Introduction
What drives individuals to risk their lives in defence of their country? The willingness to fight for one’s country is a crucial element in safeguarding the security of any society and represents the ultima ratio for maintaining sovereignty and territorial integrity. This willingness becomes especially pertinent in the face of imminent military threats. Russia’s full-scale invasion of Ukraine in February 2022 has brought defence capabilities into sharp focus, particularly among those European nations near and with close historical ties to Russia. Existing research has explored a wide range of factors that explain why some citizens are more inclined to defend their country than others.Footnote 1 These studies have identified various socio-demographic, socio-economic, political, and cultural variables that consistently drive such attitudes across diverse contexts. However, much of this scholarship has operated on the assumption of societal homogeneity and is grounded in ‘methodological nationalism’. Many countries are, in fact, ethnically diverse, and different ethnic groups often exhibit varying attitudes towards the state and its policies,Footnote 2 support distinct political parties,Footnote 3 and hold divergent views on political alliances and external actors.Footnote 4 Disparities in loyalty to the state between majority and minority groups can become particularly evident in high-stakes situations, such as defending the country during an armed conflict, where personal risk is significant. In multi-ethnic societies, national minorities, who often constitute a substantial proportion of the population, play a crucial role in a country’s defence capabilities. From a security perspective, their engagement and loyalty are key factors in shaping a nation’s resilience. Yet, the existing literature often overlooks this ‘minority dimension’, leaving a gap with significant practical implications. What shapes ethnic minorities’ (un-)willingness to fight for their country?
We study the case of Latvia, which we consider a crucial case to explore this question.Footnote 5 (1) The country exhibits a salient ethnic cleavage, and (2) the issue of national self-defence holds great political and societal relevance. Like other countries that experienced Soviet occupation and Russification policies, Latvia has a significant Russian-speaking population, with roughly one-third of its residents using Russian as their primary language at home. Several studies indicate that ethnolinguistic identity plays a crucial role in shaping political viewsFootnote 6 and voting behaviourFootnote 7 in the country. A significant factor contributing to this ethnolinguistic divide is the differing historical perceptions of the Soviet Union and the varied attitudes towards neighbouring Russia.Footnote 8 The consumption of Russian state media, particularly among Russian speakers, continues to be prevalent despite an official ban in 2022, perpetuating and intensifying the cleavage.Footnote 9 In the wake of Russia’s war against Ukraine and its aggressive actions towards other countries formerly occupied by the Soviet Union, concerns about Russian expansionism have intensified in Latvia. A major focus of these concerns has been Russia’s potential to exploit the ethnic divide. Latvia frequently faces unfounded accusations from Russia, including allegations of discrimination against Russian speakers.Footnote 10 These accusations reflect similar rhetoric previously directed at Ukraine, where Russia framed the Ukrainian government as a ‘Nazi regime’ intent on harming and repressing ethnic Russians.Footnote 11 As a result, discussions surrounding Latvia’s defence capabilities and the role of its Russian-speaking minority have become increasingly prominent in both political and societal discourse.
Despite its importance, the question of whether there are differences in the willingness to fight for the state between majority populations and ethnic minorities – and the factors shaping these differences – remains underexplored. To date, only two studies explore this among Latvia’s Russian speakers: Based on interviews, Bērziņa and Zupa conclude that Russian speakers exhibit a lower willingness to fight for Latvia compared to Latvian speakers and are generally unwilling to fight against Russia.Footnote 12 Andžāns shows that Latvian speakers are approximately 20 per cent more likely than Russian speakers to express a willingness to defend the state. He concludes that the reasons for expressing willingness to defend Latvia are largely the same among both groups, while the number of those willing to defend the state is significantly smaller among Russian speakers.Footnote 13 Two key shortcomings persist: (1) No study to date has systematically analysed individual-level data on attitudes and political behaviour to explore how different socio-demographic, socio-economic, political, and cultural factors influence the willingness to fight among both majority and minority populations. (2) Existing research tends to treat ethnic minority attitudes as homogeneous social collectives, thereby neglecting the possibility of significant within-group variance. It remains unclear whether within-group variance is more significant than between-group variance, that is, whether there are greater disparities within the groups than between them. This is particularly crucial given recent findings highlighting notable internal divisions within these groupsFootnote 14 and suggesting a (potential) convergence of attitudes and preferences between Latvian speakers and Russian speakers,Footnote 15 implying that the ethnolinguistic divide may be less pronounced than previously believed.
