Do European Capitals of Culture Influence Attitudes Towards and Identification with the European Union?

Knoll, David Article — Published Version Do European Capitals of Culture Influence Attitudes Towards and Identification with the European Union? JCMS: Journal of Common Market Studies Provided in Cooperation with: John Wiley & Sons Suggested Citation: Knoll, David (2024) : Do European Capitals of Culture Influence Attitudes Towards and Identification with the European Union?, JCMS: Journal of Common Market Studies, ISSN 1468-5965, Wiley, Hoboken, NJ, Vol. 63, Iss. 4, pp. 1376-1389, https://doi.org/10.1111/jcms.13695 This Version is available at: https://hdl.handle.net/10419/323870 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. http://creativecommons.org/licenses/by-nc/4.0/ Do European Capitals of Culture Influence Attitudes Towards and Identification with the European Union? DAVID KNOLL Department of Politics and Public Administration, University of Konstanz, Konstanz Abstract The European Capitals of Culture (ECoC) programme, a cornerstone of EU Cultural Policy, is designed to foster a sense of Europeanness amongst citizens and is described as a legitimizing tool for the EU. Although ECoCs cumulatively form the largest cultural event in Europe, their effectiveness in achieving these goals remains understudied. I contribute the first assessment of the effect of ECoCs on attitudes towards and identification with the EU by comparing ECoC regions before and after the event to non-winning competitors. Results from differencein-differences regression on Eurobarometer data show that positive effects of ECoCs on attitudes towards and identification with the EU are limited both spatially to the hosting region and temporally to the immediate aftermath of the event. Thus, ECoCs are a tool to bring the EU closer to its citizens, but their effects could be amplified further by strengthening the European dimension of the events. Keywords: attitudes towards the EU; EU cultural policy; European Capitals of Culture; identification with the EU Introduction Designed to foster a sense of Europeanness amongst citizens, the European Capitals of Culture (ECoC) programme is arguably the best known and most extensive building block of EU cultural policy (Patel, 2013, p. 541). As such, it is primarily designed to ‘increase European citizens’sense of belonging to a common cultural area’(European Commission, 2024). Since 1985, this ‘artistic counterweight to the economic and commercial integration of Europe’(Mittag, 2013, p. 41f) is a title awarded to up to three cities per year in three predesignated countries. Over time, the programme evolved from a purely cultural festival to a multifaceted, yearlong mega-event entailing dozens of high-brow and mass-level cultural happenings. This includes art exhibitions, street art or festivals (Garcia and Cox, 2013, pp. 73, 78) and involves a multiplicity of stakeholders ranging from artists to citizens, cultural organizations and even embassies (McAteer et al., 2013, p. 44). To prepare for the coveted title, host cities invest considerably in upgrading their infrastructure, including the restoration of historical sites, roadworks, developing rundown areas and building cultural venues (Steiner et al., 2015, p. 376). Thus far, scholarly attention has been paid to investigating the urban and economic impacts of ECoCs (see, e.g., Balsas, 2004; Gomes and Librero-Cano, 2018; Hudec et al., 2019), but we know much less about the programme’s effect on European identity and attitudes towards the EU. Previous conceptual (Lähdesmäki et al., 2021) and qualitative (Fage-Butler, 2020) work on ECoCs linked them to the formation of a European identity, and on the 25th anniversary of the ECoC programme in 2010, it was also celebrated JCMS 2025 Volume 63. Number 4. pp. 1376–1389 DOI: 10.1111/jcms.13695 © 2024 The Author(s). JCMS: Journal of Common Market Studies published by University Association for Contemporary European Studies and John Wiley & Sons Ltd. This is an open access article under the terms of the Creative Commons Attribution-NonCommercial License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited and is not used for commercial purposes. for improving the European Union’s image (Mittag, 2013, p. 39). Still, a systematic evaluation of the connection between Capitals of Culture and citizens’European identity and EU-related attitudes is missing. To address this gap, this research note asks: what is the effect of European Capitals of Culture on citizens’attitudes towards and identification with the EU? Based on existing research, I first argue that ECoC improve (a) attitudes towards and (b) identification with the EU. This is the result of both increased intergroup contact for citizens that comes with hosting an ECoC, as well as the economic benefits that the programme brings about. I further argue that in light of the programme’sofficial goal of highlighting the cultural diversity of Europe and fostering a common sense of belonging, ECoCs increase citizens’association of the EU with cultural diversity. Relying on the competitive national selection process for ECoCs, I use a difference-in-differences (DiD) design in which citizens in ECoC regions are compared to citizens in regions with cities that competed for the title but ultimately did not win it. This strategy uses citizens in competitor regions as comparison group for the citizens in ECoC regions and is applied to 19 Capital of Culture rounds between 2005 and 2019. The results support the second part of the argument, showing that ECoCs foster associations of the EU with cultural diversity, thereby achieving the programme’s core aim of highlighting the rich cultural heritage and diversity of Europe. Findings on the first part of the argument are mixed though: the programme has no discernible impact on people’s image of the EU itself but can increase trust in the EU (attitudinal effects), as well as identification as an EU citizen. However, effects are limited geographically to the actual host city (as opposed to the wider region) and limited in time to the immediate aftermath of the event. Ultimately, results demonstrate that Capitals of Culture can fill the EU’s motto ‘United in Diversity’with life but point to important scope conditions of these effects. This research note speaks to the literature on EU cultural policy, as well as on attitudes towards and identification with the EU more broadly. The present work contributes the first large-N, comparative assessment of the ECoC programme’s attitudinal and identity-related effects, a thus far understudied aspect of one of Europe’s largest cultural events. In doing so, it adds a new angle to the evaluation and future development of the ECoC programme as the flagship of European cultural policy. I. ECoCs and Attitudes Towards the European Union What are expectations regarding the effects of ECoCs on EU-related attitudes and identities? Invented to show tangible results after an inconclusive Greek Council presidency in 1983, Greek minister of culture and actress Melina Mercouri initiated the programme as an art-based counterpoint to the developing common market (Mittag, 2013). Besides their economic benefits, the literature describes ECoCs as a legitimizing tool for the European Union (Fage-Butler, 2020; Lähdesmäki et al., 2021), a ‘cheap tool for marketing’for the EU (Palonen, 2010, p.104) and a way to sell the community to its citizens (Sassatelli, 2002, p. 436). An unnamed member of the European Parliament poignantly summarized the necessity for a European cultural policy as a legitimizing tool for the European Union by saying ‘no one falls in love with a common market’, (as cited in Shore, 2004, p. 28). The Commission of the European Communities recognized this as Do European Capitals of Culture Influence Attitudes Towards and Identification with the European Union? 1377 © 2024 The Author(s). JCMS: Journal of Common Market Studies published by University Association for Contemporary European Studies and John Wiley & Sons Ltd. early as 1987, when they noted somewhat verbosely that ‘[…] being part of a European culture […] is one of the prerequisites for the solidarity which is vital if the advent of the large market and the resulting radical changes in living conditions within the Community is to secure the popular support it needs’(Commission of the European Communities, 1988, p. 5). Even though the programme evolved from a purely cultural celebration of Europe towards an economic and touristic booster-event for run-down cities, bringing Europeans closer together remains an official goal and each application needs to outline how they plan to fill the required ‘European dimension’with life. The idea of ECoCs as a legitimizing tool for the European Union ties in with well-established identity-based as well as utilitarian theories of EU support. First, from an identity-based perspective, citizens come to like the EU better after ECoCs because they feel more European as a result of being exposed to European symbols (cf. Billig, 1995; Manners, 2011; Shore, 2013) and positive intergroup contact (Allport et al., 1954) with fellow Europeans. European flags, the EU constantly looming in the background as the initiator of the programme, news coverage of events, speeches, debates, co-operations with other