Introduction
Trust, or the belief that others will cooperate when one cooperates (Kydd Reference Kydd2007; Rathbun Reference Rathbun2011), is crucial for peace and conflict resolution, as distrust in the intentions of an adversary can constitute or aggravate a security dilemma (Jervis Reference Jervis1978). Is it possible to increase trust among nations involved in an international crisis? Scholars of political psychology and international conflict disagree on whether emphasizing a superordinate identity that includes different nations, or individual national identities, encourages the public’s trust toward a foreign adversary. This paper asks whether affirming (i.e., bolstering of a positive source of identity; Steele Reference Steele1988) a superordinate identity or a national identity affects trust differently (including whether such affirmations backfire) in the context of an ongoing conflict with a more powerful aggressor state.
Scholarship on superordinate identity mostly focuses on an overarching, common identity shared by different groups as a vehicle to reduce bias between groups in conflict (e.g., Gaertner and Dovidio Reference Gaertner and Dovidio2000). Existing research that builds on this literature, however, often overlooks the power differential between parties, constituent to the superordinate identity. The more powerful group (often, the aggressor state) tends to actively promote such an identity as part of a strategy to gain dominance over the weaker group, sometimes in combination with military coercion.
On the other hand, national identity is a subjective or internalized sense of belonging to the nation (Huddy and Khatib Reference Huddy and Khatib2007; Huddy Reference Huddy, Huddy, Sears and Levy2013). National identity describes attachment to one’s nation that excludes foreign counterparts. Some social scientists argue that emphasizing national identities may be more helpful for international peace (Chung Reference Chung2022a; Glover Reference Glover, McKim and McMahan1997; Miller Reference Miller1995; Taylor Reference Taylor and Hall1998). According to this research, strong, independent national identities provide a sense of security, dignity, and contentment in people’s perception of their membership, which forms the basis for international peace. Furthermore, recent experimental evidence suggests that an affirmation of national identity that fulfills and replenishes the perceived worthiness of one’s nation can move relations with outsiders in a positive direction, through a reduction in biased and defensive response toward them (Chung and Woo Reference Chung and Woo2015; Gunn and Wilson Reference Gunn and Wilson2011).
Given these competing expectations for increasing trust in international relations, this paper asks whether the potential effects of stressing superordinate vs. national identity differ by the varying levels of national chauvinism individuals hold. In contexts where aggressor states attempt a recategorization of identity as a common identity, it is important to identify the conditions under which this strategy may backfire. For instance, individuals who might potentially exhibit the strongest resistance would likely hold preexisting beliefs that their nation is superior to others (i.e., national chauvinists). Yet, when receiving a national identity affirmation, it is likely that those with a sense of national superiority will feel the least amount of resistance to such a message. We therefore examine whether superordinate and national identity affirmations particularly backfire and work respectively among these individuals. We test these competing expectations in a cultural context of Russia-Ukraine post-2014 conflict, which represents a clear power difference between actors.
Borrowing from the literature on information processing, we argue that chauvinistic individuals – in the less powerful group – experience worldview-conflict when facing arrangements that support a common, superordinate identity shared with the outgroup. This cognitive dissonance may evoke resistance to trusting the outgroup. On the other hand, as affirming a national identity involves bolstering its positive sources, exposure to national identity affirmation would appear as worldview-conforming information to strong chauvinists. With little reason to resist the affirmation, chauvinists would thus be in a good position to experience any positive effects of affirmation to their full extent. We therefore hypothesize that individuals low-in-chauvinism will exhibit greater trust after the superordinate identity affirmation, while high-in-chauvinism individuals are expected to increase trust after national identity affirmation.
We test this argument in the context of Ukraine. Mutual animosity between Ukrainians and Russians skyrocketed as a result of the events of 2014: mass protests in Kyiv in 2013–2014 ousted a pro-Russian government; Russia annexed Crimea in March 2014 and fueled the Donbas War in eastern Ukraine in 2014–2015. This war became the bloodiest armed conflict in Europe since the Bosnian War (OHCHR 2017), and the region continues to see regular ceasefire violations. Compared to September 2012, Ukrainians’ positive opinions of Russians dropped by 34 percentage points and their negative views rose by 30 points in 2020 (Levada Center 2020). Ukraine’s democratic politics may leave its leaders as hostages of public opinion where any pro-peace policy is viewed as appeasement of a hostile neighbor. To understand if social science can offer any solutions to overcome this situation of distrust, which inhibits conflict resolution, we conducted survey experiments in Ukraine. We randomly assigned Ukrainian participants into three groups, where participants received treatments that affirmed either their national (Ukrainian) or a superordinate (Eastern Slavic) identity, or a control condition, and then measured their trust toward the Russian government and people.
Our first finding is that neither hypothesis is consistent with the full sample: we fail to reject the nulls of no effect. Our exploratory analysis reveals that these null effects are due to region-specific heterogeneous effects that identity affirmation has on Ukrainians in the West vs. the Southeast. Region of residence strongly shapes political attitudes and support for the government in Ukraine, even more than ethno-linguistic patterns or other demographic determinants (Barrington Reference Barrington2002). We find that affirmation of a common Eastern Slavic identity that includes an adversary backfires among high-in-chauvinism individuals residing in the West macroregion of Ukraine (i.e., the more anti-Russian region where, as we show, individuals on average hold more negative attitudes toward Russia compared to the rest of the country).
In addition to individual learning of trust toward Russia due to regional variation of whether people are already generally inclined to be more pro- or anti-Russian, the regions might have social influence on how accepting participants are to the treatment. In other words, region of residence could be causing differences in receptiveness to the treatment based on how much individuals believe others in their communities might share the view toward which the treatment sways them. For example, chauvinists in the relatively pro-Russian, Southeast region might be easier to move to more pro-Russian views through national identity affirmation because they sense that such a view would fit with the community. This result suggests that our expectation of worldview-conflict among high-in-chauvinism individuals when the superordinate identity is affirmed is conditional on baseline negative attitudes toward the outgroup within an individual, or the perceived negative attitudes toward the outgroup within the ingroup.
