Translating the complexity of EU differentiation: a comparative analysis of news coverage in Germany, Denmark, and the UK

ABSTRACT

This article explores whether and how journalists cover various forms of differentiation in the European Union (EU) in the news. Differentiation is a key feature of European integration, viewed by some as a facilitator of democracy and by others as a mechanism leading to dominance. Yet we know only little how its complex and controversial aspects are communicated to the citizens by journalists in their role as public mediator and critical watchdogs. Drawing on a selection of news reports published between 2015 and 2020 in nine media outlets from Germany, Denmark, and the UK, we find that the media frequently covers EU differentiation in a range of forms that extends beyond the conventional academic focus on territory-based differentiated integration. The majority of this coverage presents a balanced view, yet nuances emerge with respect to the forms of differentiation discussed and the origins of commentary on these matters. Overall, these results point to the key role of the media as translators and watchdogs of the complexity of EU governance.

KEYWORDS:

• European Union
• differentiation
• democracy
• dominance
• news media

Introduction

The European Union (EU) is more differentiated than its name might suggest (Fossum 2019; Holzinger and Schimmelfennig 2012; Schimmelfennig, Leuffen, and Rittberger 2015). For example, the EU has its own currency, although not all its members have adopted the euro. Some EU member states are not willing to integrate into the eurozone, such as Denmark, where the opportunity to do so has been rejected by their citizens in referendums. The governments in some other member states, including Bulgaria, are committed to adopting the currency, but these countries are made to wait until they meet the EU’s criteria before they can integrate into the common currency. At the same time, the euro is in use in a number of non-member states in Europe, with or without the EU’s agreement.¹

This complex form of integration is nothing specific to the euro or to other territory-based differences between nation states in Europe. Indeed, while territorial differentiation is the most well-known form of differentiation in the EU, it is not the only one, and the Union is characterised by a broader spectrum of differentiation (Fossum 2019). This includes law-making differentiation, where legislative processes and the roles of various actors are tailored to accommodate diverse needs, altering the existing legislative workflows. Another form is competence-based differentiation, which modifies the functions or roles in the political system. Lastly, rights-based differentiation adjusts the rights or obligations of the citizens. Together, these forms of differentiation make up the complex approach to integration in the EU.

EU differentiation is not only complex, it is also controversial and thus prone to politicisation (Bellamy, Kröger, and Lorimer 2022; Fossum 2019; Lord 2021). It is seen, on the one hand, as conducive to democracy as it allows for divergence and accommodation of differences in preferences. Indeed, without differentiation in common currency, for example, unwilling Member States would have to join the eurozone – or else the common currency would not exist at all, despite the wishes of the majority of the EU members. On the other hand, other scholars consider EU differentiation as leading to dominance because it creates insiders and outsiders of integration, excluding the latter from the decision-making process that affects them (Eriksen 2019). Returning to the eurozone example, differentiation means that willing Member States can be prevented from participating in the common currency until they meet certain criteria imposed by the others. How the media assess differentiation in the EU, therefore, can be critical for the public opinion to be formed in the future.

In this article, we explore whether and how journalists report on various forms of differentiation in newspaper coverage of EU affairs. Existing research shows that media reporting on EU affairs contributes to what people know and also think about the EU (Gattermann and de Vreese 2020; Marquart et al. 2019; Vliegenthart et al. 2008), yet media coverage of EU affairs is often irregular and incomplete (Richter and Stier 2022). Although the EU itself has increasingly been involved in public communication (Altides 2009), many remain sceptical about the effectiveness of its communication efforts (Hillebrandt 2021; Özdemir and Rauh 2022; Rauh 2021). Rauh (2021), for example, finds that the European Commission’s press releases are significantly more difficult to comprehend than articles in political sections of newspapers. Therefore, at a time when public knowledge of differentiation remains low (de Blok and de Vries 2023; Stahl 2021), newspapers coverage of EU affairs could play a critical role in bridging the communicative gap between European publics and EU institutions by facilitating public opinion formation based on plurality of opinions and knowledgeable discussion of the implications that EU governance and differentiation have for EU citizens’ lives (Michailidou and Trenz 2023).

To explore the role of media in covering differentiation in the EU, we study a selection of news reports published in nine outlets from three countries with the highest degree of differentiated secondary legislation in the EU: Germany, Denmark, and the UK. Our sampling period amounts to three months in total, constructed from six events spanning the years from 2015 to 2020.² These events are all broadly related to Justice and Home Affairs, a highly differentiated policy area that is likely to generate debates in the public sphere. We therefore study a most likely case to find coverage and framing of differentiation in the media, based on qualitative coding of sentences in news articles and a quantitative analysis of the consequent dataset. We find a substantive and overall balanced coverage, underscoring the media’s pivotal role as a facilitator of deliberative discourse on differentiation in the EU. Overall, our study speaks to a growing body of literature on the relationship between differentiation and public opinion in the EU (Leuffen, Schuessler, and Gómez Díaz 2022; Schimmelfennig, Leuffen, and de Vries 2023; Schraff and Schimmelfennig 2020).