European cities and artists, city partnerships or the highlighting of an ECoC’s European heritage are frequently used to fill the ‘European dimension’, which every ECoC is required to include in its programme, with life (Garcia and Cox, 2013). To the extent that the ECoC programme ‘contributed to the Europeanization of cities around Europe through requiring them to find ways to narrate and present themselves as European’(Lähdesmäki et al., 2021, p. 56), it makes Europe salient in the host cities. Literature on experiencing the politics of everyday Europe (e.g., Kuhn, 2015; Recchi et al., 2019) helps to explain the mechanism of identity construction through the power of symbols, pointing to the role of repetition of cultural practices, normalization and deliberate labelling in shaping the transnational social reality of Europe. McNamara (2015, p. 82) concludes that through ECoCs, ‘[…] Europe is diffused into the everyday fabric of the cities, in a subtle and banal set of representations matched with various practices celebrating these European capitals’. The mechanism ultimately works similar to the one linking the Erasmus student exchange programme to EU support (see, e.g., Mazzoni et al., 2018). Second, from a utility-based perspective, citizens come to like the EU better after experiencing an ECoC because they benefit from it economically. Previous research found positive effects of ECoCs on local and regional economic development, despite their high costs (Falk and Hagsten, 2017; Gomes and Librero-Cano, 2018). If individuals ascribe this economic improvement to the EU as the initiator of the programme, utilitarian theories of EU support (Foster and Frieden, 2021; Gabel, 1998) predict an improvement in attitudes towards and identification with the EU, as people realize the benefits brought about by the event. 1 As the events are tied to a specific host city or region, the first hypothesis reads: H1 : Citizens in Capital of Culture regions exhibit (a) improved attitudes towards and (b) higher identification with the EU compared to citizens in non-hosting regions. 1 As those benefits can materialize not only during hosting but also during the preparation phase (Gomes and LibreroCano, 2018), effects of announcing that a city will host an ECoC are tested later on as well. David Knoll1378 © 2024 The Author(s). JCMS: Journal of Common Market Studies published by University Association for Contemporary European Studies and John Wiley & Sons Ltd. From a policy evaluation perspective, ECoCs should also increase citizens’association of the EU with cultural diversity. Former Commission president Barroso echoed the stated goal of ECoCs to highlight the diversity within the European cultural space, when he praised the programme as being an ‘[…] illustration of the EU’s commitment to cultural diversity and also how culture can unite people within Europe’(European Commission, 2010). If ECoCs achieve the goal of highlighting the diversity that is part of the EU’s motto ‘In Varietate Concordia’, individuals in Capitals of Culture should associate the EU more strongly with cultural diversity than citizens on other regions, as the former are directly exposed to the diversity these events entail. Note that this association is not directional, but associational: it does not qualify whether a person approves of cultural diversity as a value or not; it merely refers to the mental link. H2 : Citizens in Capital of Culture regions perceive the EU to be more culturally diverse than citizens in non-hosting regions. II. Research Design This research note uses DiD regression on Eurobarometer survey data from 2005 to 2019 to investigate the attitudinal effects of hosting Capitals of Culture. Since 2005, the EU predesignates up to three countries per year to host an ECoC (European Commission, 2005), configuring the selection as several independent nation-wide competitions between cities for the title each year (Mittag, 2013, p. 45). Hence, I compare attitudes in ECoCs regions to their competitor regions before and after the event. This strategy has also been used to investigate the economic effects of ECoCs (Falk and Hagsten, 2017; Gomes and Librero-Cano, 2018), as well as life satisfaction in ECoCs up until the early 2000’s (Steiner et al., 2015). However, this study is the first to use the competitive ECoC selection process to investigate EU-related outcomes using individual-level survey data, which are obtained by cumulating and unifying 26 Eurobarometer waves from 2005 to 2019 (listed in Table S1). This represents the time-frame between the onset of nationalized competitions for the title and the beginning of the Covid pandemic that