On the contrary, highlighting Ukrainian national identity enhanced trust toward Russia for high-in-chauvinism Ukrainians who already held a relatively more positive baseline attitudes toward Russia (living in the pro-Russian, Southeastern region of Ukraine). This result suggests that our expectation of no worldview-conflict among high-in-chauvinism individuals when the national identity is affirmed is conditional on baseline positive attitudes toward the outgroup. It suggests the potential to reorient strong chauvinistic attachment to one’s nation into a more inward-looking national attachment that does not entail comparison or superiority over others.
The contribution of this research is twofold. First, by measuring the causal impacts of competing approaches of national identity and superordinate identity affirmation on trust toward an adversary, we help identify the scope conditions of each strategy. Prior work has compared the effectiveness of these identity affirmation approaches at boosting trust (Chung and Pechenkina Reference Chung and Pechenkina2023); here, we attempt a deeper examination of each approach’s limits, given individuals’ prior beliefs. Defining the scope conditions helps policymakers anticipate which framing devices will gain support or backfire among which constituents when promoting a cooperative policy with an adversary.
Second, we test the effects of identity affirmation in the context of Ukraine-Russia Relations, where the level of hostility had been described as a “divorce” by Kuzio (Reference Kuzio2017) prior to the time of data collection and drafting of this article. While many social scientists have studied the armed violence between Ukraine and Russia, far less attention is being paid to the constraints the public opinion of these states imposes on conflict resolution. Previous work has analyzed the prospects of social identity affirmation in areas of fragile peace where war has subsided, such as Northeast Asia (Chung Reference Chung2015), Bosnia and Herzegovina (Čehajić-Clancy et al. Reference Čehajić-Clancy, Effron, Halperin, Liberman and Ross2011), and settings of domestic tension such as Canadians’ guilt recognition toward aboriginals (Gunn and Wilson Reference Gunn and Wilson2011). By contrast, this study conducted in 2020 manipulated identities in areas where the armed conflict was recent and ceasefire violations were ongoing, serving as a harder case for uncovering the effects of identity affirmation on trust. The uncovered limited effects thus hold greater promise for the policy implications of identity affirmation on cooperation and eventual peace for other similar cases.
Motivation
Common Identity and Trust in International Relations
How can groups in conflict overcome distrust? Various studies have advocated the promotion of an overarching sense of commonality, universalism, or homogenization that downplays extant group identities as a way to boost intergroup trust (Putnam Reference Putnam2007; Riek et al. Reference Riek, Mania, Gaertner, McDonald and Lamoreaux2010). The rationale is that once people previously belonging to different groups come to recognize each other as common ingroup members, bias toward the previous outgroup will dissipate.
The theory builds on the foundation that people tend to categorize themselves and others into various social groups. This creates ingroup-outgroup distinctions (Tajfel and Turner Reference Tajfel, Turner, Worchel and Austin1979; Turner et al. Reference Turner, Hogg, Oakes, Reicher and Wetherell1987), which breed ingroup favoritism and intergroup bias (Brewer Reference Brewer1999; Dovidio and Gaertner Reference Dovidio, Gaertner, Mackie and Hamilton1993). Therefore, strong attachment to existing groups reinforces the fault lines of group differentiation, aggravating intergroup conflict (Kinder and Kam Reference Kinder and Kam2009; Mansfield and Mutz Reference Mansfield and Mutz2009; Rousseau and Garcia-Retamero Reference Rousseau and Garcia-Retamero2007). By contrast, a reconceptualization of “we” that now includes “us” and the former “them” extends and transmits ingroup favoritism to members of the former outgroup.
Specifically, theories in social psychology such as the common ingroup identity model postulate that recategorization, where participants’ group identities are replaced with a more superordinate group, integrate previous outgroup members under a common identity of “We-ness” (Gaertner et al. Reference Gaertner, Rust, Dovidio, Bachman and Anastasio1994). It has been argued that this can help reduce prejudice and thus improve intergroup relations. According to the common ingroup identity model, recategorization into an overarching identity can be activated by “increasing the salience of existing common superordinate memberships” or by “introducing factors,” such as common values, that are perceived to be shared by the groups (Gaertner and Dovidio Reference Gaertner and Dovidio2000). Similarly, contact theorists argue that positive interaction between members of different groups can improve intergroup relations, eventually leading to a recategorization of ingroup and outgroup where both are included into a larger, superordinate group (Allport Reference Allport1954).
Despite the agreement between the common ingroup identity model and contact theory that a superordinate identity is conducive to intergroup peace, the feasibility of establishing a superordinate identity is not straightforward, especially in international politics. While the theories of recategorization of identities are built on the assumption that social categorization into different groups breeds ingroup favoritism and outgroup stereotypes, thereby fostering intergroup prejudice (Jackson and Smith Reference Jackson and Smith1999; Mummendey, Klink, and Brown Reference Mummendey, Klink and Brown2001; Simon, Kulla, and Zobel Reference Simon, Kulla and Zobel1995), it is uncertain how these may carry over to international relations. For example, studies conducted with individuals in member states of the European Union have found that relatively new, overarching European identity has not replaced extant national identities (Polyakova and Fligstein Reference Polyakova and Fligstein2016).
