The Politics of Imperialism in Britain

While the British Empire shaped the world, it remained surprisingly absent as a major political cleavage within Britain itself. The Conservative Party strongly supported imperialism, declaring as late as 1950 that it was ‘the party of Empire’ and ‘proud of its past’, and viewed the empire as ‘the surest hope in our day’ (cited in Barnes Reference Barnes, Seldon and Ball1994, 337). In contrast, the Labour Party’s position was complex and ambivalent. While anti-imperialist sentiment began to emerge in the British left in the 1920s and 1930s, these voices had little influence on the parliamentary Labour Party (Rich Reference Rich1990). The Labour government that came to power after World War II did accept the independence of Asian colonies such as India and Burma, but this was a pragmatic rather than ideological consideration. Indeed, the African colonies, in particular, were intended to remain under British tutelage (Hyam Reference Hyam2006). Attention also shifted to the Commonwealth, an effort to maintain British influence in former colonies (Patel Reference Patel2021). In summary, throughout twentieth-century British politics, imperialism consistently enjoyed Conservative support and sympathy, while Labour’s position was conflicted, buttressed by anti-imperialist thought and Cold War realism, but also – as critics such as George Orwell argued – the dependence of working-class prosperity in Britain on continued colonial exploitation (Howe Reference Howe1993).

Although the British Empire remains largely absent in contemporary political discourse and campaigns, its imagery and symbolism is sometimes invoked by politicians, particularly from the right. Boris Johnson often used colonial references, for example, reciting Kipling’s colonial paean ‘Mandalay’ whilst in Myanmar. Rory Stewart, a former MP from the left of the Conservative party, developed a persona that Mitchell (Reference Mitchell2021) describes as mimicking the ‘imperial patrician tradition’ of TE Lawrence (that is, ‘Lawrence of Arabia’). Conservative MP and party leadership candidate Robert Jenrick made a more overt defence of empire in a 2024 op-ed: ‘many of our former colonies – amid the complex realities of Empire – owe us a debt of gratitude for the inheritance we left them’ Jenrick (Reference Jenrick2024). While imperialism is not a major axis of political contestation in contemporary Britain, symbolic references do surface, indicating that some politicians see value in mobilising imperial themes.

Historical research has identified imperial legacies across multiple policy domains, with particularly extensive work examining international relations and immigration policy. In international relations, Gildea (Reference Gildea2019) argues that an imperialist mode of thinking was used to justify interventions in Afghanistan, Iraq, and Syria. Other scholars (Bell Reference Bell2003, Reference Bell2020; Kenny and Pearce Reference Kenny and Pearce2018; Vucetic Reference Vucetic2011) have identified an ideational frame, the ‘Anglosphere’, which has circulated since the imperial era and focuses on perceived cultural and racial bonds between Britain and its former settler colonies. This has been used by elites to advance alternative visions of Britain’s global role. Related to this are arguments that Britain’s post-war immigration policy has reflected colonial ideas of racial hierarchy in its preferential treatment for white Commonwealth citizens (Patel Reference Patel2021).

In contrast, the British Empire’s legacy for democratic mass politics remains less explored. While both Gildea (Reference Gildea2019) and Kenny and Pearce (Reference Kenny and Pearce2018) draw links between Brexit and imperial ideas, the focus is on elite discourse rather than the attitudes and behaviours of the public. Indeed, more generally, while the positions of political elites on the British Empire are well-documented, public opinion towards empire remains far less understood. Although some historians have argued that British society was strongly imperialist in sentiment (for example, Hall Reference Hall2002), and the ‘loss of empire came as a profound psychological shock’ (Rich Reference Rich1990, 11), others claim widespread public indifference (Porter Reference Porter2004). Behind these debates lies an unavoidable uncertainty regarding historical public opinion, due both to the limitations of the historical lens and the paucity of survey research on the topic.

Surprisingly, the absence of research on public opinion on empire continues to the present day. Contemporary analyses of imperial nostalgia remain conceptual and impressionistic rather than empirically grounded (for example, Mitchell Reference Mitchell2021). Available empirical evidence is limited to sporadic surveys conducted by commercial pollsters. For instance, a 2014 YouGov survey revealed that 59 per cent of Britons viewed the empire favourably, compared to only 19 per cent who felt ashamed (Dahlgreen Reference Dahlgreen2014). A similar question asked in 2019 found lower levels of pride (32 per cent), but the same level of shame (this question also included a ‘neither’ option) (Smith Reference Smith2020). More recent polling in 2024 suggests continued polarisation, with 33 per cent viewing the empire as something to be proud of and 21 per cent as something to be ashamed of (Smith Reference Smith2025).

