A nostalgic narrative of ‘us’ and ‘them’

The antagonistic portrayal of social structures is arguably a central characteristic of the current climate of increased affective polarization. Affective polarization refers to the growing emotional divide between opposing political or ideological groups, characterized by heightened in-group favouritism and increased hostility towards out-groups (Iyengar et al. Reference Iyengar2019). Unlike ideological or political polarization, which focuses on disagreements over policies or ideologies, affective polarization captures the social and emotional dimensions of group dynamics (Dias and Lelkes Reference Dias and Lelkes2022; Hartman et al. Reference Hartman2022; Jost et al. Reference Jost, Baldassarri and Druckman2022). This phenomenon is evident in individuals’ negative perceptions of out-groups and positive bias towards in-groups, even when substantive ideological differences are absent (Mason Reference Mason2018). This distinction is crucial because it underscores the role of social identity in driving polarization independently of policy disagreements.

Classic partisan rivalries have evolved, with ideological identities becoming more prominent (Ellis and Stimson Reference Ellis and Stimson2012; Oshri et al. Reference Oshri, Yair and Huddy2022). Voters increasingly align themselves with ideological labels like ‘liberal’ or ‘left-wing’, rather than with specific political parties. These ideologies often divide citizens along lines similar to the Republican–Democrat divide in the United States, creating stark distinctions such as ‘right versus left’, with the latter being frequently characterized as ‘woke’ (Comellas and Torcal Reference Comellas and Torcal2023). In Europe, party affect primarily orbits around three ideological spheres: left, right and radical right (Bantel Reference Bantel2023; Kekkonen and Ylä-Anttila Reference Kekkonen and Ylä-Anttila2021). Additionally, specific issues can foster strong ‘us versus them’ dynamics, serving roles akin to partisan identities. Examples include the Brexit debate (Hobolt et al. Reference Hobolt, Leeper and Tilley2021; Tilley and Hobolt Reference Tilley and Hobolt2023), the controversy over Catalan independence (Balcells and Kuo Reference Balcells and Kuo2023) and positions regarding Covid-19 countermeasures (Bor et al. Reference Bor, Jørgensen and Petersen2023a). Other significant divides include pro- and anti-environmentalism, Nimbys against their adversaries (Devine-Wright Reference Devine-Wright2009) and conflicting stances on vaccination and abortion rights (Bor et al. Reference Bor, Jørgensen and Petersen2023b; Crawford et al. Reference Crawford2022; Wagner and Eberl Reference Wagner and Eberl2022). These identities amplify societal divisions, complicating the landscape of polarization.

In the expansion of antagonisms across social groups, populists find an opportune environment to thrive. Cas Mudde’s (Reference Mudde2004) ideational approach conceptualizes populism as a thin-centred ideology that seldom exists in isolation. Instead, it tends to affiliate itself with a broader host ideology, thereby displaying contextual sensitivity (Mudde and Rovira Kaltwasser Reference Mudde and Rovira Kaltwasser2013). This ideological framework creates a Manichaean tension within society, fundamentally dividing it into two homogeneous and antagonistic groups (Mudde Reference Mudde2004; Mudde and Rovira Kaltwasser Reference Mudde and Rovira Kaltwasser2013): the morally righteous ‘us’ juxtaposed against the evil ‘them’. Scholarly investigation often focuses on the populist dichotomy between the perceived homogeneous people and the perceived malevolent elite (for overviews, see Berman Reference Berman2021; Hunger and Paxton Reference Hunger and Paxton2022; Noury and Roland Reference Noury and Roland2020). Yet the adaptability inherent in the thin-centred populist ideology and its contextual sensitivity allows political leaders espousing populist rhetoric to broaden the definition of out-groups deemed malevolent, thus framing them as threats to their envisioned societal ideal. Traditionally, immigrants have been frequently depicted as an external source of malevolence, supposedly corroding the fabric of society (for overviews, see Shehaj et al. Reference Shehaj, Shin and Inglehart2021; van Spanje Reference van Spanje2011). However, in these times of intense group-based polarization, populists are employing new tropes to articulate their out-group narratives.

