living in such contexts9. Therefore, there is a critical need to identify effective and enduring strategies for improving intergroup relations to ultimately safeguard the cohesion, peace and well-being of human societies10.

Over the past two decades, social scientists have increasingly focused on understanding the processes and conditions that might improve intergroup relations in contexts of intergroup hostilities and ideological polarization11,12,13,14. In particular, social psychology research on intergroup relations has transitioned from a more descriptive approach that aims to understand intergroup relations to a more interventionist approach that aims to develop evidence-based intergroup interventions10,11,12,15,16,17,18,19. Broadly speaking, intergroup interventions are designed to instigate meaningful psychological changes by altering biased beliefs (such as prejudice), alleviating negative emotions (such as hatred), and mitigating behaviours that would harm or disadvantage members of social outgroups10. The overarching aim of such interventions is to foster more respectful, tolerant and equitable relations between antagonistic social groups by changing the minds and hearts of the individuals in these groups.

The development of intergroup interventions in social psychology draws inspiration from research in related disciplines. This cross-pollination includes insights from clinical psychology (for example, emotion regulation20), education (for example, malleability interventions21 and self-affirmation22), and traditional strategies for attitude and behavioural change (for example, paradoxical thinking17 and changing norms23, respectively). Furthermore, develo** and testing intergroup interventions is a collaborative effort led by scholars with diverse yet related foci. For example, some researchers focus on interventions to reduce prejudice15, some are dedicated to interventions that foster conflict resolution19,24,25 or intergroup reconciliation26, and some study interventions to tackle affective polarization (the emotional dimensions of heightened division between ideologically or politically polarized groups)12 (Table 1). Despite the seemingly discrete and relatively autonomous domains in which intergroup interventions are developed, evaluated and put into practice, these interventions all seek to identify effective ways to improve intergroup relations and so contribute to the creation of more peaceful, tolerant and cohesive societies10.

Table 1 Examples of intergroup interventions

The field of intergroup interventions is consistently evolving as researchers and practitioners identify novel interventions or optimize existing ones. There has been a notable transition from traditional interventions, which have primarily focused on techniques such as direct intergroup contact27 or perspective-taking28, to broader, more cost-efficient and sophisticated interventions, such as moral exemplar interventions29. Methodologically, there has been an increase in studies that field test interventions in real-world contexts, moving beyond WEIRD (Western, educated, industrialized, rich and democratic) settings, and studies are incorporating a diverse range of outcome measures beyond conventional attitudinal measures30,31 including measures of real-life behaviour32, support for policies33 and perceptions of broader intergroup dynamics31. Finally, intervention tournaments34, in which several distinct interventions (with participants drawn from the same population) are compared against a single control group, have been introduced.

However, the field lacks a standardized procedure for designing and evaluating interventions, and there is no systematic and integrated approach for addressing critical challenges related to intervention effectiveness15,35. Indeed, despite the fact that most existing psychological intergroup interventions have received empirical support for their effectiveness on a short-term basis, the long-term impact, applicability and effectiveness of interventions in the real world remain somewhat uncertain15,35. Therefore, there is a need to identify key factors that might enhance intervention effectiveness and durability as well as the social-psychological processes that facilitate or hinder successful intervention outcomes.

In this Review, we consolidate existing research on intergroup interventions from four domains within social psychology (prejudice reduction, conflict resolution, intergroup reconciliation and affective polarization). First, we provide an overview of existing categorizations of intergroup interventions and propose an additional taxonomy of interventions based on the level (individual, relational or societal) of social-psychological processes they aim to address. Next, we discuss the importance of identifying meaningful intervention goals (what), crucial characteristics of intervention recipients (who) and key contextual features (where) for optimizing intergroup interventions. Finally, we suggest two key psychological barriers that might hinder successful implementation of intergroup interventions (motivation and conformity) and strategies for surmounting those challenges.

Categorization of intergroup interventions

There are many ways to categorize intergroup interventions. Intergroup interventions can be categorized on the basis of their form, as experience-based, skills-based or information-based36,37. For example, contact30,31,38 or self-affirmation interventions22,39 provide individuals with a valuable and new experience (such as interacting with individuals from different social groups or reminding individuals of their self-worth, respectively) from which they might infer a conclusion that differs from pre-existing perceptions and beliefs. Emotion regulation40 or perspective-taking41,42 interventions teach participants a skill (such as how to take the perspective of others) through guided training. These skills, in turn, can ‘unfreeze’ existing intergroup perceptions and beliefs. Group malleability19,21 or moral exemplar26 interventions present people with information that contradicts their existing intergroup perceptions and beliefs. For instance, learning about morally exemplary behaviours of outgroup members challenges the belief that outgroups are inherently immoral, thereby facilitating perceptions that not all members of an outgroup are bad people43.