disrupted the regular cycle of events and includes about 21,500 respondents across 19 rounds (Table 1). Some features of this approach merit discussion. First, high politicization and the influence of experts in the selection process (Patel, 2013) cast doubts over the similarity of winning and losing cities. Still, Table 1shows that ECoC cities represent a diverse mix of capitals and peripheral cities, including poorer and wealthier places alike. Second, endogeneity could arise if ECoCs have higher levels of EU support than competitors to begin with. The DiD approach circumvents this problem by looking at changes between cities, thereby accounting for baseline differences. Third, other events such as the Euro-crisis also affect EU support. However, as long as events do not affect ECoCs and competitors differently, these shocks are absorbed in the within-country comparison resulting from the national competitions. Fourth, some competitor cities host local, scaled-back versions of their envisioned ECoC programme despite losing the competition (Richards, 2015). What sounds like a confounding event in fact strengthens the approach as it makes hosts and competitors more alike, with differences being reduced Do European Capitals of Culture Influence Attitudes Towards and Identification with the European Union? 1379 © 2024 The Author(s). JCMS: Journal of Common Market Studies published by University Association for Contemporary European Studies and John Wiley & Sons Ltd. to the EU looming in the background as the name-sake patron of the actual ECoC. Fifth, one would ideally compare ECoCs only to the closest competitors, not all contenders. Such a hard ranking is usually not created, but where possible, only close competitors (as identified by Gomes and Librero-Cano, 2018, tab. 3) are used as comparison group. Finally, even if cities are not chosen as-if randomly from a pool of close competitors, the award does come at random for the average citizen (Steiner et al., 2015). To measure attitudes towards the EU, the outcome for H1a, I use two variables: first, a dummy indicating if a person ‘tends to trust the European Union’(=1) or not (=0) and, second, a 5-point scale of respondents’image of the EU (‘In general, does the EU conjure up for you a very positive, fairly positive, neutral, fairly negative or very negative image?’), with higher values indicating a better image. To measure respondent’s European identity (H1b), a dichotomized version of the 4-point item ‘You feel you are a citizen of the EU’is used. To investigate H2, I rely on a dummy constructed from the question ‘What does the EU mean to you personally?’. It equals one if ‘cultural diversity’was Table 1: Overview of European Capitals of Culture. City Country Year Notes Cork Ireland 2005 Patras Greece 2006 Dropped (no competition) Luxemburg Luxemburg 2007 Dropped (no subnational data) Sibiu Rumania 2007 Dropped (no competition) Liverpool United Kingdom 2008 Stavanger Norway 2008 Dropped (not EU) Linz Austria 2009 Dropped (no competition) Vilnius Lithuania 2009 Dropped (no competition) Essen Germany 2010 Pécs Hungary 2010 Istanbul Turkey 2010 Dropped (not EU) Turku Finland 2011 Tallinn Estonia 2011 Guimarães Portugal 2012 Dropped (all competitors in same region) Maribor Slovenia 2012 Marseille France 2013 Košice Slovakia 2013 Umeå Sweden 2014 Riga Latvia 2014 Mons Belgium 2015 Dropped (no subnational data) Pilsen Czech Republic 2015 Wroclaw (Breslau) Poland 2016 San Sebastián Spain 2016 Aarhus Denmark 2017 Paphos Cyprus 2017 Leeuwarden Netherlands 2018 Valletta Malta 2018 Dropped (no subnational data) Matera Italy 2019 Plovdiv Bulgaria 2019 Source: Data: Gomes and Librero-Cano (2018, tab. 3); GESIS for disaggregation information (https://www.gesis.org/en/ eurobarometer-data-service/survey-series/standard-special-eb/study-overview). David Knoll1380 © 2024 The Author(s). JCMS: Journal of Common Market Studies published by University Association for Contemporary European Studies and John Wiley & Sons Ltd. chosen from a non-exclusive list of elements, and zero otherwise. Consequently, I rely on logistic regression for trust, identification and association with cultural diversity and ordered logistic regression for image. The DiD estimator as the main independent variable is the interaction of two dummies: first, time measures if respondents were interviewed before (0) or after (1) the event took place in their country. As Eurobarometer waves are usually fielded in spring and late autumn, the autumn wave before an ECoC-year (e.g., wave 62.0 from late 2004 for the 2005 ECoC of Cork) is the pre-event measure, and the autumn wave near the end of the ECoC-year represents the post-event measure (wave 64.2 from late 2005 for Cork). By then, most events of the ECoC have passed. 