In this study, we empirically examine whether cultural divisions help explain differences in the willingness to fight between majority and minority group members. While it is well known that historical memory and collective narratives shape cultural identities, their impact on defence-related attitudes has received limited empirical attention so far. We contend that in divided societies, where interpretations of the past are often disputed, cultural memory becomes a crucial factor in influencing commitment to national defence. Socio-psychological research indicates that perceptions of the past can influence perceptions of in-group threats – an effect particularly pronounced in high-stakes situations. Building on this, we argue that diverging interpretations of the Soviet past mediate the willingness to fight among majority and minority group members in Latvia. Consequently, we expect Russian speakers, who tend to hold more positive views of the Soviet past, to demonstrate lower levels of willingness to defend Latvia in the event of a military conflict. This relationship is influenced by differing threat perceptions: the higher level of Soviet nostalgia amongst Russian speakers diminishes the perceived security threat posed by neighbouring Russia, which in turn reduces the willingness to fight for the country.
Drawing on novel data from three waves of an original public opinion survey conducted in Latvia in 2022, 2023, and 2024, and employing a series of logistic regression models, we find robust empirical evidence for a ‘minority effect’: members of the Russian-speaking minority in Latvia, on average, show lower willingness to fight for the country in a hypothetical war than Latvian speakers. In line with our theoretical expectations, our models show that different perceptions of the Soviet past substantially mediate this ethnic disparity in defence willingness. We find that Russian speakers not only show significantly higher levels of Soviet nostalgia but also tend to have lower perceptions of Russia as responsible for the ongoing war in Ukraine. Our mediation models indicate that it is such different historical perceptions of the Soviet past that reduce the likelihood of blaming Russia for the war and, in turn, shape the willingness to fight for the country. These findings remain consistent when accounting for various socio-demographic and socio-economic factors. Our study has important practical implications for national security planning in multi-ethnic countries.
Individual-level predictors of willingness to fight
Existing literature offers a range of explanations for the factors that influence citizens’ willingness to fight for and defend their country, which can be broadly grouped into socio-demographic, socio-economic, political, and cultural approaches. Among (1) socio-demographic factors, gender consistently emerges as a significant determinant, with men generally exhibiting a greater willingness to fight and defend their country compared to women.Footnote 16 Age has been shown to be another factor, as younger individuals are generally more inclined to engage in combat due to lower opportunity costs and increasing social status with age.Footnote 17 This may also determine the impact of education.Footnote 18 More educated citizens often have alternative status options, raising the opportunity cost of joining the army.Footnote 19 Higher education also implies greater knowledge of foreign policy and war-related consequences, potentially influencing individual motivations. Marital status might also influence the willingness to fight, with married individuals typically exhibiting a lower propensity for personal risk-taking.Footnote 20 However, empirical evidence for socio-demographic explanations remains rather mixed.Footnote 21
Inglehart et al. highlight the (2) socio-economic drivers of the phenomenon.Footnote 22 Their seminal work demonstrates an empirical link between life opportunities, ‘pro-choice’ values, and the propensity to fight: in high-opportunity societies, individuals are less inclined to engage in conflict. However, this relationship is not linear. Instead, it has been observed that citizens in countries with more equal income distribution exhibit a greater willingness to fight compared to those from nations with high income inequality. For instance, citizens in Nordic countries tend to show a strong inclination to defend their countries despite having high life opportunities. Furthermore, personal willingness to fight is inversely related to income inequality at the national level. In countries with high income inequality, higher-income individuals are less willing to fight for their country than their lower-income counterparts. Conversely, in nations with lower income inequality, the willingness to fight is less differentiated between the rich and the poor.Footnote 23
A third strand emphasises the influence of (3) political factors. A widely discussed factor is citizens’ general trust in state institutions, including the government, parliament, and the judiciary.Footnote 24 Levi, for instance, suggests that citizens who view their government as fair and who believe that others are not exploiting their contributions are more likely to defend the state.Footnote 25 In this line, Wang and Eldemerdash find evidence that individuals are more willing to fight when they perceive that others will also fight.Footnote 26 Moreover, confidence in the military and trust in security guarantees have been identified as key predictors of national defence effectiveness.Footnote 27 This is intuitive, as a higher level of (perceived) military capability enhances the likelihood of successful outcomes and decreases individual costs. Current political circumstances and policies play a crucial role, too, as a country’s engagement in ongoing (territorial) conflicts,Footnote 28 ‘wars of choice’,Footnote 29 and the experience of conscriptionFootnote 30 have been linked to individuals’ motivation to defend their nation.