Cross-national data from the same 2019 YouGov poll conducted across several European countries suggests that British opinion is somewhat more positive towards empire compared to other former colonial powers, but not uniquely so: 32 per cent of the British public felt pride in their empire, ranking second only to the Netherlands (50 per cent) among European colonial powers, and well above France (26 per cent), Belgium (23 per cent) and Spain (11 per cent) (Smith Reference Smith2020). These more positive views found in Britain align with comparative historical research arguing that Britain’s relatively seamless transition from empire enabled more positive imperial narratives to persist in national consciousness (Buettner Reference Buettner2016). What comparative evidence we have thus positions Britain among the more empire-positive former colonial powers, whilst suggesting that imperial nostalgia represents a broader phenomenon across post-imperial societies with varying intensity depending on particular historical experiences of imperial competition and decolonisation.

Despite this comparative context, the true extent and nature of British imperial attitudes remains uncertain, and whether they are politically consequential is entirely unknown. The objective of this paper is to address these issues. First, we describe how we conceptualise public attitudes towards empire and theorise their connections to political preferences.

Conceptualising Imperial Nostalgia

Understanding public attitudes towards empire requires drawing on three interconnected theoretical traditions. First, we conceptualise imperial attitudes as a form of collective memory, which is a shared narrative about a nation’s history. However, the moral complexity of imperial histories – encompassing national greatness on one hand and violence, oppression and expropriation on the other (Gildea Reference Gildea2019) – means that public reactions likely include both positive and negative evaluations. We therefore engage with two additional literatures: research on political nostalgia to understand favourable imperial attitudes and work on memories of historical violence to conceptualise hostile reactions to imperial legacies.

Collective memories are ‘memories that are shared by a group and that are of central importance to the group’s identity’ (Abel et al. Reference Abel, Umanath, Wertsch and Roediger2017, 290). They function more as ‘myths’ than literal memories, however, as they are reimagined and reinvented across time (Verovšek Reference Verovšek2016).Footnote ¹ Moreover, collective memories tend to be expressed through narratives that tell a story of the nation – who it is, how it came to be, and what it values (Bell Reference Bell2003). In this way, they can be seen as a form of political culture, in that they are ‘concerned with the cultural constitution of political identities and activities’ (Olick Reference Olick1999, 336–7). Imperial eras are particularly likely to feature as collective memories because they embody both national glory, through imperial dominance, and national trauma, through the loss of empire – qualities that Volkan (Reference Volkan2001) identifies as central to the construction of group identity.

To link these collective memories of empire to political preferences and behaviour, we turn to the concept of collective (or political) nostalgia. This is the belief that one’s group experienced a golden age that has now been lost (Tannock Reference Tannock1995; Wildschut et al. Reference Wildschut, Sedikides, Arndt and Routledge2014). Nostalgia is an affectively charged orientation, infused with wistfulness, melancholy, and fondness for the past. This affective charge allows nostalgia to transform beliefs about the past into a motivating force that shapes how individuals and groups understand their identities and political priorities. The concept of nostalgia therefore allows us to link evaluations of the past with preferences in the present.

The concept of nostalgia is increasingly used by social scientists to describe and explain various forms of the politics of the past. Most prominently, researchers have used nostalgia to explain electoral choice in Europe and the United States, especially for right-wing populist parties (Gest et al. Reference Gest, Reny and Mayer2018; Goidel et al. Reference Goidel, Goidel and Kellstedt2024; Lammers and Baldwin Reference Lammers and Baldwin2020; Smeekes et al. Reference Smeekes, Wildschut and Sedikides2021; Smeekes and Lubbers Reference Smeekes and Lubbers2024; Steenvoorden and Harteveld Reference Steenvoorden and Harteveld2018). This research typically focuses on nostalgia for some halcyon period within living memory, that is, the ‘good old days’ (Smeekes et al. Reference Smeekes, Wildschut and Sedikides2021), and explains its political effects through relative deprivation: negative appraisals of present circumstances compared to a fondly remembered past (Gest et al. Reference Gest, Reny and Mayer2018; Versteegen Reference Versteegen2024).