Interventions have also been classified by their goals or what they aim to achieve24. According to this scheme, interventions aimed at prejudice reduction such as contact30 or perspective-taking44 would be categorized together despite differences in their form. Similarly, interventions that emphasize commonalities between political groups45 or dialogue trainings centred on taking the perspective of political outgroups46 would be classified as interventions aimed at reducing affective polarization. Other taxonomies of intergroup interventions categorize interventions according to whether they focus on changing the individual to promote bottom-up changes or whether they focus on institutional changes to promote top-down changes25.

Finally, researchers have classified interventions on the basis of whether they are brief, cost-effective and easy to implement (light-touch interventions)47 or whether they are more complex, requiring additional conditions for implementation15. For example, instructing participants to read three pages of stories about moral exemplars lasting between 5 and 10 minutes is a light-touch intervention32. By contrast, a contact-based intervention where participants from different social groups are brought together to discuss stories of moral exemplars face-to-face during a one-day-long field workshop38 uses similar materials but is not considered light touch.

Adding to the existing taxonomies of intergroup interventions described above, we propose a multi-level nested classification of interventions based on the level of social-psychological processes or barriers they aim to address: individual, relational and societal (Fig. 1). For example, interventions aimed at reducing defensiveness in the face of threatening information39 or facilitating beliefs about group malleability in settings of intergroup conflict19,48 would be considered individual-level interventions; contact-based interventions30,38 aimed at minimizing social distance between different social groups would be considered relational-level interventions; and institutional messaging such as norm-based interventions49, interventions communicating institutional signals50 or various types of media intervention51 would be considered societal-level interventions. This categorization integrates existing typologies, map** onto both bottom-up and top-down approaches while including interventions that take different forms (information-based, skills-based and experience-based) that can range from light-touch to complex within each level.

Fig. 1: Multilevel nested taxonomy of intergroup interventions.
figure 1

Interventions are classified according to the level (individual, relational or societal) of social-psychological processes or barriers they aim to address. These levels complement, interact and align with one another such that interventions that target one level might produce observed changes at another level. The scale from individual to societal corresponds to the notion of bottom-up versus top-down interventions, and interventions within each level can range from light-touch to complex.

Furthermore, intervention levels are nested to reflect the fact that these different target levels are not contradictory or antagonistic to each other, as is sometimes portrayed within the field of intergroup interventions by raising the question of the optimal level interventions should target. Instead, these levels complement, interact and align with one another. Indeed, these levels reflect both the level at which interventions intervene (that is, the independent variable) and the level at which interventions produce observed changes (that is, the mediator or dependent variable), which are not necessarily the same. For example, interventions that target individuals’ beliefs and emotions about others (individual level) might lead to changes in how people interact and behave towards members of adversary groups (changes observed at the relational level). Such an interaction between individual (the target level) and relational levels (the level of observed change) might then in conjunction produce broader changes at the societal level. By contrast, interventions that aim to change policies or norms (societal level) can set the stage for more positive intergroup interactions at the relational level.

Importantly, targeting solely one level or producing changes only at one level is insufficient to achieve sustainable improvements in intergroup relations. For example, adopting policies that encourage intergroup interactions (for example, student exchange programs) or that might foster commonalities between groups (for example, creating national sports’ teams in ethnically divided contexts) might not improve intergroup relations unless individuals express some degree of openness and willingness to embrace the other. Understanding that the levels at which intergroup interventions intervene are not necessarily the same levels at which change is observed might inform tailoring of more effective interventions, because different levels might be relevant for intervention design versus intervention evaluation. Such a categorization might also elucidate how to scale and implement interventions, because certain levels might be more suitable for effective intervention implementation.

Finally, existing intergroup interventions draw on different psychological theories, potentially leading to fragmented and idiosyncratic predictions. Indeed, some interventions (such as self-affirmation interventions39,52) drew insights from existing and well established psychological theories during the developmental phases, whereas other intergroup interventions (such as moral exemplar interventions26) were developed by testing initial intuitions in the real world prior to connecting the empirical observations to existing theories of human functioning. Recognizing that effective and sustainable improvement in intergroup relations requires targeting and producing changes at different yet interrelated levels might serve as a critical initial step toward theoretical advancement. For example, integrating social identity and contact theories could lead to comprehensive predictions about how social identification at the individual level might influence intergroup dynamics and outcomes of contact-based interventions at the relational level. Similarly, integrating emotion regulation theory (which offers predictions and insights at the individual level) with intergroup emotions theory (which offers explanations pertinent to the relational level) might lead to more informed predictions and identify mechanisms that drive intervention effectiveness. Ultimately, this multi-level taxonomy could pave the way for theoretical and empirical developments that address broader and understudied questions pertaining, for example, to the optimal alignment between different interventions and different recipients or societal circumstances.