Explanations that bridge political and (4) culturalist perspectives emphasise the importance of nativism, ethnonationalist preferences, or national pride. Studies have repeatedly shown that individuals who feel prouder of their country also tend to report a higher willingness to fight for their country.Footnote 31 Furthermore, different historical experiences and perceptions of the past have been shown to influence individuals’ willingness to defend their country. For instance, citizens of countries that frequently experience conflict, such as India or Israel, often show a higher willingness to fight for their country.Footnote 32
Cultural disparities and the willingness to fight
Threat perceptions and orientations towards collective security are not solely shaped by individual preferences, group grievances, or anticipated material benefits. Instead, they are deeply influenced by social and cultural identities. Cultural and social identities shape individual historical memory, leading people from different ethnic backgrounds to interpret history in distinct ways.Footnote 33 The formation and maintenance of social identity require the reduction of complexity, communication and sharing within social collectives, and continuous reproduction. In this process, identity markers including religion, languages, traditions, and rituals play key roles, as do experiences of history (‘historicity’). Socially shared historical narratives and constructions of the past primarily serve symbolic reproduction as they bridge the temporal divide between past and present and sustain collective identity over time.Footnote 34 When a particular social identity becomes salient, individuals tend to recall and evaluate historical events in ways that are emotionally resonant and shared among in-group members but may diverge significantly from the collective memory of out-group members. In this way, historical narratives and historical consciousness shape intergroup dynamics, which are always situated within broader historical, cultural, and political contexts. This dynamic is particularly pronounced – and politically consequential – in multi-ethnic societies marked by salient cultural cleavages.
Socio-psychological research shows that divergent historical narratives and interpretations of the in-group’s past can result in intergroup biases and different perceptions of in-group threats.Footnote 35 Intergroup bias leads individuals to favour in-group members while harbouring negative feelings towards out-group members.Footnote 36 Importantly, such effects persist even when individuals have never personally experienced the events of ‘chosen trauma’ or ‘chosen glory’ that constitute the collective memory of the in-group.Footnote 37 Intergroup biases emerge particularly strongly in high-stakes scenarios and under conditions of perceived existential threat.Footnote 38 In such situations, the collective perception of threat amplifies in-group favouritism and out-group distrust. For instance, reminders of mortality increase people’s tendency to prefer those similar to themselves while devaluing those who are different.Footnote 39
These heightened biases have direct political consequences: they affect individuals’ willingness to make costly sacrifices, such as risking their lives for a group perceived as distinct from their own. Because collective memory and the sense of collective continuity shape how social groups perceive in-group/out-group relations and assess present-day threats, culturally mediated interpretations of the past significantly influence individuals’ commitment to fight for their country. Consequently, different or even competing historical narratives should explain variations in willingness to fight for the country between majority and minority members. We argue that in divided societies with salient cultural cleavages, citizens’ willingness to fight is profoundly shaped by cultural identity and historically mediated threat perceptions. While this identity-based logic has been applied to explain why individuals join rebel groups or militias,Footnote 40 it has rarely been tested in relation to individuals’ willingness to fight in state armies and interstate conflict.
As in many other multi-ethnic societies, the ethnic cleavage in Latvia is deeply rooted in such contrasting historical narratives. Members of the majority and minority communities hold differing views of the country’s past and exhibit contrasting attitudes towards key historical milestones. Latvian speakers primarily focus on the suffering experienced during Soviet occupation, emphasising the state’s legal continuity and viewing Russia’s aggression against Ukraine as an existential security threat. In contrast, a significant portion of Russian speakers, heavily influenced by Russia’s revisionist narratives, tend to interpret history through the lens of Soviet and Russia’s achievements and share Russia’s espoused enmity towards Ukraine and ‘the West’ at large and the EU, NATO, and the United States in particular. In fact, Russia’s narratives have actively promoted Soviet nostalgiaFootnote 41 and have consistently portrayed Latvia as a failing state and as an entity that heavily discriminates against Russian speakers.Footnote 42 Both of these – a positive view of the Soviet past as well as not seeing Russia as the aggressor in Ukraine – can be seen as signs of idealising the Soviet Union and its successor as the Russians’ symbolic homeland. Furthermore, in Latvia, Russia’s war of aggression is widely perceived by the in-group members, that is, the population in our setting, as precisely such a high-stakes scenario.Footnote 43 This perception is heightened by the minority out-group’s cultural and kin-state ties to the aggressor, further amplifying out-group distrust.