Populist politicians exhibit a propensity to delineate distinct in- and out-groups. Conversely, non-populist politicians utilize communication strategies to construct a narrative aimed at an ideal societal framework; however, the adaptability of such strategies is often constrained by ideological limitations in comparison to the more flexible approach of populists. Scholars specializing in populist communication have highlighted the prevalence of emotionally charged rhetoric in numerous communication studies (Aalberg et al. Reference Aalberg2016; Widmann Reference Widmann2021). This kind of rhetoric is designed to evoke profound emotional responses, aiming to resonate deeply with the audience while consistently emphasizing the sovereignty of the people (Schumacher et al. Reference Schumacher, Rooduijn and Bakker2022). Populist rhetoric inherently revolves around asserting the existence of a severe political crisis that poses an existential threat to ordinary citizens. This rhetoric vehemently vilifies the out-group, allocating blame to them (Schumacher et al. Reference Schumacher, Rooduijn and Bakker2022). Recent studies have underscored the efficacy of nostalgia as a potent vehicle employed within this narrative construction (Elçi Reference Elçi2022; Smeekes et al. Reference Smeekes, Wildschut and Sedikides2021; Van Prooijen et al. Reference Van Prooijen2022).

Populists need a compelling narrative to channel their antagonistic depictions, which they often achieve by contrasting the perceived decay of the present – where purported enemies abound – with an idealized, pristine past where such issues were non-existent, essentially employing nostalgia. Nostalgia, derived from the Greek words nostos (home) and algia (pain or sorrow), encapsulates a complex blend of sentiments, representing both a pain of loss and a repository of pleasant memories. Regardless of its more positive connotation as a pleasant memory or the poignant aspect as a pain of loss, nostalgia fundamentally revolves around the notion of loss (Pickering and Keightley Reference Pickering and Keightley2006). At its core, nostalgia can be understood as an individual-level emotion: a tender yearning for one’s past, usually focusing on the innocence of childhood or the formative years of adolescence (Marchegiani and Phau Reference Marchegiani and Phau2013). This psychological interpretation predominates in individualistic frameworks. However, within the realm of the social sciences, the emphasis often shifts to collective nostalgia. Defined as an emotive attachment or fond feelings associated with perceiving oneself as a member of a group, collective nostalgia is grounded in idealized perceptions of the history or past events of this group (Baldwin et al. Reference Baldwin, White and Sullivan2018). These events often have direct connections to the in-group’s shared experiences. Recollection of such events is commonly transmitted through collective memory, perpetuated within families, documented in literature or disseminated through mass-media narratives (Kao Reference Kao, Gürhan-Canli, Otnes and Zhu2012). The emergence of collective nostalgia is frequently linked to a perception among individuals that their group enjoyed a superior status or quality of life in the past compared to the present circumstances (Cheung et al. Reference Cheung2017). Despite its abstract nature, collective nostalgia tends to exhibit a remarkable consistency among individuals within the same cultural group and generation, reflecting shared experiences and cultural influences (Kao Reference Kao, Gürhan-Canli, Otnes and Zhu2012).

Nostalgia, with its intricate blend of sentimental reminiscence and a sense of loss, is intertwined with the dynamics of the in- and out-groups, delineated in the seminal work of Tajfel and Turner (Reference Tajfel, Turner, Austin and Worchel1979). The concept of collective nostalgia (Smeekes et al. Reference Smeekes, Verkuyten and Martinovic2015, Reference Smeekes, Wildschut and Sedikides2021; Van Prooijen et al. Reference Van Prooijen2022), shaped by idealized perceptions of the past, often finds resonance within group identities. These nostalgic yearnings, rooted in shared historical events and cultural experiences, can reinforce the distinctions between in-groups and out-groups. Nostalgia often implies more than a simple longing for the past – it also frames the present as a deviation from this ideal state. This deviation requires explanation, and scapegoating emerges as a complementary mechanism by which blame for societal decline is attributed to specific out-groups. By constructing these groups as obstacles to reclaiming the past, scapegoating translates the implicit contrasts of nostalgia into explicit narratives of blame. This process not only reinforces the in-group’s cohesion but also sharpens its boundaries by portraying out-groups as threats or antagonists. We argue that political elites make hay of the nostalgic discourse by framing these identities as either an in- or an out-group and implicitly or explicitly shifting blame to the out-groups for unwanted changes in society.

Thus, the interaction between nostalgia and scapegoating has significant implications for affective polarization (Iyengar et al. Reference Iyengar2019). Nostalgia intensifies emotional attachment to the in-group by evoking pride and solidarity, while scapegoating amplifies hostility towards out-groups by framing them as responsible for societal decline. Together, these mechanisms contribute to affective polarization by deepening emotional divides between groups, strengthening in-group favouritism and exacerbating out-group hostility (Iyengar et al. Reference Iyengar2019). Importantly, while nostalgia can function independently of scapegoating, its polarizing effects are potentially heightened when coupled with explicit out-group blame.