In addition, an important factor in the promotion of a common ingroup identity is that the more powerful group often actively promotes such an identity to gain dominance over the less powerful group. In extreme situations, this also involves military aggression. Any effect of a superordinate identity on trust then must be discussed in the context of such politicized intentions. Experts note that Russia’s actions and words have often revealed similar intentions, arguing that the rhetoric of senior Russian officials indicate perception of Ukraine as a weaker sub-entity of Russia (Charap and Darden Reference Charap and Darden2014). In one example of use of provocative and destabilizing language, Vladimir Putin himself questioned whether Ukraine was a real country in a conversation with former US president George W. Bush (Charap and Darden Reference Charap and Darden2014). This article’s findings apply to this specific type of superordinate identity (Eastern Slavic identity, in this case), propelled in a conflict by the more powerful group as a strategy to achieve dominance. We examine the effectiveness of this strategy in the context of a common Slavic identity that includes both Ukraine and Russia. Identifying the conditions under which promotion of a superordinate identity can backfire in the relatively less powerful state can enhance our understanding of why aggressor states’ attempts to submerge existing national identities into an overarching common identity that includes the defender state may be less than successful.
Researchers have noted that Russia’s annexation of Crimea and attempts to further dismember the Ukrainian state pose something of an imperialist challenge, using coercion and force to take control of and destabilize a smaller neighbor state (Allison Reference Allison2014). Considering the long history of Russian dominance over its neighbors and its permanent seat on the UN Security Council, Russia’s use of force has been compared to wars by dominant powers that attempt to assimilate a neighbor. Indeed, Russian elites have justified annexation, violent conflict, and resistance to diplomatic resolutions (Onuch, Hale, and Sasse Reference Onuch, Hale and Sasse2018 refer to the Minsk I and II accords as an example). This situation further increases implausibility for a superordinate identity that would successfully and effectively expand Ukrainians’ self-identification to include Russia, the current aggressor. Moreover, research has found that aggravation of conflict often leads to further consolidation of national identities (Connor Reference Connor1994; Fearon and Laitin Reference Fearon and Laitin2003; Posen Reference Posen1993).
National Identity and Trust in International Relations
Despite globalization and the rise of international institutions, national identities are stronger than ever – a world without them is unthinkable. Huddy and Del Ponte (Reference Huddy, Del Ponte, Gustavsson and Miller2019) note that national identities will continue to be an influential force in international politics, performing positive functions such as working as social glue (Miller and Ali Reference Miller and Ali2014), creating a collective identity for citizens, and serving as a source of political legitimacy.
In addition, tenets of liberal nationalism and recent studies on identity affirmation propose a contrasting approach: strong national identities may, in fact, promote international peace and cooperation. According to liberal nationalists, independent national identities provide a sense of dignity in people’s membership, which secures a foundation for peaceful and cooperative interaction with outgroup members (i.e., other countries) (Glover Reference Glover, McKim and McMahan1997; Miller Reference Miller1995; Tamir Reference Tamir1995; Taylor Reference Taylor and Hall1998). Research has also found that emphasizing national identity may promote international reconciliation by boosting trust, guilt recognition, and positive images between groups (Chung Reference Chung2022a; Gunn and Wilson Reference Gunn and Wilson2011).Footnote 1 These results suggest that recategorization into a superordinate identity by downplaying national identities is not necessary for conflict resolution.
In particular, the current situation of the Ukrainian context adds promise and confidence to the workings of a strong and salient national identity, while raising skepticism regarding the plausibility of an effective (in terms of improving intergroup relations) superordinate identity. Kulyk (Reference Kulyk2016) argues that the recentevents of Russian aggression have led to a dramatic reemphasis and strength of Ukrainian national identity, where individuals show increased assertion of “self-identification as Ukrainian, greater pride in being a citizen of the Ukrainian state, stronger attachment to symbols of nationhood”, etc. (p. 588); these observations are consistent with the psychological process of assigning blame to out-group that harms members of one’s in-group, as was demonstrated in the Palestinian context (Pechenkina and Argo Reference Pechenkina and Argo2020).
National Chauvinism and Worldview-Conflict When Experiencing Superordinate or National Identity Affirmation
We suggest that the effects of each of the competing approaches (of emphasizing a superordinate or national identity) diverge, depending on the type of attachment one holds toward their nation. Specifically, the promotion of a superordinate identity backfires among individuals with a sense of nationalistic superiority while highlighting Ukrainian national identity boosted trust toward Russia even among the more chauvinistic respondents.
How is affirming national identity different from national chauvinism? The former highlights positive sources of pride of one’s nation, but without comparison toward other countries. National chauvinism, on the other hand, refers to perceived national dominance and superiority over other countries, which can spill over into animosity toward other countries (De Figueiredo and Elkins Reference de Figueiredo and Elkins2003; Dekker, Malová, and Hoogendoorn Reference Dekker, Malová and Hoogendoorn2003).
Researchers have differentiated various types of national attachment. For example, among individuals that strongly associate their selves with their nation, the nature of that attachment can range from a purely inward-looking sense of love that does not entail a sense of dominance over others (e.g., an analogy often drawn here is one of families and children – love for one’s children does not involve a sense of dislike for others’ children; in this sense it is purely inward-looking). This kind of separation has been noted by researchers as certainly plausible, as national identity is an attachment to one’s country with no necessary implication for how one feels towards other countries (Huddy Reference Huddy, Huddy, Sears and Levy2013; Mummendey, Klink, and Brown Reference Mummendey, Klink and Brown2001; Sniderman, Hagendoorn, and Prior Reference Sniderman, Hagendoorn and Prior2004). In other words, while national identity can generate a sense of liking for co-nationals, this does not necessitate hatred or prejudice toward outsiders (Hopkins Reference Hopkins2001; Huddy and Del Ponte Reference Huddy, Del Ponte, Gustavsson and Miller2019).
To test the effect of this inward-looking attachment that emphasizes national identity salience (without comparison toward outgroups), we borrow from theories of identity affirmation. According to self-affirmation theory (Steele Reference Steele1988), affirming people’s self-identity can decrease biased or defensive response toward others. This is done by bolstering a positive value of self-identity that is irrelevant to the issue at hand (e.g., with regard to response toward others, the affirmed value should be unrelated to relations with those people). As the self is composed of several different domains, by bolstering an alternative value important to one’s identity, that individual’s sense of self-worth is replenished. In this way, affirmation is said to leave the individual in a more content and secure state to deal with threats to the self’s image in an evenhanded way, without reverting to defensive response (Čehajić-Clancy et al. Reference Čehajić-Clancy, Effron, Halperin, Liberman and Ross2011).