Drawing on our knowledge of the case, we propose that considering the salient ethnic divide and the observed differences in threat perceptions from Russia between Latvian speakers and the Russian-speaking minority, there is a significant difference in national defence commitment. Because the majority–minority cleavage is significantly rooted in contrasting interpretations of the Soviet past, we contend that these historical perceptions are central to shaping threat assessments and, consequently, willingness to fight for the country. We derive the following hypotheses, which guide our empirical analysis:
H1: Russian-speaking individuals in Latvia are less likely to express a willingness to fight for the country compared to the Latvian-speaking majority population.
H2: Divergent perceptions of the Soviet past mediate the effect of ethnic identity on individuals’ willingness to fight for Latvia.
H3: Divergent perceptions of Russia as a security threat mediate the effect of Soviet nostalgia on individuals’ willingness to fight for Latvia.
Research design
Data collection
To test our hypotheses, we utilise data from three waves of original face-to-face public opinion surveys conducted in Latvia across 2022, 2023, and 2024 (n = 3,016). All three surveys were conducted by a reputable public opinion company, SKDS, on nationally representative samples during the following periods: 13–24 May 2022; 3–12 June 2023; and 16–26 March 2024. The 2022 survey included 1,010 respondents, the 2023 survey included 1,002 respondents, and the 2024 survey included 1,004 respondents. We merge the three survey waves based on our core variables, with the 2023 wave incorporating several additional questions for further analysis. All interviews were conducted as part of an omnibus survey, employing a random probability sampling design with stratification based on administrative-territorial divisions.
Our sample comprises 62 per cent Latvian speakers and 37 per cent Russian speakers, closely reflecting official population statistics (Table 1).Footnote 44 In Latvia, as in other multi-ethnic societies, patterns of self-identification are fluid. Russian is spoken by not only ethnic Russians but also other minorities such as Ukrainians and Belarusians. While this cultural complexity adds layers to the social fabric, the salient cleavage between Latvians and ethnic Russians remains a defining feature. This allows for a meaningful empirical differentiation between majority and minority group members, without oversimplifying the social dynamics at play. In light of this, our sample includes only respondents who identify as either Latvian or Russian speakers. We categorise respondents according to their response to the question: What is the main language spoken at home? Minority members (Russian speakers) are operationalised as respondents holding Latvian citizenship who report Russian as their primary language spoken at home. To ensure the robustness of our findings, we re-run our analyses using an alternative classification of minority status based on self-reported nationality (Latvian, Russian).
Table 1. Sample composition (unweighted).
Table 2. Soviet nostalgia and willingness to fight among citizens in Latvia.Footnote 53
* Note: *p **p ***p < 0.01
Variables
Our key dependent variable is citizens’ willingness to fight for their country in the event of a hypothetical attack. Our variable willingness_fight_latvia gauges attitudes towards armed resistance that the state or society might offer in a violent conflict. The measure is derived from the following survey question: ‘Of course, we all hope and believe that there will not be another war, but if it were to come to that, would you be willing to fight for Latvia?’ As we are interested in respondents’ general attitudes, we refrain from specifying a potential invader. Nevertheless, given the current geopolitical context and Latvia’s history, it is reasonable to assume that many respondents interpreted the question with Russia in mind as the most imminent threat to the country.Footnote 45 Responses were originally recorded on a 6-point Likert scale, including options from ‘Definitely yes’ to ‘Definitely not’, with additional categories for ‘Don’t know’ and ‘No answer/ refused’. For our main models, we re-coded the responses into a 1–4 scale and created a binary variable for willingness, where respondents selecting ‘Definitely yes’ or ‘Yes’ were coded as willing and those selecting ‘Definitely no’ or ‘No’ as unwilling. To test the robustness of our findings, we re-run our main analysis using ‘Don’t know’ and ‘No answer/refused’ as a combined reference category in multinomial regression models.Footnote 46 Perceptions of history, our key mediating variable, are measured through respondents’ level of agreement with the statement: ‘Life in the Soviet period was better than now’ (Soviet nostalgia), using a four-point scale ranging from 1 (strongly disagree) to 4 (strongly agree).