A complementary literature examines how political actors across the political spectrum deploy nostalgic rhetoric in campaigns (Bonikowski and Stuhler Reference Bonikowski and Stuhler2022; Elçi Reference Elçi2022; Menke and Wulf Reference Menke and Wulf2021). This research demonstrates nostalgia’s mobilising power when elites frame political appeals around idealised pasts, suggesting that nostalgic sentiments can be both measured as mass attitudes and activated more implicitly through elite communication – a distinction relevant to understanding how imperial nostalgia might operate both as a latent public opinion and as a mobilisable set of symbols.

Related is work on ‘authoritarian nostalgia’ in Eastern Europe and East Asia, which highlights how former authoritarian regimes are viewed positively by many citizens shortly after democratic transitions (Boyer Reference Boyer2006; Kim-Leffingwell Reference Kim-Leffingwell2024; Neundorf et al. Reference Neundorf, Gerschewski and Olar2020). Nostalgia is generally not directly measured in this literature, but is inferred as a mechanism (cf. Kim-Leffingwell Reference Kim-Leffingwell2024). But relative deprivation appears to be a key factor in these circumstances, with dissatisfaction with present circumstances coupled with relatively positive appraisals of previous regimes.

Imperial nostalgia has received far less attention from researchers who measure public opinion and test its political implications. Elçi (Reference Elçi2022) presents Turkish respondents with experimental treatments manipulating nostalgia for either the Ottoman Empire or the secular Kemalist period, finding that Ottoman nostalgia increases populist attitudes while Kemalist nostalgia does not. Bizumic and Duckitt (Reference Bizumic and Duckitt2018) develop a six-item battery measuring pro-British Empire attitudes, though this was used as a criterion variable for validating their authoritarianism scale rather than a subject of inquiry on its own terms.

Beyond this empirical gap, existing literature also lacks a clear concept of imperial nostalgia as distinct from general political or authoritarian nostalgia. Nostalgia for imperial pasts shares some features with these other forms, but differs in two important ways. First, imperial eras are beyond the personal experience of most, if not all citizens. Relative deprivation between the imperial era and the present is therefore unlikely to be the psychological mechanism at play, as it may be for other forms of nostalgia. Rather, imperial attitudes reflect mythologised national narratives about periods beyond living memory rather than lived comparisons. Second, the moral complexity of imperial histories likely generates far more ambivalence than nostalgia for recent political periods, creating a need to reconceptualise how we measure and understand these attitudes.

This moral complexity stems from the inherently violent and coercive nature of imperial projects. Researchers have examined European public reactions to colonial atrocities (Leach et al. Reference Leach, Branscombe, Wohl, Newman and Erber2013; Licata and Klein Reference Licata and Klein2010), as well public responses to other forms of historical violence perpetrated by the national ingroup, such as antisemitic pogroms (Charnysh Reference Charnysh, Jeffrey, Subotić and Welch2023) and the subjugation of native peoples in North America and Australia (Gunn and Wilson Reference Gunn and Wilson2011; Rotella and Richeson Reference Rotella and Richeson2013; Williams Reference Williams2000). This body of work demonstrates how people avoid accepting guilt for their forebears’ actions (Leach et al. Reference Leach, Branscombe, Wohl, Newman and Erber2013; Wohl et al. Reference Wohl, Branscombe and Klar2006) and seek to deny or minimise historical atrocities (Bonnot et al. Reference Bonnot, Krauth-Gruber, Drozda-Senkowska and Lopes2016). Presenting a national history as morally fraught challenges individuals by threatening deeply held national identities (Branscombe and Wann Reference Branscombe and Wann1994; Leach et al. Reference Leach, Branscombe, Wohl, Newman and Erber2013) and provoking cognitive dissonance (Charnysh Reference Charnysh, Jeffrey, Subotić and Welch2023). Indeed, while colonial violence is now widely condemned in principle, open acknowledgements of colonial violence or criticisms of a nation’s imperialism have provoked vigorous reactions in the past (Buettner Reference Buettner2016).