More recently, scholars have applied self-affirmation theory to study the effects of affirming social identities in groups (Sherman and Cohen Reference Sherman and Cohen2006). Research suggests that affirming social identities can be an effective strategy for group members to better respond to group-based threats without reverting to group-serving bias (Sherman et al. Reference Sherman, Kinias, Major, Kim and Prenovost2007). The proposed mechanism is similar to that of self-affirmation in individuals. Just as reflecting on positive values significant to one’s own identity allows individuals to face information threatening to a positive image of the self, group-affirmation may operate in an equivalent way for entire social groups. When applied to national identity, theories of affirmation imply that bolstering a positive source of national identity can be used as a strategy to encourage improvement in relations with other nations. Studies find that affirming national identities increases guilt recognition between Northeast Asian states with a history of conflict and cooperative foreign policy attitudes (Chung Reference Chung2023; Chung Reference Chung2022b).
To affirm a group identity, experimental psychologists rely on affirmation tasks. For instance, participants are asked to choose from a list of positive values a single one, which they think is the most important to the identity at hand and explain why. It is notable that all listed positive values are inward-looking; in other words, they in no way refer to the outgroup (in our case, that would be Russia). For example, to affirm the Ukrainian national identity, the participants would choose one value from a list of positive values unrelated to Ukraine’s relations with Russia and explain why the chosen value is the most critical part of the Ukrainian national identity. This exercise aims to put participants in a mindset where the national identity is made salient and the perceived worthiness of one’s national identity is replenished, but not in a way that entails comparison or superiority over another country.
Recent studies have tested the implications of identity salience in an international context by focusing on their effects on international trust (Chung Reference Chung2015). We move this literature further by examining the effects of identity affirmation on trust for individuals with varying levels of chauvinism. Individuals may view their country to be dominant or superior than other countries regardless of the actual power hierarchy between one’s country and another. We therefore utilize measures of chauvinism to assess perceived national superiority. As a psychological assessment, chauvinism may not coincide with actual positions of status. Rather, chauvinism has been found to be associated with individual traits such as authoritarian tendencies (Huddy and Khatib Reference Huddy and Khatib2007) and social dominance orientation or a preference for status hierarchy and inequality (Huddy and Del Ponte Reference Huddy, Del Ponte, Gustavsson and Miller2019).
Psychological research on information processing indicates that individuals resist updating based on information that deviates from their prior beliefs (Cohen Reference Cohen2003; Kunda Reference Kunda1990; Nickerson Reference Nickerson1998). Since chauvinistic individuals perceive their country to be superior to other countries, and favor hierarchy and inequality, encountering a request to recategorize one’s own nation and another (especially one that is not an ally) into a larger group where the two become equal members would conflict with their worldview. Individuals high in chauvinism would likely resist the idea of a common, superordinate identity. Especially because superordinate identities are frequently advocated in wars by imperial or dominant groups when they are in conflict with newly independent or minority groups, a type of wars that is not uncommon, we expect that the resistance of chauvinists in the weaker power against the idea of a common ingroup identity would be strongest among those who already hold extremely negative attitudes against the more dominant power. The theoretical expectation in the context of Ukraine-Russia relations would thus be that promotion of a common Eastern Slavic identity is unlikely to boost trust toward Russia among chauvinistic individuals, particularly among those with strong negative priors against Russia.
On the other hand, Ukrainians with low levels of chauvinism should experience little worldview-conflict when their Eastern Slavic identity is affirmed. Since, according to the common ingroup identity model, affirming a superordinate identity is expected to increase trust toward the outgroup, we expect this effect to be most pronounced among less chauvinistic individuals. Therefore, we expect that gains in trust toward Russia among Ukrainians whose Eastern Slavic identity is affirmed will be concentrated within the subgroup with lower levels of national chauvinism.
Hypothesis 1: Individuals with lower levels of national chauvinism exhibit more trust toward the outgroup if their superordinate identity (i.e., the identity they share with the outgroup) was affirmed, relative to the control group.
By contrast, chauvinistic Ukrainians are expected to experience less worldview-conflict when their national identity (rather than Eastern Slavic identity) was affirmed. In the sections above, we clarified that national chauvinism is not the same as national identity affirmation. The two do not necessarily go hand-in-hand – individuals, for example, may hold a strong inward-looking attachment to their nation, or be affirmed of their national identity, but not particularly think their country is superior to others. However, among individuals who already hold highly chauvinistic tendencies, it can be expected that national identity affirmation would not present a worldview-conflict. Rather, for those who already view one’s country as superior, being asked to think about their national identity’s positive values (even if they are purely inward-looking) would likely provide the comfort of receiving congenial, worldview-conforming information.
We therefore propose that, for strong chauvinists, unlike the effect of affirming a superordinate identity, affirming national identity will have the effect of increasing trust toward the outgroup.
Hypothesis 2: Individuals with higher levels of national chauvinism exhibit more trust toward the outgroup if their national identity (i.e., the identity that excludes the outgroup) was affirmed, relative to the control group.
Research Design
Sample
The survey was fielded by the Kyiv International Institute of Sociology (KIIS) in Ukraine in late May–June 2020 as a phone interview with a random sample of 1,058 individuals. Although 2,000 respondents were interviewed, almost half did not pass the manipulation check in the experimental conditions (the manipulation checks are described below), which is why our resultant sample includes 325 control observations and 371 respondents in the Eastern Slavic identity treatment and 362 respondents in the Ukrainian national identity treatment.Footnote 2 Footnote 3
The survey was described to participants as a study on public opinion. Participants in the study were over 18 years of age. The data collection was terminated after 2,000 complete survey units were acquired.