We include several potential confounders that capture the four different explanatory approaches for defence attitudes outlined above. First, we account for socio-demographic factors. For gender, we code a dummy variable for male respondents. Age is incorporated as a numerical measurement. Additionally, marital status is included with a dummy variable (single), where 1 indicates that respondents are single, that is, not married, not living with a partner, or widowed. Variables to account for socio-economic explanations include the self-reported personal income (avg_pers_income), the highest level of education attained (education), and dummy variables accounting for unemployment (unemployed) and rural habitation (rural).
We account for political factors using three additional variables: satisfaction with democracy (satisfaction_democracy) is measured using a 0–10 ordinal scale based on responses to the question, ‘On the whole, how satisfied are you with the way democracy works in Latvia?’ Our items measuring respondents’ general trust in the judicial system (trust_judicialsystem) and their trust in the country’s armed forces (trust_army) are assessed on a 1–4 scale.Footnote 47
The majority–minority divide in willingness to fight
Latvia exhibits a pronounced cultural divide – a divide that largely falls along majority–minority lines. Our survey data shows that, on average, Russian-speaking respondents report lower satisfaction with democracy (Δ
$\overline x\,=\,-0.86$), reduced trust in the judiciary (Δ
${{\bar x}}$ = −0.19), and lower levels of national pride (Δ
${{\bar x}}$ = −0.57) compared to majority respondents.Footnote 48 This divide also extends to questions of national defence. Figure 1 illustrates responses to our willingness to fight for Latvia item among both majority and minority group members. We find that, on average, Russian speakers are less willing to fight for their country (Δ
${{\bar x}}$ = −0.61). A Welch t-test confirms that the difference in average fight willingness between Latvian speakers (
$\overline x \, = \,2.86,\,\sigma \, = \,0.972$) and Russian speakers (
$\overline x=2.25,\sigma=0.993$ is statistically significant with a t-value of 14.689 and a p-value < 0.01. An ANOVA further supports this finding (F(1, 2492) = 218.7, p < 0.01), with a within-group variance of 0.96.
Figure 1. Differences between Latvian speakers and Russian speakers in willingness to fight.Footnote 49
Notably, this pattern remains consistent over the investigation period 2022–2024.Footnote 50 While an intra-class correlation (ICC) of 0.081 indicates that most of the variance occurs within groups rather than between the majority and the minority group, we interpret these findings as strong support for H1. Ethnic identity matters, but it is, unsurprisingly, not the only factor influencing individuals’ willingness to fight for their country.
Logistic regression models confirm the critical role of minority status and cultural identity in shaping these attitudes (Figure 2): respondents who speak Russian at home exhibit a significantly lower willingness to fight for Latvia (β = −1.092, p < 0.001). Russian speakers have roughly 66 per cent lower odds of being willing to fight for Latvia, compared to the majority members, controlling for socio-economic and socio-demographic factors. This also holds for the ethnicity variable coded based on self-reported identity of Latvian citizens (β = −0.987, p < 0.001) and remains consistent when including different sets of control variables and year-fixed effects to account for temporal variation.
Figure 2. Factors influencing the willingness to fight among citizens in Latvia.Footnote 51
Soviet nostalgia and the willingness to fight
To test H2, we include our variable accounting for positive perceptions of the Soviet era in logistic regression models. In our baseline model (Model 1), we include gender, age, marital status, and income as controls. Model (2) adds education and unemployment status. Model (3) extends the baseline model by incorporating respondents’ satisfaction with democracy and trust in the judicial system and the army – variables we included in the 2023 wave of our survey. In line with our expectations, we find that across different model specifications, nostalgia for the Soviet era emerges as a powerful predictor of lower willingness to fight for Latvia. Respondents who agree more strongly with the statement that ‘life was better during Soviet times’ are significantly less likely to indicate willingness to take up arms for the country’s defence in case of an attack. Our baseline model shows a significant and substantial negative coefficient for this variable (β = −0.702, p < 0.001). This result remains robust in more extensive models that control for socio-economic variables (Model 2: β = −0.711, p < 0.001) and political factors, including satisfaction with democracy, trust in the country’s judicial system, and confidence in the armed forces (Model 3: β = −0.497, p < 0.001).
The consistent size and significance of the coefficient across models suggest that Soviet nostalgia is not merely a proxy for age, education, socio-economic deprivation, or dissatisfaction with current governance. Rather, historical memories and contrasting interpretations of the past are crucial drivers in their own right and have a direct attitudinal effect on individuals’ willingness to fight for their country.