Building on our argument, we contend that nostalgic narratives are not merely rhetorical devices but pivotal tools in the populist strategy to deepen societal divisions, exploiting affective polarization by framing out-groups as obstacles to reclaiming an idealized past. These narratives wield significant emotional power, fostering a sense of in-group unity through shared memories of a glorified history while simultaneously amplifying hostility towards those cast as responsible for societal decline. By intertwining nostalgic appeals with scapegoating, populist actors transform collective longing into a mechanism of exclusion, reinforcing intergroup boundaries and intensifying animosities. Existing research on populist rhetoric has significantly advanced our understanding of how out-groups are constructed and mobilized, particularly in contexts where immigrants or elites are depicted as threats to societal cohesion (van Spanje Reference van Spanje2011). However, much of this work has centred on these traditional or static out-groups, often overlooking the fluid and evolving nature of these categorizations in response to shifting political and social dynamics. Recent developments illustrate how nostalgic narratives are increasingly coupled with scapegoating to target a broader spectrum of social groups (Balcells and Kuo Reference Balcells and Kuo2023; Bor et al. Reference Bor, Jørgensen and Petersen2023b; Wagner and Eberl Reference Wagner and Eberl2022). This adaptability allows populist rhetoric to reframe grievances and identify new out-groups, reinforcing affective polarization in diverse and evolving contexts. By examining the dynamic interplay between nostalgia and scapegoating, we highlight their dual role in shaping polarization. Together, these mechanisms extend beyond political rhetoric, deepening societal divisions and posing challenges to democratic cohesion. Understanding how these narratives are mobilized in evolving contexts provides critical insights into the mechanisms driving contemporary polarization. We therefore test the following hypotheses:

Data, measurement and method

Experimental set-up

To examine the affective consequences of nostalgic and scapegoating messages on social in-groups and out-groups, we conducted an experiment designed to cue participants with nostalgic values and scapegoats. The participants were randomly assigned to different conditions within this framework, as shown in Figure 1.

Figure 1. Overview of Experimental Procedure

The first step in the experimental set-up involved classifying participants into ideological groups based on their self-identified political orientation. Participants reported their ideological stance on a five-point Likert scale, indicating whether they considered themselves very progressive, progressive, centrist, conservative or very conservative. This classification was critical because specific objects of nostalgia which people look back on fondly are known to differ between ideological positions (Lammers and Baldwin Reference Lammers and Baldwin2018; Stefaniak et al. Reference Stefaniak2021). Individuals who identified as ‘centre’, ‘conservative’ or ‘very conservative’ were grouped into the ‘conservative’ category, while those who identified as ‘progressive’ or ‘very progressive’ were assigned to the ‘progressive’ category. While we acknowledge that ideological in-/out-grouping can occur at various levels of granularity (from party factions to broader ideological camps), our study focuses on the primary division between the progressive and conservative ideological spheres. This division is substantiated by research indicating that European citizens, even in highly fragmented contexts, tend to align affectively within two main blocks (Bantel Reference Bantel2023; Hagevi Reference Hagevi2015; Kekkonen and Ylä-Anttila Reference Kekkonen and Ylä-Anttila2021; Simón Reference Simón2020).

For each ideological group, we developed tailored messages to reflect the narratives most relevant to their respective ideological positions. These narratives were derived from two primary sources: open-ended responses collected earlier in the study and existing literature in sociology and marketing. The open-ended responses provided direct insights into the themes and symbols the participants associated with nostalgia (see Van der Velden et al. Reference Van der Velden2023), while the literature offered a broader theoretical foundation to shape these messages. Specifically, insights from studies on the modalities of nostalgia, the role of belonging and the sociocultural dimensions of nostalgic messaging informed the construction of the experimental stimuli (Hartmann and Brunk Reference Hartmann and Brunk2019; May Reference May2017; Pickering and Keightley Reference Pickering and Keightley2006).

The experimental design followed a 2 × 2 factorial framework. The messages varied along two key dimensions: whether they invoked nostalgia or not, and whether they included a scapegoating element or not. This set-up resulted in four distinct experimental groups, allowing us to disentangle the effects of nostalgia and scapegoating independently and in combination. The participants of the two groups were presented with messages specifically tailored to resonate with their ideological perspective (right-wing conservative or left-wing progressive). These messages emphasized either nostalgia or scapegoating or both nostalgia and scapegoating or neither nostalgia nor scapegoating, depending on the assigned condition.