Given this complexity, we propose that imperial attitudes run from positive views and emotions about empire (nostalgia) to negative views and emotions (aversion), with indifference and ambivalence occupying a midpoint between these poles. The antithesis of nostalgia is therefore not mere indifference, but affectively charged, hostile views of national myths, such as feeling ashamed of the British Empire. This conceptual framework suggests that while some may view the imperial past nostalgically, few would desire empire’s actual return, and many others may feel genuine aversion to these same historical periods.Footnote ²

As we have established, attitudes towards empire are best understood as a form of collective memory: they are narratives or myths that are as much about the nation as they are about the past. The concept of nostalgia provides a way to link these national narratives to contemporary political preferences, building on recent research showing nostalgia’s growing influence in electoral choice across Western democracies. Like general nostalgia, there are two sides to imperial nostalgia that can be analysed: the beliefs and attitudes of the mass public, and the rhetoric and language of elite campaigns. However, the moral complexity of imperial histories requires expanding beyond nostalgia alone to encompass the full spectrum from nostalgic to aversive reactions.

Data and Research Design

Our data come from an original panel survey conducted by YouGov with a sample of adult residents of Britain (excluding Northern Ireland).Footnote ³ We adopt two complementary approaches to understand the nature and consequences of imperial nostalgia among the British public and test the political salience of elite rhetoric on imperial themes.

First, we measure imperial nostalgia using two batteries of survey items to examine whether the British public hold coherent and meaningful attitudes towards the British empire.Footnote ⁴ We assess the psychometric properties and consistency of our measures, then examine their associations with evaluations of political parties and electoral choice. This observational approach allows us to determine whether imperial attitudes represent a distinct dimension of public opinion and whether these orientations are politically consequential for understanding party support and voting behaviour.

Second, we designed a conjoint experiment to examine the political salience of elite positions on empire by testing how candidates’ nostalgic versus critical stances influence respondents’ electoral preferences.Footnote ⁵ This experimental approach allows us to test whether political rhetoric about the imperial past shapes electoral preferences and how the effects of this rhetoric vary depending on respondents’ pre-existing attitudes. Together, these approaches provide both measurement validation and causal inference about the political consequences of imperial nostalgia in contemporary British politics.

Patterns of Imperial Nostalgia in the British Public

We measure public views of empire using two batteries of questions. The first asks respondents to evaluate the British Empire using seven questions, drawn in part from the questions fielded by YouGov (Smith Reference Smith2020) and Bizumic and Duckitt (Reference Bizumic and Duckitt2018).Footnote ⁶

The second battery measures respondents’ emotional reactions to the British Empire. Respondents are asked ‘to what extent’ they feel each of ten emotions ‘when you think about the British Empire’: embarrassment, shame, guilt, anger, pride, nostalgia, sadness, disgust, happiness, and satisfaction.Footnote ⁸

Our data show that opinions are divided regarding the British Empire. Significant minorities support both pro- and anti-imperial positions on most of the seven items (see Figure 1). On only two questions does opinion clearly fall in one direction, and it is an anti-imperial direction in both cases: majorities agree that the British Empire did commit atrocities and majorities disagree that they wish for Britain to still have an empire.

Figure 1. Responses to the imperial attitudes questions.

Note: Each bar shows the weighted distribution of responses for one of the seven questions in the imperial attitudes battery using the second-round survey. The net percentage of the sample offering a nostalgic (anti-imperial) view is shown in the panel on the left. See the supplementary materials for the respective figure using the third-round data.

These findings reveal the ambivalence in attitudes to empire that we theorised earlier, with substantial constituencies holding both positive and negative views of empire depending on which aspects are emphasised. The net percentage of respondents offering a pro-British Empire opinion varies dramatically from −50 per cent when asked about responsibility for ‘atrocities’ to +21 per cent regarding ‘civilising effects’ in the second-round survey (−52 per cent and +28 per cent respectively in the third round). This considerable variation demonstrates that imperial attitudes are strongly shaped by framing – here, question wording.

Consistent with this result, respondents also display a significant level of uncertainty in their opinions regarding empire. Between a quarter and 40 per cent of respondents selected the ‘neither agree nor disagree’ or ‘don’t know’ options. This can be seen more clearly in the second battery, on emotional reactions to the British Empire (see Figure 2), where pluralities select ‘not at all’ in response to the questions.