To test whether Soviet nostalgia mediates the relationship between ethnic identity and the willingness to fight, we conduct a causal mediation analysis. First, we estimate a mediator model to examine how ethnic identity predicts the mediator, Soviet nostalgia. Second, we estimate an outcome model to assess how both ethnic identity and the mediator influence reported defence willingness. We then use nonparametric bootstrapping (1,000 simulations) to assess the Average Direct Effect (ADE), Average Causal Mediation Effect (ACME), and the total effect.Footnote 52 We specify the models as follows:
Mediator model
\begin{equation*}life\_better\_soviet\_time{s_i} = {\beta _0} + {\beta _1}russian\_speaker{s_i} + {\beta _2}mal{e_i} + {\beta _3}singl{e_i} + {\beta _4}ag{e_i} + {\beta _5}incom{e_i} + { \in _i}\end{equation*}Outcome model
\begin{equation*}willingness\_fight\_latvia = {\gamma _0} + {\gamma _1}russian\_speaker{s_{i{\text{ }}}} + {\gamma _2}mal{e_i} + {\gamma _3}singl{e_{i{\text{ }}}} + {\gamma _4}ag{e_i} + {\gamma _5}incom{e_i} + {v_i}\end{equation*}The results strongly support the mediating role of nostalgic beliefs on respondents’ willingness to fight for their country (Figure 3). We calculate an ACME of −0.072 (p < 0.001): in other words, being a Russian speaker correlates with more positive perceptions of the Soviet past, which, in turn, reduces willingness to defend the country in the case of an attack. The total effect of Russian ethnic identity on defence willingness is estimated at −0.214 (p < 0.001), demonstrating a substantial overall association between being Russian-identifying and reduced willingness to defend the country. We find that approximately one-third of the total effect is mediated through Soviet nostalgia. The ADE is significant at −0.142 (p < 0.001), suggesting that ethnic identity matters on its own even when controlling for the mediator. In other words, a positive perception of the Soviet past substantially, but not fully, explains the attitudinal gap between Russians and non-Russians in their willingness to defend Latvia.
Figure 3. Mediation effects of Russian ethnicity on willingness to defend Latvia through Soviet nostalgia.
We conduct three robustness tests to assess the reliability of our mediation analysis: (1) First, we replace our main identity variable again with the alternative measure of minority status based on self-reported nationality, while keeping the dependent variable and the set of control variables unchanged.
The results, shown in Table 4 in the supplementary material, are robust: the ACME is still negative and statistically significant (−0.065, p < 0.001). (2) Second, we add two variables to the set of controls (gender, marital status, age, income) in our mediation models, which may be potential confounders: Individuals with a preference for nativism and ethno-nationalism may be more likely to demonstrate a higher willingness to defend their country. To operationalise nativism, we match respondents’ party preferences with the nativism scores provided by the Populism and Political Parties Expert Survey.Footnote 54 Second, we add a dummy variable that indicates whether respondents reside in Latgale, the easternmost region of Latvia with the highest share of Russian speakers among Latvia’s historical regions, which borders Russia. This accounts for the possibility that both geographic proximity to the potential invader, divergent historical perceptions, and economic situation explain the reduced defence willingness among minority respondents.Footnote 55
The mediation effect remains robust: the ACME remains negative and statistically significant (−0.039, p < 0.05). (3) Third, we examine whether it is national pride, rather than historical memories, that mediates the relationship between minority status and defence willingness. National pride is measured on a 1–4 scale in response to the question: ‘How proud are you to be a national of Latvia?’ While the ADE remains negative and statistically significant (−0.342, p < 0.001), we do not find any mediation effect (ACME = 0, p = 1) in this model. This indicates that national pride does not serve as a mediating factor in explaining the relationship between Russian minority status and reduced willingness to defend Latvia. This contrasts sharply with our primary finding that divergent historical memories account for a substantial portion of this relationship and highlights the specific importance of historical perception rather than general national attachment in shaping defence attitudes between majority and minority members.
Soviet nostalgia, threat perceptions, and the willingness to fight
To test H3 and the mediating role of threat perceptions on the willingness to fight, we ask about respondents’ agreement with the statement ‘Russia is responsible for the war in Ukraine’, coded on an ordinal scale of 1–4. Given the close political, historical, and geographic relevance of the Russo–Ukrainian war for Latvia, assessments of Russia’s responsibility provide a suitable indicator of perceived external threat. To empirically assess the role of different threat perceptions, we extend our mediation analysis by incorporating this variable as a second mediator. We specify a sequential mediation model with three pathways: (1) Soviet nostalgia as a direct mediator between ethnic identity and defence willingness; (2) threat perception as an additional mediator between ethnic identity and defence willingness; and (3) a sequential pathway in which Soviet nostalgia influences threat perception, which in turn affects the willingness to fight. We use nonparametric bootstrapping with 1,000 simulations to estimate standard errors and confidence intervals for all indirect effects.