EU in the news

The EU can enter the news in various ways. Often, the EU and its main actors appear backstage in domestic politics or in economic news, for example when EU law and regulation applies (Trenz 2005). As such, the EU is part of every-day reality of its member states and does not need to be specifically addressed or further explained. At the frontstage, EU news is generally created as part of foreign news coverage by a specialised group of journalists, some of them temporarily sent to Brussels as correspondents to cover the EU (Hannerz 2004; Meyer 2010; Siapera 2004; Terzis 2008). In this context, there arises a need not only for conveying information but also for providing explanations regarding the interests at stake, the rationale behind the necessity for policies, and the increasing desirability of coordination.

EU news coverage is generally characterised by a mismatch between supply and demand (Michailidou and Trenz 2023). The supply chain of EU news is not only held in motion by the work of EU correspondents but also by specialised press agencies and by the press offices of the EU who deliver ready-made textual and visual material for the free use of journalism (Aldrin 2013). Demand for EU news is channelled through the daily news selection by the copy editors in the home offices of the national news organisation, who only reserve limited space for print or TV news from Brussels (Meyer 2010). This is in line with the perceived preferences for news of national audiences who devote only little time and attention to news from Brussels (Martins, Lecheler, and de Vreese 2012). This draws attention to the selection criteria that are applied by national news organisations to sell EU news to their audiences. Research on EU news has shown that several filters apply in the selection and framing of EU news: a nationalist filter, where EU news are highlighted whenever national interests are at stake at the cost of a European common good orientation; a conflict filter, where EU news are highlighted when simple conflict lines between states can be drawn at the cost of complex coalitions and often shifting majorities; and a negativity filter, where negative EU news about scandals or mismanagement are highlighted over positive or neutrally frame news (Galpin and Trenz 2019).

These selection and framing filters in the working routines of EU journalism have consequences for the coverage of complex issues such as differentiation in the EU. Journalists, in general, are translators of the complexity of the world of politics that is explained in a common language understood by lay people. Journalists, in other words, de-differentiate. They round up and summarize instead of covering details. Differentiation, in turn, means a loss of newsworthiness. For EU news, a trade-off applies between the complexity of a news story that is needed to inform about facts and to engage in critical opinion-making and the limited attention of audiences. The higher the information value of an EU news story, the lower its market value to be sold and reach out to broader audiences. This reduces the likelihood to cover EU differentiation in the news. Differentiation can however also be seen as an opportunity for a specialised corps of expert journalists to fill a market niche for critical news about EU politics (Michailidou and Trenz 2023). This responds to specific demands for information and critical engagement by smaller segments of the audience. In response to EU differentiation, also media markets and audiences would differentiate. Especially EU correspondents would take a new role as driver of critical discourse about the EU and a watchdog of differentiation in the EU.

We therefore arrive at two possible role models of professional journalism confronting an increasingly differentiated EU. First, journalists as translators of the complexity of EU governance would mainly de-differentiate and select and frame EU news stories according to criteria of relevance for their national readers. Second, journalists as watchdogs of the complexity of EU governance would need to differentiate themselves and respond to specialised demands of critical niche publics. By looking in the following at how EU news coverage addresses EU differentiation, we do not consider both role models of EU journalism as exclusive but rather wish to demonstrate how both function of a selective filter and translator and as a critical watchdog are intertwined in the daily work of news making with different emphasis across countries and news outlets.

Research design and data

Our qualitative content analysis draws on a purposeful sample of news reports on EU affairs that is constructed by underlying three selection criteria. First, we limited the time frame for analysis to cover 90 days around six important events in the EU across six years from 2015 to 2020—one event for each year, with seven days immediately before and after each event to the time frame. Table 1 provides the list of the chosen events and dates.

Table 1. List of events and dates behind the time frame for analysis.

Event	Date	Period
Speech by Angela Merkel on Germany’s open border policy	31 August 2015	From 24 August 2015 to 7 September 2015
Proposal by the Commission, to reform the Common European Asylum System	13 July 2016	From 6 July 2016 to 20 July 2016
Joint statement by Visegrád Group, titled “Strong Europe — Union of Action and Trust”	2 March 2017	From 23 February 2017 to 9 March 2017
Informal Council meeting on internal security, migration, and Brexit	19 September 2018	From 12 September 2018 to 26 September 2018
Migration policy in the strategic agenda 2019–2024	20 June 2019	From 13 June 2019 to 27 June 2019
Migration Pact statement by Ursula von der Leyen	16 September 2020	From 9 September 2020 to 23 September 2020

The events in Table 1 are all broadly related to the area of Justice and Home Affairs, and specifically to the topic of migration through internal and external borders of the EU. Justice and Home Affairs is a highly differentiated policy area in the EU (Comte and Lavenex 2021; Duttle et al. 2017).³ It has the highest share of differentiation in terms of secondary legislation, with one or more member states opting out of about 42% of the legislation in this area between 1958 and 2012 (Duttle et al. 2017). Focusing on periods around important events in Justice and Home Affairs increases the likelihood that our dataset includes observations where differentiation might be covered.