Figure 2. Responses to the imperial emotions questions.

Note: Each bar shows the weighted distribution of responses for one of the ten questions in the imperial emotions battery using the second-round survey. The stem of the question read ‘When you think about the British Empire, to what extent do you feel…?’ The percentage of the sample holding a pro- or anti-imperial emotion – defined as selecting the response options a ‘moderate’ or ‘great’ extent – is shown in the panel on the left. See the supplementary materials for the respective figure using the third-round data.

Figure 3 illustrates the overall direction of attitudes and emotions regarding the British Empire. Across all questions, net opinion is slightly unfavourable towards the empire, with anti-imperial attitudes and emotions outnumbering pro-imperial ones. However, anti-imperial views are never close to a majority position when considered across our seven-item battery. Many respondents hold ambivalent or neutral views, accounting for over 40 per cent on the attitudinal battery and nearly 30 per cent on the emotional battery.Footnote ⁹ Despite this, we find that a significant portion of the British public is nostalgic for the British Empire: in the second-round survey, 24 per cent reported net nostalgic attitudes, and 34 per cent expressed net nostalgic emotions. These proportions remained consistent in the third round, at 23 per cent and 34 per cent, respectively.

Figure 3. Overall opinion regarding empire.

Note: Each panel shows the average direction of opinion across the items in the attitudes and emotions batteries as measured in the second-round survey. Neutral values are defined as 2.5 to 3.5 on the imperial attitudes scale (range 1–5) and −0.5 to 0.5 on the emotion difference scale (range −3–3). The figure uses composite mean scales for illustration, whereas confirmatory factor analysis (CFA)-derived scales (see supplementary materials) are used elsewhere in the paper.

Finally, in Figure 4, we illustrate how imperial nostalgia, averaged across all seven questions in the attitudinal scale, varies across demographic and geographic groups. We see that imperial nostalgia increases steadily and significantly with age while decreasing markedly with education. Men are more nostalgic than women, and White British respondents are more nostalgic than non-White respondents, though ‘Other White’ respondents are the least nostalgic overall. Geographically, residents of London and Scotland exhibit the lowest levels of imperial nostalgia, with the latter echoing the more critical views of the Belgian empire in Flanders as compared with Wallonia (Buettner Reference Buettner2016). In summary, imperial nostalgia shows demographic patterns consistent with it being a form of cultural conservatism: respondents are more nostalgic of empire to the extent that they are older, male, White British respondents, have lower levels of education, and live outside London and Scotland.Footnote ¹⁰ In the next section, we examine the convergent and divergent validity of our imperial nostalgia measures by testing whether they represent a distinct political orientation or simply reflect broader patterns of cultural conservatism.

Figure 4. Group differences in overall imperial nostalgia.

Note: The dots show the average level of imperial attitudes within the respective demographic group, pooled across both waves and weighted, and the horizontal bars indicate the 95 per cent confidence intervals. The shaded regions show the weighted distributions of the data within each subgroup. A composite mean scale is used in this figure. See the next section for psychometric details and the CFA-based scale used in subsequent analyses.

The Distinctiveness of Imperial Nostalgia

Having explored the patterns and distributions of our imperial attitude and emotions batteries, we consider now whether and to what extent we can speak of imperial nostalgia as a dimension of British public opinion. Our analysis of the psychometrics of the two scales (see supplementary materials) suggests that both are reliable and valid, but with the imperial attitudes scale having slightly better measurement properties, we consider it in isolation from this point forward.

To test the convergent and divergent validity of our imperial attitudes battery, we employ exploratory factor analysis (EFA). In addition to our seven items, we include thirty-six other items that measure other political attitudes and values, including left-economic and authoritarian-libertarian values, hostile sexism, immigration opinion, support for liberal democracy, populist attitudes, national pride, and chauvinistic nationalism. The results are reported using a heatmap in Figure 5.

Figure 5. Heatmap of exploratory factor analysis loadings.

Note: The heatmap shows the loadings from an exploratory factor analysis (EFA) of multiple attitudinal survey items. Eight dimensions are used, as indicated by a parallel analysis. EFA employs minimum residual estimation, promax rotation and pairwise polyserial correlations. Only loadings ≥ |0.30| are reported in this figure. Most items are from round two, except national pride and chauvinistic nationalism (round three).