The importance of what, who and where

Each target level presents distinct opportunities and challenges. For example, interventions directed at the relational level might grapple with the challenge of motivating individuals to engage with members of antagonistic groups35, yet have the potential to instigate meaningful changes31,38. Interventions that target changes within individuals33,39,53 might be particularly beneficial in contexts where institutional support for relational-level interventions (such as intergroup contact) is lacking, and thereby offer a starting point for future initiatives at broader levels. Interventions targeting the societal level (such as institutional messaging interventions) have the potential to induce widespread changes owing to their high scalability. Importantly, intergroup interventions can be effective regardless of whether they target the individual, relational or societal level (Box 1). The critical factor is the extent to which the target level is informed by the goal of an intervention (what), recipient characteristics (who) and the contextual conditions in which it is to be implemented (where).

Although relevant recipient characteristics and contextual conditions might occur across different levels (individual, relational and societal), the intervention target level should not be conflated with critical moderators at each level. For example, political beliefs are an important recipient characteristic to which interventions ought to be adjusted to be effective54. At the same time, interventions targeting the societal level might include content depicting specific political beliefs messages to mitigate affective polarization by highlighting commonalities between political groups45. In other words, political beliefs might be an important recipient characteristic while also serving as an entry point for societal-level interventions.

Most interventions so far have been developed and implemented with broad and generic goals and fail to account for the social context or recipient characteristics. As detailed below, choosing meaningful outcomes and selecting and tailoring interventions to specific audiences and the social contexts in which people live and interact is crucial for develo**, evaluating and implementing effective and durable intergroup interventions across all target levels.

Intervention goals

Intergroup interventions are broadly geared towards instigating psychological or behavioral changes to improve intergroup relations. However, this overarching goal can encompass a variety of specific objectives10, such as reducing prejudice, increasing support for social change, reducing negative emotional barriers such as hatred or increasing positive emotional sentiments such as hope. Frequently, both researchers and practitioners lack clarity regarding their specific objectives, erroneously treating distinct goals as interchangeable. However, interventions that focus on one goal (for example, reducing prejudice) might hinder the promotion of other goals (for example, increasing support for social change)55,56,57,58. Moreover, not all interventions fit all goals, and achieving a specific objective might not be practical or meaningful in all situations. For example, in conflict settings, social groups might coexist peacefully following political resolution of the conflict and therefore the primary aspiration could be reconciliation. By contrast, in contexts where social groups live separately amid ongoing armed conflict, the foremost objective might be cessation of violence. Thus, distinct settings necessitate different goals that align with the specific motivations and needs of the social groups involved.

Furthermore, the motivations and needs of involved social groups might not align with the goal of improving intergroup relations, particularly in contexts where maintaining intergroup distance serves to safeguard ingroup security. Intergroup intervention researchers inevitably take a position within any given conflict by selecting and defining what constitute desirable outcomes from an intergroup relations perspective. Recognizing this inherently normative nature of intervention goals is important and should be considered during goal identification, intervention development and implementation.

Given such complex dynamics between different desired outcomes, selecting the desired, meaningful and specific outcome should be the first step when designing a psychological intergroup intervention. Variations in context and intergroup dynamics necessitate different goals, which, in turn, require appropriate interventions. However, determining the most meaningful goal for intergroup interventions can be a complex endeavour. After determining the goal of an intergroup intervention, researchers need to decide whether they want to change attitudes, behaviours59 or perceptions of intergroup dynamics31. Often, there is a discrepancy between an individual’s attitudes and their actual behaviour (the principle–implementation gap)60. For example, contact interventions in conflict-affected contexts are effective in promoting positive changes in attitudes towards adversary groups but are ineffective in changing actual behaviours towards the outgroup, such as attending a dinner with outgroup members or donating to a business run by an outgroup member30. Thus, different strategies might be required to achieve different outcomes. For example, behaviour-focused interventions can leverage social norms to catalyse change. By accentuating prevalent behaviours within a particular social context, these interventions effectively guide individuals to engage in accepted or desired behaviours61,62,63. By contrast, interventions that address meta-perceptions (what people think others think about their ingroup and social reality in general)64,65,66,67 can recalibrate attitudes by influencing individuals’ perceptions of others’ attitudes towards their own group, thereby subverting stereotypes and prejudicial inclinations67.