To ensure the robustness of the findings, additional analyses were conducted to account for the potential influence of the participants who identified as centrists. These analyses involved running the data both with and without centrists, and the results demonstrated that the inclusion or exclusion of this group did not substantively affect the outcomes (see the Supplementary Material, Online Appendix C). For further details about the experimental design and procedures, see the Supplementary Material, Online Compendium and Online Appendix A.

We randomly assigned the participants to see just one of four different potential campaign messages (see Figure 1), telling them that a party close to their position was testing campaigning material for the next Dutch elections. Figure 2 demonstrates that the participants who voted for more conservative parties (as ranked by Kieskompas), displayed in the top lines of Figure 2, also positioned themselves as more conservative, and the voters who supported more progressive parties positioned themselves as more progressive. An example of the treatment text for a left-wing voter, including nostalgia and a scapegoat, reads as follows (see Online Appendix A for all the treatment texts):

Looking back, we remember a time when solidarity, workers’ rights and respect for the environment were paramount. People lived a slow and peaceful life in harmony with nature, valuing community over individualism. Today, we see a world plagued by pollution, polarization and stress, caused in large part by large corporations, contamination lobbyists and neoliberal policymakers. It’s time to hold them accountable again for their actions and reclaim those values of solidarity, respect and peace that made our world a better place. Let’s create a world where workers’ rights are protected again, the environment is re-cherished, and the well-being of all people is prioritized again over the interests of a select few.

Figure 2. Baseline of Self-Placement into Ideological Positions

Affective consequences of a nostalgic narrative

First, we delve into the descriptive analysis of attitudes towards social groups across political ideology. Figure 3 shows the attitudes towards social groups held by members of conservative and progressive groups, which provide a baseline for the subsequent analysis. It indicates a clear pattern of affective polarization: the left panel shows clearly that right-wing voters (pink dots) are rated most negatively by individuals identified as ‘very progressive’ and most positively by individuals identified as ‘very conservative’. Similarly, the right panel shows that left-wing voters (green dots) are rated most positively by individuals identified as ‘very progressive’ and most negatively by individuals identified as ‘very conservative’. This aligns with the definition of affective polarization (Iyengar et al. Reference Iyengar2019), as the conservative participants evaluated their in-group (i.e. right-wing voters) more positively than their out-group (i.e. left-wing voters), while the progressive participants evaluated their in-group (i.e. left-wing voters) more positively than their out-group (i.e. right-wing voters).

Figure 3. Descriptive Analysis of Social Groups

Beyond the left–right distinction, our examination of the sentiments of right-wing conservative supporters towards other social groups revealed a similar trend: greater warmth towards Christians, multinationals (many of which have a significant presence in the Netherlands), neoliberal politicians and fellow right-wing voters. However, nuanced variations exist within these sentiments. Notably, the most linear variation in warmth towards right-wing voters can be seen as one moves from left-wing progressive party supporters to right-wing conservative party supporters. It is noteworthy that Christians garnered a generally positive sentiment across the political spectrum. Additionally, materialistic individuals, neoliberal politicians and multinationals did not receive uniformly positive evaluations, even from right-wing party supporters.

Turning our attention to the right side of Figure 3, a similar narrative unfolds. The more politically left-leaning the participants, the more positive their sentiments towards social groups associated with progressive ideologies. Notable entities receiving favourable evaluations include climate activists, immigrants, Muslims, individuals living on benefits and other left-wing voters. However, nuances in sentiment exist within this category as well. Some groups, such as individuals living on benefits, received relatively neutral evaluations from right-wing supporters. Conversely, climate activists, immigrants, Muslims and left-wing voters were viewed with varying degrees of antipathy by the right.

Figure 4 shows the distribution and average affective sentiment towards the social groups of the message per experimental treatment. Social in-groups were more liked than out-groups – with scores of five and lower for the disliked social group and six and higher for the liked social group. The distributions indicate that, on average, the progressive voters liked their in-group more than the conservative voters liked their in-group, regardless of the experimental treatment. However, there was no difference between progressive and conservative voters regarding their social out-groups: they both disliked them equally. Again, there was little variation in levels of support for the social out-group between the messages. This analysis establishes a crucial baseline for our subsequent examination of the effects of the treatment, as it outlines the reference point for changes in sentiments that may arise.