We see that a single dimension of imperial nostalgia emerges clearly and distinctively in the EFA, with minimal overlap with other items measuring cultural conservatism or national pride. At the level of the factors, modest to strong correlations emerge between the imperial nostalgia factor and the national pride/chauvinism (r = 0.60), authoritarian values (0.63), hostile sexism (0.51) and left-economic values factors (−0.55). This suggests that while imperial nostalgia is related to other political attitudes – particularly those measuring cultural conservatism – it represents a distinct and coherent construct. Notably, its separation from measures of national pride and chauvinism indicates that nostalgia for empire is not reducible to a broader sense of patriotic attachment but instead taps into a specific ideological perspective.

Nostalgia and Party Choice

We now turn to an examination of the links, if any, between imperial nostalgia and electoral preferences. We begin with a heatmap showing bivariate correlations between respondents’ support for each of the major parties and various important political attitudes, including our measures of imperial nostalgia (Figure 6).

Figure 6. Correlates of party evaluations.

Note: Cells show the bivariate correlation between opinions listed in rows and self-assessed likelihood of ever voting for the party listed in columns. Darker blue cells indicate stronger absolute correlations. The column labelled ‘Mean’ shows the mean absolute correlation across the five party support items.

Imperial nostalgia has moderate correlations with support for all parties. As one might expect, it is negatively associated with support for parties of the left (Labour, Liberal Democrats and Greens) and positively associated with support for the right (Conservatives and Reform). Perhaps more surprising is the strength of the associations: imperial nostalgia is overall the second strongest correlate among all the covariates we include, after preferences regarding relations with the EU. Imperial nostalgia is more strongly correlated with party evaluations, on average, than such well-established predictors of party preference as left-economic and authoritarian values and immigration opinion. This result is intriguing given the minimal attention the British Empire received in electoral campaigns compared to issues like taxation, spending, and immigration.

We turn to regression models to further examine the relationship between imperial nostalgia and electoral preferences. Table 1 presents seemingly unrelated regression (SUR) models predicting respondents’ 11-point ratings of their likelihood to vote for the five most popular British parties. Imperial nostalgia remains a significant predictor of evaluations for three of the five parties we consider. It has a stronger association with support for the Conservatives than support for Labour, which is consistent both with the Conservatives’ long-standing pro-empire stance and Labour’s more ambivalent position. However, imperial nostalgia is an even stronger predictor of support for challenger parties on both the right (Reform) and the left (Greens), highlighting its potential as a disruptive dimension of political contestation. Since both parties position themselves as alternatives to the political establishment, imperial nostalgia may become an increasingly salient political force if these parties continue to grow in prominence.Footnote ¹¹

Table 1. Party support regressions

*p < 0.05. Seemingly unrelated regression results using wave two data, with weights applied and standard errors in parentheses. Error terms allowed to correlate across models. Missing values are imputed using FIML. Models also include an intercept, political attention, age, ethnicity, gender, education, religion, social grade, home ownership and region.

In Table 2 we include a measure of general nostalgia: a question asking ‘for people like me, life in our country is better today than it was fifty years ago’.Footnote ¹² The associations between imperial nostalgia and party evaluations remain similar to those presented in Table 1 when this potential confound is included, although the nostalgia-party evaluation associations are slightly stronger for the Conservative Party and much weaker – effectively zero – for the Labour Party. These findings suggest that, while some overlap exists between general and imperial nostalgia, the latter retains substantial predictive value in understanding party support across Britain’s political landscape.

Table 2. Including general nostalgia

*p < 0.05. Seemingly unrelated regression results using wave two data, with weights applied and standard errors in parentheses. Error terms allowed to correlate across models. Missing values are imputed using FIML. Models also include the covariates listed in Table 1.

Table 2 furthermore shows that general nostalgia and imperial nostalgia have different patterns of association with party evaluations. While imperial nostalgia is positively associated with both Conservative and Reform parties, general nostalgia is negatively associated with Conservative approval but positively linked to Reform support. This echoes existing literature demonstrating that general nostalgia is particularly associated with radical right support (Gest et al. Reference Gest, Reny and Mayer2018; Steenvoorden and Harteveld Reference Steenvoorden and Harteveld2018). Imperial nostalgia, in contrast, demonstrates broader mainstream appeal spanning both the conventional right (Conservatives) and also the Greens. These divergent patterns underscore that imperial nostalgia operates as a distinct political orientation.