Furthermore, individuals can have different goals concerning the outgroup, depending on their own social identity, perspectives or experiences. For example, in emotion-regulation training aimed at reducing interracial conflict, individuals who belong to low-power groups might be motivated to increase anger towards the outgroup to maintain their need for status and power, whereas high-power group members might be motivated to reduce anger and increase guilt to maintain their need for a good moral image and social acceptance68. One study found that reappraisal led to greater restoration of emotional well-being after upsetting political events (compared to participants who were not instructed to reappraise their emotions), but it simultaneously reduced intentions to engage in political action69. Indeed, emotion regulation can hinder rather than promote engagement in collective action across diverse socio-political contexts70. Thus, emotion regulation might be an effective strategy if the goal of the intervention is to decrease negative emotional experiences, but not if the goal is to promote collective action. These findings emphasize the necessity of carefully selecting and aligning an intervention goal with an intergroup intervention.

Finally, different intervention goals might conflict with each other. For instance, promoting reconciliation might require acknowledging past wrongdoings and addressing human rights violations. By contrast, the goal of ending violence might require prioritizing immediate security measures over addressing moral wrongdoings. For example, perspective-taking interventions can increase empathy (relative to the baseline) and subsequently reduce prejudice41. At the same time, this increased empathy can be associated with lower levels of guilt71, subsequently decreasing support for policies that advocate social change and other outcomes that require individuals to make concessions for moral wrongdoings39,72. Thus, researchers and practitioners must carefully weigh the trade-offs and nuances associated with different goals and identify the most fitting intervention for each situation.

In summary, to be effective, goal selection for intergroup interventions must take into account the dynamic — and sometimes opposing — interplay between desired outcomes, the recipients and the context. Not all interventions fit all goals at all times and across all situations, and in some cases different relevant goals might be contradictory. Identifying the goal will guide intervention selection and determine the appropriate target level for that intervention.

Recipient characteristics

Intergroup interventions are aimed at diverse individuals, each with their own perspectives, experiences and attitudes. What might resonate or be effective for one person or group of people might not have the same impact on another. Some individuals might already hold more inclusive or accepting views, making them more amenable to change, whereas those with deep-seated biases might be more resistant to change.

According to one framework37, the individual-level parameters that are critical for optimizing the fit between individual characteristics and motivations with intergroup interventions include ideology and values relevant to the intergroup context, dominant emotional sentiments, and individual differences rooted in personality traits. Although there are inherent challenges in empirically identifying such subgroup or heterogeneous treatment effects, particularly in studies that rely on small sample sizes, tailoring intergroup interventions to these recipient characteristics is essential for enhancing their effectiveness and durability over time.

Individual ideology

In the context of intergroup relations, ideology is the lens through which individuals experience relations with other groups37, sha** their beliefs and attitudes about those groups. In intergroup conflict situations, political ideology is one of the strongest predictors of conflict-related attitudes and policies73,74 including inter-partisan affective polarization67,75. For interventions to be effective, these ideological differences must be recognized, because the psychological needs and motivations associated with each ideology determine the type of content that resonates with them76,77.

Interventions can be tailored to psychological needs associated with specific ideologies. For instance, right-leaning ideologies often align with motivations for order, stability and structure, reflecting a desire to preserve existing societal norms and traditions. These motivations stem from a need to minimize uncertainty and maintain a sense of security within an established worldview76,78,79. Consequently, individuals who endorse right-leaning ideologies might tend to respond more positively to interventions that employ a binary, well-defined and even one-sided approach80 than those who endorse more left-leaning ideologies. By contrast, left-leaning ideologies are often associated with motivations for change, equality and progress, leading individuals who endorse left-leaning ideologies to prioritize inclusivity and societal transformations that promote fairness and social justice76. Interventions tailored to individuals who endorse left-leaning ideologies, who exhibit an inclination for novelty and stimulation, are therefore likely to work best when they incorporate elements of surprise and ambiguity81,82,83.

Alternatively, intervention content can be framed to align with an individual’s deeply held moral values and beliefs84 (moral reframing), which should foster engagement and receptivity to the intervention84,85,86. Indeed, moral reframing of the content embedded within intergroup interventions increased the effectiveness of interventions for individuals who endorsed both right-leaning and left-leaning ideologies87,88.