The study was approved by the Institutional Review Board of […] university (Protocol #[…]). Participants were asked to provide consent before the survey and were notified they could withdraw at any time during the survey. We also informed participants that they may not withdraw their information after the survey is over, since we do not collect any identifiers and may not determine which data are whose. Participants received no participation fee, as is standard for KIIS surveys.
The KIIS used software to generate random mobile telephone numbers; after removing non-existing phone numbers, 2,000 phone numbers were randomly selected and contacted,Footnote 4 resulting in a non-stratified random sample.
Survey instrument: Dependent variables
Table 1 presents descriptive statistics for all measures.
Table 1. Descriptive Statistics
Two dependent variables (all responses are recorded on a 5-point scale of response options) differentiate respondents’ trust toward the Russian government as opposed to trust the Russian people. These trust questions are adapted from the general trust questions in the World Values Survey.
Trust toward the Russian government is a composite measure (the scale reliability coefficient of Cronbach’s alpha equals 0.72) of two items: i) how much respondents trust the Russian government; ii) whether respondents believe that the Russian government would exploit Ukraine for its own benefit, or treat Ukraine fairly.
Trust toward Russian people is a composite measure (the scale reliability coefficient of Cronbach’s alpha equals 0.67)Footnote 5 of two questions: i) how much respondents trust the Russian people; and ii) how selfish or kind the respondents believe that the Russian people are.
Importantly, Cronbach’s alpha assumes unidimensionality (Cortina Reference Cortina1993, 102–103), which is why we verify via the principal component analysis whether the items all measure the same underlying component. Each pair of items indeed only has a single dimension, therefore composite measures are appropriate in this case.
After generating composite indexes, we have also used a linear transformation to rescale the range of these measures from a min of 1 and max of 5 into a min of 0 and a max of 1, which eases substantive interpretation. Substantively, one’s level of trust toward Russian people is dramatically higher (average of 0.4 on a 0 to 1 scale, the modal outcome is the middle value of 0.5) than the average trust of Ukrainians toward Russian government (0.19 on a 0 to 1 scale, the modal outcome is 0).
Survey instrument: Independent variables
Measuring national chauvinism
The level of respondents’ national chauvinism is captured by the question of how much better the world would be if people from other countries were more like Ukrainians (Kosterman and Feshbach Reference Kosterman and Feshbach1989). Responses are recorded on a 5-item scale from “Not better at all” to “Extremely better.” The resultant measure Chauvinism is right-skewed with a mean of 2.4 and a modal outcome of 1.
Experimental conditions of group identity affirmation
To test our hypotheses, we need to interact one’s level of chauvinism with social identity affirmation. Social identity affirmation is done via a survey experiment with two treatments and a control group, such that 1/3 of the sample received the group affirmation of the common Eastern Slavic identity; 1/3 received the group affirmation of the national Ukrainian identity; and 1/3 served as a control that follows the same structure as the treatments but does not allude to identity, instead asking about dessert preferences (between-subjects design).
Treatment 1: Superordinate (Eastern Slavic) identity affirmation:
-
1. There are many positive aspects about being Eastern Slavic. Please choose only one of the following items that you think is the most important value for Eastern Slavs:
-
• family; beauty; patience; hard work; liberty
-
-
2. Why did you choose the value you chose above as the most important to Eastern Slavs? Why do you think that value is important to Eastern Slavs? Please explain your choice in 1–2 sentences.
-
• Open answer
-
-
3. How is the value you chose above expressed among Eastern Slavs? Please answer in 1-2 sentences or give an example.
-
• Open answer
-
-
4. Manipulation check: The task on values made me think about:
-
• Things Eastern Slavs value about themselves / Things Eastern Slavs do NOT value about themselves.
-
This manipulation check is a simple yet straightforward way of verifying whether participants were paying attention to the task and thus thinking about values associated with said identity. In addition, this measure has been used as a reliable manipulation check in previous research that uses experimental treatments of identity affirmation (Chung Reference Chung2022a).
Treatment 2: National (Ukrainian) identity affirmation
This treatment and the manipulation check are identical to treatment 1, except all instances of “Eastern Slavic or Eastern Slavs” are substituted with “Ukrainian.”
Control
-
1. Marmeladki (jelly beans) are a chewy candy. The following is a list of flavors of marmeladki. Please choose only one of the following flavors that you think will be tastiest.
-
• Sizzling Cinnamon / Tropical Mango / Apple Jack / Blueberry Balloon / Tutti-Fruitti
-
-
2. Please explain why you think the marmeladki (jelly beans) you chose will be tastiest in 1–2 sentences.
-
• Open answer
-
-
3. When you imagine the taste of the marmeladki (jelly beans) you chose, what do you think it would taste like compared to the others you did not choose? Please explain your choice in 1–2 sentences.
-
• Open answer
-
-
4. Manipulation check: The task of jelly beans made me think about:
-
• Flavors I would like / Flavors I would NOT like
-
Except for the type of identity (Eastern Slavic or Ukrainian), treatments 1 and 2 are identical. The control group follows a similar procedure but does not consider anything related to identity.
The control task on jelly beans is borrowed from earlier psychological experiments that test the impacts of identity affirmation (Critcher, Dunning, and Armor Reference Critcher, Dunning and Armor2010). It is crucial in the design of the study that the control task is similar in structure to the treatment – however, it is substantively unrelated to treatment (in our case, group values associated with group identities). In summary, the control task on dessert/candy preferences is substantively irrelevant to treatment conditions, yet mimics the structure of the treatment exercise.
In summary, assuming randomization was done correctly (it was, as discussed below), any intergroup differences in trust towards the Russian government/people should be attributable to the treatment’s interaction with chauvinism. Furthermore, following the best practice of measuring moderating variables pre-treatment (Montgomery, Nyhan, and Torres Reference Montgomery, Nyhan and Torres2018), the questions on chauvinism precede treatment.