Finally, we analyse respondents’ preferred choice of party (rather than their voting likelihood ratings for all parties, which we used previously). Party choice is measured using two items asking respondents whether they are likely to vote if a general election were to be held tomorrow and, if so, which party they would choose. We combine these data to create a single qualitative variable with eight party choices (including an ‘other’ option) as well as a ninth category indicating if a respondent would not vote.

We fit a random forest predictive model to this variable, including as features all our attitudinal covariates, drawn from all three survey waves, as well as the set of demographic variables we have available. Random forests aggregate an ensemble of decision trees that each analyse a random subset of variables. They are particularly well-suited for predicting party choice, as they allow for the modelling of complex, interactive, and non-linear relationships that traditional regression methods may not detect (Montgomery and Olivella Reference Montgomery and Olivella2018). We focus here on the impact of our set of covariates on the overall model’s predictive accuracy (that is, as indicated in the variable importance estimates in Figure 7). These show the amount the model’s Brier score (predictive accuracy) would be reduced if the values of each variable were randomly shuffled across respondents, with higher variable importance scores signifying stronger predictive power.

Figure 7. Variable importance for predicting UK voter intentions.

Note: This figure presents the variable importance scores from random forest models predicting respondents’ vote intentions, which were measured in the third-round survey. The scores are calculated using the Brier score, which measures the mean squared error between the predicted probabilities and the actual outcomes. Our random forest models achieved a Brier score of 0.47, indicating reasonable predictive accuracy. The variable importance scores show the amount the model’s Brier score would be reduced if the values of each variable were randomly shuffled across respondents; higher variable importance scores are better. The top fifteen variables are presented and are ranked in descending order of importance. The left figure includes variables from all three survey rounds (N = 663); the right figure includes variables only from the third round (N = 1,664).

As shown in Figure 7, imperial nostalgia emerges as one of the most important predictors of respondents’ voting intentions. It has similar predictive power to left–right economic values and stronger predictive power than immigration opinion, authoritarian values, and other known correlates of British vote choice. Only preferences regarding EU relations are a stronger predictor. This finding is consistent with the earlier correlation and regression results in reinforcing the central role that imperial nostalgia appears to play in British political behaviour.

In this section, we have used a variety of models and specifications to show that imperial nostalgia is an important correlate of party support and a powerful predictor of vote intentions. It is of comparable importance to economic values and immigration opinion, being stronger in certain specifications and weaker in others. However, unlike these well-established political issues, imperial nostalgia remains a neglected topic in the analysis of British (and European) politics. Our findings suggest that this neglect is misplaced, with imperial nostalgia potentially an important orientation for structuring citizens’ political views.

Yet these results raise a further question: would voters actually respond to political rhetoric about the imperial past? As we have noted, empire is not a major feature of contemporary British campaigns and political discourse. However, just as the study of general nostalgia encompasses both mass opinion (for example, Gest et al. Reference Gest, Reny and Mayer2018) and campaign effects (for example, Menke and Wulf Reference Menke and Wulf2021), a complete account of imperial nostalgia’s political consequences requires examining how citizens respond to nostalgic versus critical elite positions on empire. We investigate this question experimentally in the next section, testing whether politicians who take stances on the imperial past – either positively or negatively – face electoral rewards or penalties from voters.

The Consequences of Elite-Level Nostalgia

In this section we analyse a paired conjoint experiment designed to test how hypothetical political candidates’ positions on the British Empire causally affect respondents’ preferences. In our experiment, the parliamentary candidates are presented as having taken one of three stances regarding the empire: a nostalgic, ‘civilising effect’ position; a critical, ‘atrocities’ position; and an intermediate position that endorses both points of view.Footnote ¹³ Respondents were presented with three pairs of candidates, and were asked to choose between and rankFootnote ¹⁴ each pair of would-be MPs. Seven other MP attributes (gender, ethnicity, education, age, occupation, political party, and tax-and-spend positions) are also included in the conjoint design. These are chosen primarily because respondents may infer other important characteristics from a candidate’s imperial stance; for example, they may infer that a pro-empire candidate is more likely to be a Conservative or to hold conservative economic views. Our inclusion and randomisation of these attributes reduces the likelihood of these unobserved confounding inferences.