Dominant emotions

Emotions influence attitudes, motivations and behaviours in intergroup contexts89. Emotions can also develop into overarching sentiment or inclinations when reacting to outgroups90. Emotional sentiments differ between individuals. Emotions signal what matters to people, and many different emotions can be relevant in the context of intergroup relations (for example, empathy, guilt, remorse, pride and humiliation). Therefore, interventions should identify the most dominant emotion in an intergroup setting and then seek to address the core appraisal associated with this emotion37. For example, individuals who predominantly experience despair or lack of hope91 might be receptive to interventions that offer novelty and change; individuals who are dominated by fear, which is often linked to feelings of personal vulnerability and perceived inability to cope with threats92, might benefit from interventions that provide reassurance and affirmation37; individuals characterized by high levels of anger stemming from the perception of the unjust behaviour of others93,94 might benefit from interventions that align with their inclination to take action to rectify the situation; and individuals who are dominated by hatred might benefit from interventions that contradict the appraisal of the outgroup’s character as fixed and evil95.

Personality characteristics

Recognizing diverse personalities among intervention recipients might enable more personalized and consequently more effective interventions37. Personalization in this regard entails aligning interventions with the unique needs and values of individuals based on their relatively stable personality characteristics, such as the Big Five personality dimensions96. Indeed, the Big Five traits have been associated with intergroup attitudes and biases97,98,99. The fit between an intervention and an individual can be based on congruence, where interventions match an individual’s values and goals, or on incongruence, where interventions address an individual’s needs37. For example, extraverted individuals are sensitive to rewards and social attention100 as well as activities, challenges, excitement and pleasure101. Consequently, they might respond positively to interventions that challenge their existing beliefs and offer social rewards. Agreeable individuals value communal goals and interpersonal harmony102 as well the welfare of people close to them101. They might therefore favour interventions that promote solidarity, well-being and social obligations to benefit people they cherish. Conscientious individuals value achievement, competence, order, self-discipline and efficiency101,103, and might therefore respond positively to interventions that promise stability and consistency, or that potentially increase their capabilities. Neurotic individuals are sensitive to threats and uncertainty104,105, and so might be responsive to interventions that offer reassurance and alleviate their fears. Finally, individuals who are high in openness101,103 might be receptive to interventions that offer novelty and correspond with their values of diversity and universalism.

Social context

Most intergroup interventions have been developed and evaluated without systematically accounting for the nuances of different social contexts. However, what might be effective in one specific context might not be effective in another. Here we describe four context-level parameters that might be critical for advancing the study and practice of intergroup interventions, based on two related lines of social-psychological theory and empirical research: prejudice reduction and intergroup contact.

Degree of conflict and perceived threat

Social environments that pose realistic or perceived threat can hamper a person’s sense of control because their ability to effectively influence societal circumstances is limited106. Thus, realizing or anticipating that future events will be aversive (such as intergroup conflict) can threaten the human need for control107. From a social-psychological perspective, belonging to social groups functions not only as a tool for survival but also as a resource for satisfying psychological needs108,109,110. In situations of intergroup threat and hostilities, group memberships become especially important because they can safeguard individual-level and group-level existence and survival111. The high level of importance and relevance of group membership under conditions of threat can lead to exclusive social categorizations (us versus them) and intergroup bias including outgroup discrimination112,113. Thus, degree of conflict and perceived threat is a major external factor that influences individual attitudes and behaviours towards outgroups108,114.

Consequently, whether an intervention will be effective for altering attitudes and behaviours might depend on the degree of intergroup threat in a specific context where people live. For example, interventions that do not directly threaten the importance of group identity or those that preserve the clarity of group boundaries might be more effective in contexts where group identities are particularly salient, such as contexts of high conflict and threat, compared to contexts with less conflict and threat. Additionally, interventions that provide a sense of safety, certainty and meaning might be more effective in high-conflict or high-threat contexts where these particular needs are threatened compared to low-conflict or low-threat contexts. In fact, emphasizing threat reduction can decrease affective polarization in the USA115. Research systematically examining the impact of realistic and perceived intergroup threats on the effectiveness and outcomes of intergroup interventions is currently insufficient, presenting an important area for future investigation.