Survey Instrument: Controls
Since respondents were randomly assigned to each of the three experimental groups, these groups should be comparable on average on both observable and unobservable characteristics. To verify this claim, we collect relevant information on observables (about sex, age, settlement type) to ensure that randomization was done properly.
Empirical Strategy
First, employing the demographic attributes, we verify whether the experimental groups are comparable on average. We estimated differences-in-means between treated and control units with respect to treatment assignment; all differences are substantively and statistically negligible, with exception of age: older individuals were more likely to pass the manipulation check for the NI treatment. This means passing the manipulation check question did not correlate with any of the sociodemographic attributes except for age in the NI treatment.
Second, since our experimental groups are largely comparable, we employ differences-in-means to uncover average between-group differences in trust levels (between the three groups of respondents) obtained from two-tail two-sample t-tests.
Third, as our main analysis, we estimate the same differences-in-means via linear probability models, adding controls. Since our sample is fairly small and non-stratified, it is possible that by chance we ended up with individuals who were more susceptible to treatment in the treated conditions. The KIIS reweighed the sample based on four attributes (macroregions, type of settlement, age, gender) in accordance with 2019 data collected by the Central Election Commission of Ukraine and the State Statistics Service of Ukraine; we used these weights in the regression analyses; all takeaways remain.
Finally, to explore the heterogeneous treatment effects contingent on chauvinism levels, we use split sample analyses where we re-estimate said linear probability models on subsamples of western, center-northern, and southeastern residents of Ukraine. This approach helps explain why treatments had virtually no effect in the full sample.
Analysis
Our other work demonstrates that exposure to national identity affirmation and superordinate identity affirmation fail to increase trust toward an outgroup in the context of an ongoing crisis in Russia-Ukraine relations (Chung and Pechenkina Reference Chung and Pechenkina2023). This article asks whether subjects with varying levels of chauvinism process the two competing identity affirmation treatments differently, which, perhaps, helps explain why treatments (without conditioning on other attributes) fail to increase trust.
Table 2 presents differences-in-means for split samples: individuals whose answers ranked as below median on the chauvinism measure are compared to those at or above median levels, i.e., those respondents who believe that the world would be better if other countries are more like Ukraine.Footnote 6 We observe that all differences are substantively and statistically negligible.
Table 2. Differences-in-Means in Trust toward Russian Government (Composite) by Treatment
Table 3. The Effect of National Identity (NI) and Superordinate Identity (SI) Affirmations on Trust toward Russian Government and Russian People
Note: * p < 0.1, ** p < 0.05, *** p < 0.01.
Next, we employ linear regression models as a robustness check to re-estimate all differences-in-means as interaction effects between treatment conditions and chauvinism measures, adding control covariates and weights.Footnote 7 Table 6 reports these OLS estimates. Models 1–2 estimate the impact on trust toward Russian government; Models 3–4 use trust toward Russian people as the dependent variable. All coefficients have an intuitive interpretation (they report the amount of change in trust on a 0 to 1 scale). The interaction terms require further visualization, because Chauvinism is an ordinal scale, implying that the effects of treatments may have a statistically discernible impact for certain values of Chauvinism. This analysis (not shown) reveals that neither the NI nor the SI affirmation treatments affect trust in our sample for any level of chauvinism. These findings are not consistent with hypotheses 1 and 2.
To sum up, our results are inconsistent with the two expectations we set out to test. To understand whether region-specific heterogeneous effects cancel each other out in the full sample, we present exploratory results based on split samples specific to the macroregions of Ukraine.
Exploratory Analysis Using Region-Based Split Samples
In Ukraine, different history split along the Dnieper River has resulted in different ethnic and linguistic compositions across its macroregions, which also shapes political attitudes (Barrington Reference Barrington2002). This regional contrast explains why underlying attitudes toward Russia differ by macroregions as summarized in Table 4. We observe that the West of Ukraine is 17 points more (on a 100-point scale) anti-Russian and 0.20 points (on a 4-point scale) more chauvinistic, while Southeastern Ukraine is 15 points more pro-Russian and 0.12 points less chauvinistic (although only statistically discernible in one-tail test) than the rest of the country. Therefore, it is intuitive that social identity affirmation’s effect on trust varies by these baseline attitudes toward Russia, for which we use respondents’ macroregions as a proxy.
Table 4. Differences-in-Means in Attitudes toward Russia and in Chauvinism by Macroregion
Table 5 presents linear regression models with controls and weights. Models 5–7 estimate the impact of the National Identity (NI) affirmation on trust toward Russian government and Models 8–10 on trust toward Russian people. All coefficients have an intuitive interpretation (they report the amount of change in trust on a 0 to 1 scale). To fully interpret the interaction between one’s levels of Chauvinism and exposure to the national identity affirmation, we use Figures 1 and 2, which chart the predicted probabilities for the Western, Central-Northern, and Southeastern regions of Ukraine for each dependent variable respectively.
Table 5. The Effect of National Identity (NI) Affirmation on Trust toward Russian Government and Russian People in Macroregion-Specific Subsamples
Note: * p < 0.1, ** p < 0.05, *** p < 0.01.
Figure 1. uses the estimates from Models 5, 6, 7 of Table 5 to predict levels of trust toward the Russian government when we vary region, chauvinism, and the NI treatment. Overall, these visualizations reveal a major difference between the pro-Russian Southeastern region and the rest of the country. That is, in Western and Central-Northern Ukraine, respondents’ exposure to the national identity treatment lowers trust in Russian government as one’s chauvinism increases. By contrast, one’s chauvinism interacts differently with the national identity affirmation in the pro-Russian region of Southeastern Ukraine: the treated subjects are more likely to report higher trust in Russian government as their level of chauvinism rises.