Figure 4. Affective Sentiment for Social Groups

To test whether a nostalgic narrative of ‘us and them’ has polarizing consequences, we present the results of our pre-registered experiment. The left-hand panel in Figure 5 demonstrates the support for social in-groups based on the treatment message. In general, neither the nostalgic messages (H1a) nor the scapegoating messages (H1b) resulted in a statistically significant increase in affective sentiment for the social in-groups. While the effect of the scapegoating messages (H1b) went in the hypothesized direction, it was not statistically significant. The nostalgic messages actually did elicit a statistically significant effect, but in the opposite direction to the hypothesis (H1a): being exposed to a nostalgic message (versus a non-nostalgic message) made the participants feel colder towards their in-group. Hence, we did not find any support for H1a or H1b. The analysis of our pre-registered experiment, as depicted in Figure 5, aimed to assess whether a nostalgic narrative framing of ‘us’ and ‘them’ engenders polarizing consequences, particularly concerning the support for social in-groups.

Figure 5. Predicted Affective Sentiment for Social Groups

To look deeper into this, we ran separate analyses for progressive and conservative voters per social group and found that the progressive voters were an exception (see Online Appendix C). When progressive voters received a nostalgic scapegoating message, although this did not materially affect their feelings towards groups such as left-wing voters and immigrants, it did seem to make them feel more negative about Muslims and more positive about climate activists. This could indicate that while both groups are generally seen more positively by progressives than by conservatives, climate activists are seen as an in-group that resists nostalgic and scapegoating pushes, whereas Muslims are probably seen more as strategic allies for whom at least some voters of the left signal instrumental and fragile positive feelings. Conservative voters also showed an interesting pattern. Under the nostalgic message treatment, especially when it also included scapegoating, they had a less positive response to neoliberal politicians, multinationals and materialistic people. This pattern could indicate that even if these groups are not necessarily out-groups for conservative voters, such voters may still see them as partly responsible for the decline perceived through their nostalgic lens. The implications of this pattern underscore the nuanced dynamics at play within the conservative voter demographic. In a broader context, the findings shed light on the multifaceted nature of conservative voter attitudes.

In conclusion, the lack of support for H1a and H1b in our study, as well as the subgroup analyses, provide valuable insights into the nuanced dynamics of political messaging and its differential impact on diverse voter groups. It is crucial to recognize these subtleties in order to understand the complexities of political sentiment and to avoid broad generalizations about the existence of a populist Zeitgeist based on the examined experimental conditions.

The right-hand panel in Figure 5 demonstrates support for social out-groups after exposure to nostalgic and scapegoating messages. The nostalgic messages (H2a) had no effect on how the participants felt towards their social out-groups: the coefficient is zero. The scapegoating messages (H2b), however, did have the hypothesized polarizing effect: they made the participants colder towards their out-groups. This is an indication that a ‘purely’ nostalgic message could foster more understanding towards some ideological out-groups and that any scapegoating element in a nostalgic narrative could diminish any positive affective sentiment towards the out-groups achieved by the nostalgic appeal. In essence, this suggests that people feel more affectively polarized towards their out-groups when they are exposed to a message that puts the blame on someone else.

Online Appendix D presents the subgroup analyses for conservative and progressive voters. The data show that the nostalgic messages (red coefficients) were associated with an increased affective sentiment towards right-wing voters and Christians (the latter without statistical significance). Combining the nostalgic message with scapegoating (green coefficients) decreased the affective sentiment towards out-groups: the progressive voters became more negative towards right-wing voters (without statistical significance), neoliberal politicians, multinationals and Christians (without statistical significance). Only a scapegoating message (in blue) decreased the sentiment towards right-wing voters, neoliberal politicians, multinationals, Christians and materialistic people – with only the latter being statistically significant.

Hence, while our data do not support H2a, they do support H2b. The subgroup analyses show a more nuanced pattern: nostalgic and scapegoating messages decreased the affective sentiment towards some, but not all, of the groups theorized as out-groups. In light of these results, it is essential to acknowledge the complexity of the impact of political messaging on affective sentiment towards social out-groups. Our study does not decisively indicate the presence of a populist Zeitgeist based on the examined experimental conditions. The nuanced and varied responses observed underscore the importance of considering contextual factors and the diverse nature of political attitudes within the broader population. We explored whether our results are conditional upon ideology, party support, populist attitudes, education and gender, but we did not find evidence for a relationship between nostalgia or scapegoating and affective polarization.