Degree of integration versus segregation

One limitation of existing social-psychological research is its primary focus on individual-level variables and its neglect of types of community in which people live116,117. Advances in methods and statistical analyses over the past twenty years have enabled researchers to examine the role of community types on individual orientations towards outgroups55,117. Specifically, research suggests that the effect of communities on levels of prejudice is larger than the effect of an individual’s own experiences with outgroups118. Evidence from seven large-scale survey studies suggests that living in a place where other ingroup members interact positively with members of the outgroup is associated with less prejudice than living in a place with fewer opportunities for intergroup contact, irrespective of individual contact experiences and of whether individuals know the ingroup members who experience intergroup contact118. Thus, a person living in an integrated community (context with a higher mean level of positive intergroup contact) is likely to be less prejudiced than a person with the same level of individual positive contact but living in a more segregated community. Similarly, a preprint that has not yet been peer-reviewed demonstrated that inter-ethnic reconciliation in post-war societies is impeded in more segregated communities compared to more integrated communities2) can facilitate meaningful and effective change in intergroup relations by producing tailored and effective interventions and by informing the selection of intervention targets.

Fig. 2: A procedure for designing and evaluating intergroup interventions.
figure 2

Selecting or develo** an effective intervention ought to begin with establishing a well defined, meaningful and measurable goal that addresses a clearly defined problem related to intergroup relations within a specific context (what). Once the goal is identified, it should be anchored in psychological theories or established principles of human functioning. This theory-driven foundation then guides the selection of an existing intervention or the development of a new one. At this stage, critical characteristics of the intervention recipient (who) and social context (where) are important for creating a meaningful connection between the goal, theory and intervention itself. This process of intervention design (or selection) should be followed by systematic and rigorous empirical evaluation of whether and how the intervention changes attitudes and behaviours over time when compared to those of meaningful control groups. Finally, interventions should be developed with practical, real-world applicability in mind, considering the complexities of the social context in which it will be implemented, and adapted accordingly.

Psychological barriers to implementation

Effectively implementing interventions in contexts of intergroup hostilities is difficult — it is not easy to effectively and durably change minds, emotions and behaviours (see Box 2 for an example of an effective implementation). Here we describe how to overcome two psychological barriers that might influence the impact and sustainability of intergroup intervention success in the real world: motivation (for example, resistance to change35,149) and conformity (for example, uncritical adherence to the ingroup’s narratives150).

The challenge of motivation

Individuals are often reluctant to engage in intergroup interventions and to change their deeply ingrained worldviews and perspectives. This reluctance is closely linked to the belief that one has objective views and to the tendency to attribute the need for change to others rather than oneself151. Furthermore, specific beliefs and emotions about outgroups play a substantial part in fostering a positive sense of identity149 and epistemic security150; it is unlikely that people will want to change something that has a clear and meaningful purpose. Overcoming these motivational barriers150 and making intergroup interventions more appealing to individuals51 is essential to fostering more effective implementation of intergroup interventions across contexts and for different recipients.

Directly aligning interventions with key psychological motives and values that shape how people make sense of the social world152 is the most direct approach to overcoming motivation barriers to intergroup interventions35. For example, appealing to values of social security such as the desire for safety, peace and stability152 while highlighting the destabilizing effects of intergroup violence might be effective for individuals who value safety and stability. Aligning interventions with other psychological motivations or human needs such as preservation of self-worth, positive identity and sense of belonging might ensure the effectiveness and durability of intergroup interventions39. Indeed, participants who learned about morally admirable actions (versus neutral descriptions) performed by their fellow ingroup members were more likely to endorse and engage in prosocial behaviours that benefited the outgroup, owing to the importance of moral identity32. These findings suggest that intergroup interventions that accentuate a value important to individuals and provide people with a sense of positive identity and worth might be effective in contexts where an individual’s positive identity is threatened.

In addition to traditional incentives (such as monetary compensation or academic credit) other types of incentive could be used to increase motivation to engage with an intervention. For example, convincing individuals that participating in an intergroup intervention has specific and strategic benefits such as improved leadership skills48 or interpersonal communication skills153 might boost the motivation to engage35. Indeed, participants who were told that anger impairs decision-making were more motivated to regulate their group-based anger — and expressed less outgroup hostility as a result — than participants who were presented with a message that highlighted the usefulness and benefits of anger154.

Furthermore, providing people with pleasurable or positive experiences through inspiring information26 or captivating artistic platforms such as immersive virtual reality155 or engaging games156 might capture individual attention and increase engagement with intervention content. For instance, the effectiveness of interventions that use stories featuring moral exemplars26,43 can be attributed, in part, to the feelings of inspiration they evoke157. Stories about outgroup individuals who have risked their lives to save ingroup members evoke inspiration and awe in participants, leading them to engage with the information and act157 (that is, change). These findings emphasize the importance of using information that is likely to resonate with and motivate intervention recipients. Similarly, associating social norms with already established social policies could increase the persuasiveness of social norm messages63,158,159. Finally, implementing interventions in a way that uses movies or photography to encourage thoughtful internal reflection about the intervention content might promote change to arise from within individuals38. Actively involving participants in intervention implementation and encouraging them to reflect on their own terms during the implementation phase might yield effective and lasting shifts in attitudes and behaviours.