Figure 2. uses the estimates from Models 11, 12, 13 of Table 6 to estimate the predicted levels of trust toward Russian government when we vary region, chauvinism, and Superordinate Identity (SI) treatment. Overall, these visualizations reveal a major difference between the Western region on the one hand and the rest of the country. That is, in Western Ukraine, respondents’ exposure to the superordinate identity treatment lowers trust in the Russian government among high-in-chauvinism respondents the most. Consider, for instance, an individual median in chauvinism from Western Ukraine: when randomly assigned to the control group, the predicted level of trust is 0.21 (on a 0 to 1 scale), but after the SI affirmation this level drops by 10 units to 0.11. This difference is statistically discernible (p<0.01). Further, consider a high-in-chauvinism individual residing in Western Ukraine: when assigned to the control group, they are predicted to report the level of 0.35; however, when assigned to the SI affirmation, this level drop by 28(!) units to 0.07 (indistinguishable from 0 as shown in the middle graph of Figure 2); the difference is statistically discernible (p<0.001).
Consider, for instance, an individual high in chauvinism from the relatively anti-Russian Western Ukraine: when randomly assigned to the control group, the predicted level of trust is 0.34 (on a 0 to 1 scale), but after the NI affirmation this level drops by 16 units to 0.18. This difference is statistically discernible in a one-tail test. Further, consider a high-in-chauvinism individual residing in Central-Northern Ukraine: when assigned to the control group, they are predicted to report the level of trust of 0.22; however, when assigned to the NI affirmation, this level drops by 17 units to 0.05 (indistinguishable from 0 as shown in the middle graph of Figure 1); the difference is statistically discernible (p<0.05). Finally, consider an average-in-chauvinism or high-in-chauvinism individual residing in Southeastern Ukraine: if assigned to the control group, they are predicted to report the levels of 0.19 and 0.11, respectively; however, when assigned to the NI affirmation, this level rises by 8 and 20 units to 0.27 and 0.31 respectively; both differences are statistically discernible (p<0.05). To sum up, we observe that only the subsample of the Southeastern subjects is consistent with our second expectation that individuals with higher levels of national chauvinism would exhibit more trust toward the outgroup upon affirming their national identity.
Our first hypothesis posits that individuals with lower levels of national chauvinism exhibit more trust toward the outgroup if their superordinate identity (i.e., the identity they share with the outgroup) is affirmed, relative to the control group. While the abovementioned pattern does not contradict this expectation, the difference between the control and treated units that are low-in-chauvinism is not statistically discernible, which is why we cannot conclude that hypothesis 1 is consistent with the Western Ukraine subsample.
Finally, Figure 2 also graphs the predicted levels of trust in the Central-Northern and Southeastern subsamples: none of the differences between the control and treated units in these regions are statistically discernible. To summarize, we observe that the superordinate identity treatment lowers trust toward Russian government among the median and high in chauvinism individuals in the relatively anti-Russian subsample living in Western Ukraine.
Next, consider Models 8–10 of Table 5, which estimate the same models 5–7 while employing the dependent variable Trust toward Russian people. We visualized the interaction effects (included in the online appendix) to estimate the predicted levels of trust toward Russian people for varying regions, levels of chauvinism, and exposure to the NI and SI treatments. Overall, these visualizations reveal two takeaways. First, the overall differences are greater between macroregions than within macroregions, such that the range of reported levels of trust toward Russian people in Western Ukraine is 0.29–0.39, while the range in Southeastern Ukraine is 0.4–0.52. Second, none of the differences between the control and treated units are statistically discernible at 0.05 level. This suggests that national identity affirmation influences how much trust respondents report toward the Russian government, but it does not impact the amount of trust toward the Russian people.
Similarly, we charted the predicted levels of trust toward Russian people based on Models 14 –16 of Table 6. Similarly, none of the differences between control and treated units is statistically discernible. This suggests that superordinate identity affirmation influences how much trust respondents report toward the Russian government in the Western subsample, but does not impact the amount of trust toward government in other regions or toward the Russian people.
Table 6. The Effect of Superordinate Identity (SI) Affirmation on Trust toward Russian Government and Russian People in Macroregion-Specific Subsamples
Note: * p < 0.1, ** p < 0.05, *** p < 0.01.
Discussion
The null results in the full sample should be interpreted in the cultural context of Ukraine-Russia relations. The recent memory of the high-intensity war in the Donbas and the painful (for many Ukrainians) experience of Russia’s annexation of Crimea make for a difficult testcase for increasing trust via an inward-looking psychological change. In other cases where positive situational changes came first, or elites proactively implemented a cooperative foreign policy such as détente, positive public perception toward a former adversary as more trustworthy might follow more easily.
Although we report an effect that is limited in scope, the effect of the national identity affirmation on increasing trust among strong chauvinists in Southeast Ukraine is worth highlighting. Considering the popular assumption that strong national identity itself automatically leads to aggression in international relations or is a major source of conflict (Bertoli Reference Bertoli2017; Gruffydd-Jones Reference Gruffydd-Jones2017; Mansfield and Snyder Reference Mansfield and Snyder2002; Mearsheimer Reference Mearsheimer2014; Mercer Reference Mercer1995; Schrock-Jacobson Reference Schrock-Jacobson2012; Snyder Reference Snyder2000; Van Evera Reference Van Evera1994), this finding may appear counterintuitive. But it also brings us one step closer to understanding why many other studies (such as work by liberal nationalists or psychological research on group affirmation theory), on the contrary, find positive effects of robust national identities. Bridging this divide requires understanding how affirming different levels of group identity operates for various subgroups of the population; this research is the first step in this direction as we show that these effects are not uniform.
While national chauvinism is often associated with conflict in international relations,Footnote 8 our finding that national identity affirmation may increase trust among high-in-chauvinism individuals depending on prior attitudes toward the outgroup suggests a way to curb and redirect a toxic tendency to emphasize one’s nation’s superiority into a healthier type of national attachment that increases trust toward another country.