2 Comparing two autumn waves accounts for potential seasonality effects. In the robustness tests, I also check for effects of announcing (as opposed to hosting) the event by using the announcement date as treatment. Second, ecoc is a dummy indicating if a respondent lives in the region hosting the ECoC (1) or in a region with a competing city (0). The variable is constructed from Eurobarometer region indicators, with information on competitors taken from Gomes and Librero-Cano (2018, tab. 3) and the official evaluations for more recent cities. I additionally control for a number of sociodemographic factors, namely, respondent age, gender, age when finishing full-time education, occupation status (working/student/retired/ non-working) and residency (rural/town/urban). Due to many missing values on attitudinal covariates of EU support, they are not included as controls in the regression, but their pre-event balance is discussed below and respondent’s political left–right orientation is included as a covariate in Table S7 with the correspondingly reduced sample size. Detailed information on variables, coding, cities and competitors is provided in Tables S2 and S3 and Figure F3. Before looking at the results, I discuss three underlying assumptions. First, it is necessary to verify that respondents in host and competitor regions are sufficiently similar to serve as comparison groups for each other. It is thus reassuring to see that most pre-event covariates are balanced between groups, even though splitting around 1000 respondents per country and wave in the respective regions does not guarantee representative samples on the regional level (Table S4). Given the extensive preparation before the event and the boosterist talk about benefits accompanying them, it is not surprising that the economic expectations and support for globalization are slightly more positive in host regions. Second, the DiD framework requires that both groups follow parallel trends prior to the event. Figure F1 shows that by and large, the parallel trends assumption holds. Exceptions include trends in Cyprus and Poland, which are not included in subset analysis later on for that reason. Remaining differences appear to be attributable to statistical noise. Finally, I argue that spillovers from ECoCs to competitors are not a major concern –if anything, they attenuate effects towards zero. Whilst a decrease in attitudes in competitor cities as a ‘backlash’against losing the competition is theoretically possible, it does not appear empirically. Three outcome variables remain constant or increase. Only image decreases, but does so in the host region as well, indicating an overall time trend, rather than a backlash specific to competitors. 2 Investigating longer periods of time before and after the event is not a viable option, as the regional coding becomes inconsistent across longer time spans for many countries, which is also why an event-study is infeasible in the present case. Do European Capitals of Culture Influence Attitudes Towards and Identification with the European Union? 1381 © 2024 The Author(s). JCMS: Journal of Common Market Studies published by University Association for Contemporary European Studies and John Wiley & Sons Ltd. III. Findings Afirst descriptive result of the effects of hosting Capitals of Culture on attitudes towards and identification with the EU are t-tests of post-event differences. Whilst they show significantly higher trust in the EU and higher associations of the EU with cultural diversity in host regions (p<0.001), the DiD regressions in Table 2do not appear to support the hypotheses at first glance. The models include demographic controls, a series of dummies for event rounds and, of course, the interaction between time (pre/post hosting the event) and ECoC (host region/competitor region). Looking at all 19 rounds for which data are available, the DiD estimators return insignificant for respondents’image of the EU (model 1), their trust in it (model 3), their self-identification as an EU citizen (model 5) and their association of the EU with cultural diversity (model 7). However, the right-hand models for each outcome in Table 2paint a more nuanced picture. They focus on the subgroup of ECoC rounds in which effects (positive as well as negative) are most likely to be detectable –that is, rounds where city-level effects are less prone to get diluted in the wider region being coded as ‘hosting’, until they become statistically undetectable noise. This includes (a) events where the ECoC and the region coded as