Motivational barriers can be bypassed entirely by circumventing a person’s conscious decision to participate and engage with intergroup interventions35. One form of bypassing intervention relies on subtle shifts in naturally occurring dialogue. For instance, presenting sensitive policies using passive nouns versus active verbs (for example, ‘the division of Jerusalem’ versus ‘dividing Jerusalem’) promoted conciliatory attitudes among Jewish and Palestinian people in Israel160. Motivational barriers can also be bypassed by embedding intervention content in the recipients’ natural environments. For example, in work that fostered intergroup contact by creating ethnically mixed sports teams or classrooms30,161,162, the athletes or students were not aware that they were participating in a contact intervention. Nonetheless, these interventions led to marked reductions in prejudicial attitudes and behaviours compared to control groups. In another intervention, researchers programlmed social media bots that appeared as high-status ingroup versus outgroup members (for example, individuals from racial or ethnic majority groups with many followers) to reproach users who posted hateful comments. This intervention resulted in marked reductions in racist slurs163.

Future bypassing intergroup interventions could draw inspiration from behavioural interventions that communicate information about social norms through prominent environmental signals or social referents164,165,166. Insights from nudging interventions167 that alter choice architecture (for example, through reminders and feedback) might also offer some promising bypassing tactics for intergroup interventions. However, in situations where bypassing approaches are used to motivate the audience, increase attention or increase the likelihood of engagement, important ethical issues must be kept in mind (Box 3).

The challenge of conformity

People are generally inclined to conform to the beliefs and opinions of others, which influences their perceptions, decisions and behaviours137,140,168. This tendency to conform extends to adhering to existing ingroup norms or ideas that signal desired and expected perceptions of169,170,171 and behaviours toward172,173 social outgroups. Such conformity can counteract or impede the positive effects and potential long-term success of intergroup interventions174. For example, individuals who have undergone successful transformation through an intervention might face strong resistance to their new views within their community, leading them to revert to old perspectives. Therefore, it is important to identify strategies to overcome — or at the very least minimize — the tendency to conform to prevailing norms or ideologies in contexts of intergroup hostilities or violence. We propose three specific recommendations to this end.

First, influential figures should be used to convey non-normative or unexpected information. People tend to pay closer attention to information when it comes from influential individuals175,176 and are less likely to disregard unexpected information if it comes from influential others175,176. For instance, individuals closely observe the actions enacted or espoused by influential members of their own groups, whose expressed attitudes and behaviours guide their own attitudes and interactions with members of other groups177,178,179. Thus, a willingness to express intergroup biases closely tracks the social norms perceived from fellow group members, and the extent to which those biases against certain groups are deemed to be socially acceptable180,181. Leveraging influential figures as sources of information in intergroup interventions could therefore enhance intervention effectiveness. Furthermore, using influential individuals could safeguard against resistance to new perspectives within the social community because behaviours enacted by influential others might exert high social influence.

Second, larger and more diverse samples should be targeted. When interventions target diverse participants, the influence of any single prevailing norm could be diluted. Participant diversity might also introduce a wider spectrum of perspectives, reducing the likelihood of conformity to a specific norm that might act as a barrier to effective implementation. Diverse samples might also promote a more representative understanding of the complexities within a society. Furthermore, targeting a larger number of individuals might increase the likelihood of spill-over effects to non-targeted participants182. For example, one study182 found that a reappraisal intervention reduced emotions in an intergroup context, even among non-treated participants (control group) who were not instructed to use reappraisal themselves. Moreover, the relationship between the number of treated participants and the reduction in the emotion of the control participants was exponential: when 25% of participants were treated, non-treated participants showed a reduction in negative emotions by 0.1 standard deviation. When 80% of participants were treated, non-treated participants’ negative emotions were reduced by nearly 0.3 standard deviation. The scarcity of research addressing the question of how many participants need to be treated to ensure broader effects makes this work particularly important.