The mediation models show that Russian speakers not only draw a much more positive picture of the Soviet past (β = 0.249, p < 0.001) but are also significantly less likely to view Russia as being responsible for the war in Ukraine (β = −0.363, p < 0.001). As Figure 4 illustrates, the analysis confirms all three mediation pathways.Footnote 56 First, Russian speakers’ more favourable views of the Soviet past directly reduce their willingness to fight for Latvia (β = −0.05, p < 0.001). Second, different threat perceptions are a significant mediator between Russian ethnicity and defence willingness: Russian speakers are less likely to blame Russia for the war in Ukraine, and this indirectly lowers their willingness to fight for Latvia (β = −0.074, p < 0.001). Third, we find support for a sequential mediation pathway: more positive historical perceptions of the Soviet past are associated with a lower tendency to attribute blame to Russia for the war, which subsequently results in a reduced willingness to fight for the country (β = −0.365, p < 0.001). The unmediated and direct effect of minority membership on our dependent variable is small and not significant (β = −0.052, p = 0.120). The findings remain robust when controlling for respondents’ age, gender, marital status, and income. All mediation pathways retain their negative direction and statistical significance.Footnote 57
Figure 4. Soviet nostalgia and threat perceptions as mediators of willingness to fight.
Our models provide robust empirical evidence that divergent historical memories and associated threat perceptions constitute key psychological mechanisms underpinning differences between majority and minority group members in their willingness to fight for and to defend their country in the case of attack. In Latvia, Russian speakers demonstrate a lower willingness to fight compared to Latvian speakers, a pattern closely linked to their more positive collective memories of the Soviet past. While Soviet nostalgia directly reduces the willingness to defend the country, it also indirectly shapes perceptions of external security threats. Specifically, Russian speakers are less likely to attribute responsibility to Russia for the war in Ukraine, which, in turn, further diminishes their willingness to fight for Latvia.
Conclusion
The Russian full-scale invasion of Ukraine in 2022 has attracted renewed attention to European defence capabilities. In this context, a growing body of research explores why citizens are motivated to defend their country and what factors influence these attitudes. While this literature has provided valuable insights, it has largely assumed societal homogeneity, overlooking the fact that many countries in Europe are ethnically diverse and marked by salient cultural cleavages. Minority groups may hold fundamentally different perspectives about fighting for a state dominated by the majority population. This ‘minority dimension’ represents a crucial blind spot in current assessments of national resilience and defence preparedness. To address this gap, our study focuses on individuals’ willingness to fight for their country in a crucial case. Latvia is a country not only with a significant Russian-speaking minority but also under constant pressure from its belligerent neighbour, Russia.
While from a comparative cross-case perspective, a considerable number of Russian speakers in Latvia express willingness to fight for their country, we observe a stark cultural cleavage between Latvian speakers and Russian speakers, which manifests in significantly different levels of willingness to fight for Latvia between majority and minority members in the case of a hypothetical attack. Importantly, these differences are primarily driven by cultural factors rather than socio-demographic or socio-economic variables. We argue that a key driver of this disparity lies in divergent historical narratives and collective memories regarding the Soviet past. While ethnic Latvians emphasise the suffering and repressions experienced under Soviet occupation, Russian speakers tend to highlight the perceived achievements of the Soviet Union, sharing a more positive view of this historical chapter. These contrasting historical memories directly shape present-day threat perceptions and influence how individuals interpret Russia’s belligerent actions. For Latvian speakers, Russia’s war of aggression against Ukraine represents an existential threat to their country, driving a strong willingness to defend Latvia in case of attack. In contrast, Russian speakers more often share Russia’s antagonistic stance towards ‘the West’ and, consequently, demonstrate a lower average level of commitment to fighting for Latvia.