Especially combined with current trends of strengthening national identity in Ukraine with the ongoing conflict, the finding that the positive effect of national identity affirmation may vary by the level of chauvinistic Ukrainians’ prior attitudes toward Russia adds to the plausibility and promise of national identity affirmation. Although, based on our findings, some baseline level of positive attitudes toward Russia can be considered a necessary precondition for identity affirmation to have positive effects, it should be noted that researchers have found attitudes are more malleable than identities. For example, Pop-Eleches and Robertson (Reference Pop-Eleches and Robertson2018) report that people can more easily modify their attitudes than shirt identities; thus, identity is a predictor of political attitudes rather than vice versa and people are more likely to shift attitudes to reflect their identities rather than modify their identities to match their politics. This implies that emphasizing the Ukrainian national identity while making adjustments to attitudes toward Russia suggests a more plausible route to increasing Ukrainians’ trust of Russia rather than attempts to promote at a superordinate identity.
Conclusion
Previous research disagrees on which type of identity – superordinate or national – encourages the public’s trust toward a foreign adversary. This article studies whether the efficacy of each approach in boosting trust depends on one’s prior attitudes. We argue that due to differences in chauvinism, individuals may be more or less receptive to either group identity affirmation approach. The results indicate that the two competing identity affirmation approaches generate different effects on trust in individuals with varying levels of national chauvinism and by their region of residence (which approximates existing attitudes toward an outgroup). Our main findings are that, first, national identity affirmation generates greater increases in trust toward the other country’s government among higher-in-chauvinism individuals with relatively positive attitudes regarding the outgroup. Second, the affirmation of a superordinate, common ingroup identity backfires among chauvinistic individuals with relatively negative prior attitudes toward the outgroup, presumably due to resistance against the idea of a common ingroup that includes both ingroup and outgroup.
In addition, these effects are found regarding trust toward the adversary’s government and are negligible regarding trust toward the adversary’s people. This difference in how identity affirmations interact with chauvinism to shape trust toward government vs. the people suggests that Ukrainians may assign blame for the ongoing conflict with Russia to the government more than to the Russian people.
Compared to studies that champion strong national identities as a helpful factor for international peace, the idea that advancement of an overarching, common ingroup identity may seem more intuitively understandable. That is, when individuals, formerly from different groups, are recategorized under an umbrella identity that is commonly shared, the recognition of a former “other” as now “one of us” may help lessen bias toward them. However, these desirable effects are contingent on the process of recategorization being successfully completed, and the new identity has been adequately accepted by the individuals.
This process is easier said than done. We argued that for some individuals, attempting – or, perhaps, merely considering – to combine identity with an outgroup may backfire. In particular, for people who believe their country is superior to others and hold negative prior attitudes toward an outgroup, a nudge to imagine positive elements of a superordinate identity shared with the outgroup produced no gains in trust toward that group and in some measures even decreased trust. Testing how the affirmation of a superordinate identity affects trust in various groups of people helps identify the scope conditions of the common ingroup identity model in finer detail.
On the other hand, affirming national identity increased trust toward Russia even among the most chauvinistic Ukrainians if they held relatively positive prior attitudes toward Russia. Unlike the affirmation of an Eastern Slavic identity, reinforcing a positive source of national pride via national identity affirmation (that does not entail comparison with an outgroup) is not worldview-conflicting for chauvinists who already hold strong attachment to their nation. This allows chauvinists to be more receptive of the idea of national identity affirmation, and with a replenished sense of national greatness, affirmed individuals perceived Russia as more trustworthy. In other words, national identity affirmation offered a way to increase international trust toward a foreign adversary while protecting the integrity and perceived self-worth of Ukrainians. We find, however, that this effect only applies to individuals living in the pro-Russian macroregion, implying that there might be a minimum necessary baseline positive attitude toward the outgroup for the national identity affirmation to improve intergroup relations. These limited results (i.e., differing by levels of chauvinism and region of residence) were likely shaped by the ongoing conflict in Ukraine-Russia relations, where the recency and salience of the international crisis create a challenging testcase of whether an inward-looking psychological identity affirmation may raise trust toward the outgroup. These findings clarify the limits of social identity affirmation’s effect on improving intergroup relations.
What do we make of these findings on national identity affirmation and chauvinism? To be clear, chauvinistic tendencies have been found to be harmful in international relations (studies find national chauvinism correlates with preference for militaristic foreign policy attitudes; Chung Reference Chung2023) and should thus be discouraged. While strong love and pride for the nation has been found to have positive effects in international relations, these apply to the extent that the love for the nation is inward-looking. When attachment to the nation spills over into extreme attitudes of chauvinism, where individuals perceive superiority of one’s nation over others, international peace and cooperation may suffer, as we have seen from various examples of atrocities and transgressions in history.
The result that national identity affirmation increases trust toward the outgroup among chauvinist individuals as long as they hold a baseline positive attitude toward the outgroup suggests that via national identity affirmation, even the strongest chauvinists may channel their strong love for their nation (that includes assertions of superiority and dominance) into a more constructive national pride and love that promotes peaceful coexistence and cooperation with others. In this sense, national identity affirmation may provide a way to reorient the potentially dangerous variants of excessive nationalism into a national identity that is at least innocuous and potentially helpful for international relations. Future studies should more directly test the proposed mechanism.
Acknowledgements
We are grateful to Anton Grushetsky and the entire team of the Kyiv Institute of Sociology for their work in preparing and fielding the survey.
Supplementary material
The supplementary material for this article can be found at http://doi.org/10.1017/nps.2023.49.
Financial support
This work was supported by the Department of Political Science, College of Social and Behavioral Science, University of Utah and the Spring 2020 CARE Grant from the College of Humanities and Social Sciences, Utah State University.
Disclosure
None.
Open access