Third, appropriate and engaging platforms should be used. For example, there is always a risk of backfire effects when interventions involve discussing sensitive topics such as collective responsibility for committed crimes or topics related to group privilege183. Indeed, face-to-face discussions (for example, embedded in contact-based interventions) might lead to negative and polarizing experiences in contexts of intergroup violence, particularly for topics related to conflict and responsibility for committed crimes38,184,185. Such negative and polarizing experiences often reinforce conformity to pre-existing and prevailing views and biases184. In such cases, a media-based intervention might be more effective than direct face-to-face interaction186. For instance, research in Israel found that participants exposed to a television programme that discussed sensitive content related to outgroups showed less prejudice compared to participants who were not subjected to any intervention (empty control)187. Moreover, interventions delivered through media can be engaging and interesting, thereby enhancing motivation to engage with material aimed at fostering improved intergroup attitudes188,189 and making them less likely to automatically conform to prevailing ingroup views137,140. Furthermore, entertainment–education interventions address the challenge of conformity by using fictional content, which is perceived as less intimidating than realistic (and potentially threatening) content190. Using such content is associated with less resistance from participants than content that directly and bluntly addresses prevailing views about intergroup conflict190. Finally, media interventions can be accessed virtually anywhere, thereby addressing potential logistical barriers and anxieties associated with participating in intergroup interventions that require physical presence188,191.

Summary and future directions

The past two decades have seen the emergence of innovative, effective, practical and scalable interventions for improving intergroup relations in contexts marked by intergroup hostilities and violence. However, a ‘one intervention fits all’ approach persists. In this Review, we have integrated the existing, fragmented, research on intergroup interventions, by proposing individual- and context-level features that are critical for intervention development, evaluation and optimization. A thorough understanding of the impact of these features will enable effective personalization and contextualization of existing interventions and the development of new ones. Although our Review does not present a full and integrative theory of intergroup interventions, we hope that our proposed tailored and integrated approach to intervention development and optimization will serve as an initial step in this direction. Furthermore, our approach can provide concrete and effective tools for policymakers, peacebuilding agents and any stakeholder aiming to implement or fund effective interventions geared toward improving intergroup attitudes and behaviours towards social outgroups.

Research on intergroup interventions has started to address some methodological limitations in this literature10, such as conducting studies in non-WEIRD contexts. However, there is room for further methodological improvement. First, there are still few experimental field studies that use random assignment and that track effects over time15. Future studies should address this gap to test the external validity of empirical evidence obtained in more controlled settings. Second, researchers should preregister their studies to enhance transparency in reporting and minimize the influence of post-hoc decisions15. Third, to address concerns about the generalizability and durability of intervention effects, it is essential to evaluate interventions using large, heterogeneous and representative samples over extended periods of time15. Finally, researchers should continue to incorporate outcome measures related to real-life behaviour whenever possible15,30. Although attitudinal changes are important indicators of intervention effectiveness, observing actual behavioural outcomes in naturalistic settings provides a more comprehensive understanding of intervention impact.

Although improving intergroup relations is important, it is essential to recognize that changes brought about by intergroup interventions do not automatically translate into the immediate resolution of conflicts and hostilities between groups. In other words, the effectiveness of intergroup interventions in altering attitudes and, at times, the real-life behaviours32 in contexts of intergroup hostilities does not guarantee the resolution of actual conflicts111 or the establishment of lasting peace. Furthermore, there is a lack of evidence regarding the broader impact of interventions in real-world settings. Consequently, research on intergroup interventions faces the formidable challenge of exploring the extent to which changes in attitudes and behaviours are linked to actual peacemaking on a broader societal scale. Systematic personalization and contextualization of interventions are imperative steps toward addressing these important questions.

Another important gap in the field of intergroup interventions is the lack of comprehensive and systematic understanding of the psychological mechanisms that drive observed outcomes of intergroup interventions. Future research should integrate highly controlled experimental studies (such as laboratory studies) with large-scale field studies to elucidate the mechanisms underlying effective interventions, contributing to both theoretical advancement and practical development.

Finally, addressing broad societal challenges such as peacemaking or conflict resolution involves multiple levels of change. These levels might not necessarily align in the same direction of change, and inducing changes at one level might not automatically (or ever) translate to changes at another. For instance, individuals whose attitudes towards conflict are effectively altered as a result of an intergroup intervention (for example, who show a greater inclination toward political compromises) will inevitably encounter challenges in the real world, especially in contexts marked by immediate threats or violence. In such scenarios, individuals might be tempted to revert to their initial position, influenced by dominant and often ingroup-serving narratives and interpretations. Individuals might also return to their initial position to fulfill their needs for group belonging, survival and cohesion. Future research should explore whether intergroup interventions have the potential to equip people with the psychological tools to enhance their resilience and prevent automatic adherence to group-serving attitudes and positions in the face of intergroup adversities.