hosting overlap more closely due to the fact that NUTS-2 regions are either geographically comparatively small or rounds with more detailed NUTS-3-level data available [Pilsen 2015 (CZ), Aarhus 2017 (DK), Leeuwarden 2018 (NL), Turku 2011 (FI), Tallin 2011 (EE) and Riga 2014 (LV)] and (b) events that explicitly incorporated the entire region [Maribor 2012 (SI), Essen 2010 (DE), see Srakar and Vecco, 2017; Mittag, 2013]. When compared to the time trend of the competitor regions, citizens in this subset of ECoC regions appear to have higher trust in the EU (Table 2, model 4, p<0.1), tend to identify as EU citizens more frequently (model 6, p<0.1) and also associate the EU more strongly with cultural diversity (model 8, p<0.05). These findings –visualized in Figure 1–illustrate that under favourable circumstances, the change in average predicted probabilities to trust the EU increases by 4.1 percentage points compared to the predicted change in trust in the competing regions. The predicted probability to identify as a European citizen increases by 5.6 percentage points and the probability to associate the EU with cultural diversity by about 3.9 percentage points. These effects amount to 8.2%, 12.2% and 9.5% of the respective variable’s standard deviation. The EU’s image in hosting regions remains unaffected, as model 2 shows. To sum up, there is conditional empirical support for H2 (EU = Cultural diversity) and H1b (EU identity), whereas support for H1a (EU attitudes) is mixed, with null findings for EU image and limited support for trust. The results so far use the autumn wave of the hosting year as post-event measures. To test if effects persist, I rerun the above analyses using spring wave after hosting as the post-event measure. Table S5 shows that effects are short lived and disappear less than half a year after a city’s spell as an ECoC draws to a close, and Table S13 reports substantially similar results obtained from linear models. These points to important scope conditions for the attitudinal and identity-based effects of ECoCs: first, they appear to be limited to the hosting time itself, and second, they do not spill over to the wider geographic region. Instead, they are only detectable if an entire region is exposed to the event, either David Knoll1382 © 2024 The Author(s). JCMS: Journal of Common Market Studies published by University Association for Contemporary European Studies and John Wiley & Sons Ltd. Table 2: DiD Regression Results of Hosting ECoCs on EU Attitudes. EU image Trust in EU EU citizen EU = cultural diversity (1) (2) (3) (4) (5) (6) (7) (8) ECoC 0.17*** 0.22** 0.16** 0.10 0.08 0.15 0.01 0.07 (0.04) (0.07) (0.05) (0.08) (0.07) (0.11) (0.06) (0.09) Post-event 0.01 0.09 + 0.03 0.23*** 0.06 0.19* 0.05 0.01 (0.03) (0.05) (0.03) (0.06) (0.04) (0.09) (0.04) (0.07) ECoC * Post-event 0.04 0.01 0.04 0.18 + 0.04 0.27 + 0.10 0.23* (0.06) (0.09) (0.07) (0.10) (0.09) (0.15) (0.08) (0.11) Gender 0.03 0.07 0.01 0.02 0.04 0.04 0.20*** 0.26*** (0.03) (0.04) (0.03) (0.05) (0.04) (0.07) (0.03) (0.06) Age 0.00*** 0.00 0.00*** 0.01* 0.00* 0.00 0.00 0.01* (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) Education 0.10*** 0.10*** 0.09*** 0.09*** 0.15*** 0.16*** 0.14*** 0.13*** (0.01) (0.01) (0.01) (0.01) (0.01) (0.02) (0.01) (0.01) Residency (base: town) Rural 0.04 0.10 + 0.07* 0.08 0.13** 0.06 0.14** 0.20** (0.03) (0.06) (0.04) (0.06) (0.05) (0.09) (0.04) (0.07) Urban 0.08* 0.08 0.02 0.08 0.13* 0.04 0.18*** 0.24*** (0.03) (0.06) (0.04) (0.06) (0.05) (0.09) (0.04) (0.07) Occupation (base: working) Student 0.05 0.07 0.35*** 0.32** 0.14 0.34* 0.20** 0.12 (0.06) (0.10) (0.07) (0.11) (0.10) (0.16) (0.07) (0.12) Retired 0.05 0.04 0.13** 0.25** 0.20** 0.08 0.12* 0.08 (0.04) (0.07) (0.05) (0.08) (0.06) (0.11) (0.06) (0.09) Other non-working 0.28*** 0.30*** 0.18*** 0.25** 0.42*** 0.38** 0.16** 0.27** (0.04) (0.08) (0.05) (0.09) (0.06) (0.12) (0.06) (0.10) FE for rounds Yes Yes Yes Yes Yes Yes Yes Yes Constant 0.50*** 0.97*** 0.33** 0.53** 1.88*** 1.83*** (0.11) (0.16) (0.13) (0.20) (0.12) (0.18) Pseudo R 2 0.050 0.018 0.066 0.045 0.060 0.053 0.040 0.040 Log-likelihood 27,250 9002 12,540 4493 7955 2533 10,663 3927 Observations 21,520 7493 19,381 6791 13,459 4198 20,899 7552 Notes: Robust standard errors in parentheses. Ordered logistic regression for EU image, logit for trust, EU citizen and cultural diversity. + p<0.1. * p<0.05. ** p<0.01. *** p<0.001. Do European Capitals of Culture Influence Attitudes Towards and Identification with the European Union? 1383 © 2024 The Author(s). JCMS: Journal of Common Market Studies published by University Association for Contemporary European Studies and John Wiley & Sons Ltd.