Drawing on an original survey in the country, we find robust support for our hypotheses. Russian speakers in Latvia consistently demonstrate lower willingness to fight for the country compared to Latvian speakers. This ‘minority effect’ holds against different ways of assessing respondents’ minority status and remains substantial when controlling for key socio-demographic and socio-economic variables. Our analysis confirms the critical role of historicity and collective memories in shaping the (un-)willingness to fight for the country. Specifically, idealising the Soviet era emerges as a powerful predictor. Our mediation analysis shows that Russian speakers are more likely to hold positive views of the Soviet past and less likely to blame Russia for the war against Ukraine. Parallelly, Russia is actively promoting a romanticised view of the Soviet period. In line with our assumptions, the statistical models demonstrate that these divergent historical perceptions among minority members shape their threat assessments, ultimately resulting in reduced willingness to fight for the country.
Our empirical findings suggest that people’s (un-)willingness to fight for and defend the country is deeply influenced by cultural identities and culture-related factors. While culturally rooted historical narratives are an important element of identity construction in the first place, historical memories have far-reaching implications for attitudes and political behaviour. Given that over three decades have passed since the collapse of the Soviet Union, most respondents likely lack direct memory of that time, suggesting their views are shaped by romanticised historical memories or secondary accounts, including those promoted by Russia. The pattern of lower willingness among Russian speakers holds for all age groups in our sample.Footnote 58
Our study makes two significant contributions to the growing literature on individual-level determinants of national defence willingness. First, we underscore the critical importance of ethnic diversity in understanding national defence attitudes, demonstrating that the willingness to fight for the country may fundamentally differ across ethnic groups. Second, while previous research has predominantly concentrated on socio-demographics, socio-economic factors, group grievances, and material incentives, our work highlights the pivotal role of cultural factors in shaping the willingness to defend one’s country. In the regional context of our case study, these findings have important practical implications: since the onset of Russia’s war against Ukraine in 2014, the defence capabilities of states perceived as vulnerable to Russian interference have come under intense scrutiny – particularly the Baltic states, which were formerly occupied by the Soviet Union and, especially Latvia and Estonia, are marked by multi-ethnic societies with significant Russian-speaking minorities.
While our study makes significant contributions, it also has limitations. First, it examines a single minority group within a country that exhibits a notable ethnic divide and has a notably aggressive kin-state. Consequently, our findings might not be entirely generalisable to different contexts or other minority groups. This is a fruitful avenue for future research. Second, due to the absence of comparable data before 2022, we cannot evaluate how the willingness to defend the country has changed since the beginning of the Russo–Ukrainian war. Third, the study depends on self-reported measures of willingness to defend the country, which may not accurately reflect individuals’ true willingness to engage in combat. Specifically, respondents’ actions in a real war environment could vary significantly depending on the circumstances, such as the nature and success of the invasion and the extent of resistance encountered. Despite these limitations, our findings provide valuable insights for multi-ethnic societies marked by significant ethnic divisions, in both the global north and south. This is especially relevant amidst international crises and increased geopolitical tensions, as seen in the greater Middle East and the South China Sea. From a security standpoint, promoting minority engagement and loyalty is crucial for enhancing a country’s resilience and defence readiness.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/eis.2025.10026
Acknowledgements
The authors would like to thank the editors and reviewers for their helpful comments and suggestions.
Funding statement
This publication is supported by the project ‘Factors and Actors Contributing to the Willingness to Defend One’s Own Country: The Case of Latvia, Lithuania and Taiwan (WILLDEF)’, funded by Latvia–Lithuania–Taiwan Joint Fund for Scientific Cooperation (project agreement no. LV-LT- TW/2024/6) and the Grant “Why People Would (Not) Fight for Their Own Country in a War? NATO Member States at a Cross-Section” (Grant No. RSU-ZG-2024/1-0001), which is part of the project “RSU Internal and RSU with LASE External Consolidation” (Project No. 5.2.1.1.i.0/2/24/I/CFLA/005), funded by the European Union Recovery and Resilience Facility and the budget of the Republic of Latvia.
Ethical approval
The data collection was conducted as part of the project implemented by Riga Stradins University and supervised by the Latvian Science Council.
Data availability statement
We will publish an aggregated and anonymous version of the dataset as well as the do-files to allow replication.
Dr Felix Schulte is a Senior Researcher and Head of the Cluster ‘Conflict & Security’ at the European Centre for Minority Issues (ECMI) in Flensburg, Germany.
Dr Juris Pupcenoks is a Professor of Political Science & Chair of the Political Science Department at Marist University, USA.
Dr Māris Andžāns is an Associate Professor at Riga Stradins University and the Director of the Center for Geopolitical Studies in Riga, Latvia.
Open access