Figure 8 presents the marginal means, with results from the forced-choice question displayed on the left and the profile-rating questions on the right.Footnote ¹⁵ The first result to note is that respondents’ preferences are strongly and significantly influenced by elites’ positions on the British Empire. The intermediate stance – acknowledging both the atrocities committed by the empire and its so-called civilising effects – is the most favoured position.Footnote ¹⁶ The ‘civilising effect’ stance ranks second, drawing less support than the intermediate position but more than the ‘atrocities’ position,Footnote ¹⁷ which is the least popular.

Figure 8. Conjoint experiment results, marginal means.

Note: Marginal means shown, with forced-choice results on the left and profile rankings on the right; the latter includes some ‘don’t know’ responses that have been removed prior to analysis.

These results also reveal a disconnect between public attitudes to empire and the electoral viability of candidate positions on empire. On the critical side, while a majority of the public (63 per cent in the third wave) agreed that the British Empire was responsible for many atrocities, with only a tiny minority (11 per cent) disagreeing, candidates adopting this majority position face an electoral penalty rather than a reward. Similarly, regarding empire’s positive aspects, while 47 per cent of respondents agreed that the empire had civilising effects (compared to 19 per cent who disagreed), the corresponding experimental treatment generates neutral rather than positive electoral effects for candidates. This pattern suggests a systematic divergence between mass opinion and electoral traction: whilst both critical and nostalgic views enjoy substantial public support in survey responses, translating these positions into electoral appeals proves problematic for political candidates. The asymmetry is particularly pronounced for empire-critical positions, which face stronger electoral penalties despite representing majority opinion, highlighting how criticism of imperial violence remains politically taboo even when widely held among the public.

These results are drawn from the whole sample, which may obscure important heterogeneity in how ideological and partisan subgroups respond to nostalgic (or critical) views of the British Empire. To examine this potential heterogeneity, we split our sample using three measures of cultural conservatism as well as preference for a party of the left or right and compare results of candidates’ imperial positions within and between these subgroups. These results are presented in Figure 9, where the marginal means for the left- and right-leaning subgroups are shown in the first two columns, with the differences between these groups shown in the third column.Footnote ¹⁸

Figure 9. Subgroup analysis, splitting sample by measures of conservatism.

Note: Subgroup marginal means when splitting the sample by the median values of (from top) imperial nostalgia, authoritarianism, and national chauvinism, as well as by respondents’ preference for a party of the left (Labour, Liberal Democrat, SNP, Plaid, or Green) or party of the right (Conservative or Reform). Forced-choice results presented.

As we might expect, the pro- versus anti-empire stances resonate differently with culturally conservative and culturally liberal respondents. We find that right-leaning respondents strongly dislike candidates who take empire-critical positions, but only weakly prefer candidates who take pro-empire positions. And similarly, while left-leaning respondents dislike candidates who defend empire, they only weakly favour candidates who criticise empire.Footnote ¹⁹ There is therefore an asymmetry in how stances towards imperial history shape voters’ preferences. On the one hand, anti-empire views are more polarising than pro-empire views: the ‘atrocities’ treatment creates more of a divide between left- and right-leaning voters than the ‘civilising effect’ treatment (as can be seen in the differences in the marginal means). On the other hand, conservatives are more exercised by candidates’ imperial positions than are cultural liberals: the contrast between the atrocities and civilising effects treatments is larger for conservatives than it is for liberals.

In sum, our experiment reveals two main findings. First, electoral candidates’ positions on the British Empire significantly shape voter preferences. This is evident in the whole sample but is particularly clear when we split the sample by measures of conservatism (versus liberalism). Second, the effects are asymmetric, in that anti-empire positions are more polarising than pro-empire positions. This asymmetry is due to cultural conservatives’ dislike of critical views of empire more than cultural liberals’ dislike of positive views of empire.

These results resonate with our earlier discussion of the politics of imperialism in Britain: the Conservative Party declared itself the ‘party of Empire’ (and recently took pro-empire positions) whilst the Labour Party has always been more ambivalent in its views on the British Empire. This historical asymmetry in partisan alignment with the issue of the imperial past continues to the present day, with those more on the right reacting more strongly to negativity on the imperial past than left-leaning respondents do.