As for the second sampling decision, we chose to focus on nine media outlets from Germany, Denmark, and the UK. Generally speaking, these were the top three countries in terms of legislative differentiation. However, in Justice and Home Affairs, they were at the extreme ends of the scale: while Germany did not opt out from any of the related 180 legislative acts enacted in EU secondary law in this area between 1958 and 2012, Denmark and the UK opted out of, respectively, 38% and 21% of these acts (Duttle et al. 2017). Nevertheless, Germany is said to be ‘the major driver of differentiated integration’ in Justice and Home Affairs (Comte and Lavenex 2021, 3) – although it is itself not differentiated, Germany leads the development of EU policies in this area, but allows for the sceptical countries to opt out, so that the remaining countries further integrate along the lines of their lead. As a result, our case selection follows a most-like case design in terms of salience, and we expect the media outlets from this set of countries to be more likely to report on differentiation.

When it comes to the selection of specific media outlets, we had two criteria in mind – that (1) our selection should include one centre-left, one centre-right, and one economy outlet per country and that (2) each outlet should have at least one correspondent in charge of EU affairs in the time frame under analysis. Table 2 provides the list of outlets and their correspondent with the highest number of articles published in our time frame. We then went through the website of these outlets, and collected any article written by EU correspondents. We also collected articles written by other journalists, if these were clearly about EU affairs. This resulted in 1,363 news articles, written by 251 journalists.⁴

Table 2. Case selection, and the correspondent with highest number of articles per outlet.

Country	Orientation	Outlet	Top Journalist
Germany	Centre-left	Sueddeutsche Zeitung	Alexander Mühlauer (38)
Germany	Centre-right	Frankfurter Allgemeine Zeitung	Werner Mussler (75)
Germany	Economy	Wirtschaftswoche	Silke Wettach (22)
Denmark	Centre-left	Politiken	Nilas Heinskou (34)
Denmark	Centre-right	Berlingske	Jakob Ussing (37)
Denmark	Economy	Dagbladet Børsen	Tore Keller (57)
United Kingdom	Centre-left	The Guardian	Daniel Boffey (103)
United Kingdom	Centre-right	The Times	Bruno Waterfield (88)
United Kingdom	Economy	The Financial Times	Mehreen Khan (158)

The Top Journalist column lists the journalist with the highest number of news articles (in parentheses) per media outlet.

Coding

Our unit of analysis are sentences within news articles, which can contain a diagnostic or evaluative dimension of EU differentiation, dominance, or democracy – or a combination of these elements. We identify whether each sentence refers to particular forms of EU differentiation and how such differentiation is assessed by the journalist or other actors as directly or indirectly quoted in the news. Because our qualitative coding strategy is highly resource-intensive, there was uncertainty regarding our ability to code all the articles that we collected. To avoid selection bias in the event that we would not be able to apply our codebook to all articles in our dataset, we randomly ordered the news articles within each country before beginning the coding process. The coding was then carried out by three coders, one for each language in the sample.⁵ Appendix B provides our codebook, which formed the basis for the training of the coders.

As our study is primarily about the coverage of differentiation in the media, especially with regard to dominance and democracy, we initially coded whether the sentences in news articles included a statement about differentiation, dominance, or democracy in the EU. For sentences without a reference to differentiation, we did not code any other variables. Otherwise, we continued the coding with the following variables.

For our diagnostic variable, we coded whether the sentences referred to one or⁶ more of the four forms of differentiation – as defined by Fossum (2019):

• Law-making differentiation refers to changes in the way legislative decisions are made at any level of government in the EU, such as the European Parliament becoming more or less powerful in the EU legislative process.

• Competence-based differentiation refers to changes in the functions of the political system or its actors, such as the creation of a new EU agency.

• Territorial differentiation refers to changes in the European integration at national borders, such as Ireland not being a part of the Schengen Area.

• Rights-based differentiation refers to the changes in the rights that citizens have, such as EU citizens having the right to vote in local elections but not in national elections, if they are residing in an EU member state other than their country of origin.

For our evaluative variables, we coded whether the sentences included a suggestion that (1) differentiation is politically acceptable or sustainable. Moreover, we coded whether there is a suggestion that differentiation is positively or negatively related to (2) dominance or (3) democracy. For independent variables, we coded the (a) type of differentiation in the statement, (b) the makers of these statements (journalists, citizens, or actors from the public, private, or civil society sectors), and (c) the media origin (Germany, Denmark, UK). Finally, we coded (d) whether the sentences were about migration, as our time frame was constructed around events related to the topic of migration through internal and external borders of the EU.

Measuring text at the level of sentences increases precision in qualitative coding, yet this strategy has two potential disadvantages (Roller and Lavrakas 2015). First, sentence-level measures might miss relevant contextual information provided elsewhere in the text, but not repeated in every sentence. To address this challenge, we allowed our coders to drive contextual information for a given sentence from the previous sentence. Second, analyses at the lower level increase the number of decisions that coders have to make, and therefore they are resource intensive. As explained above, to address this challenge, we used three sampling steps to reduce the number of news articles in the dataset.

To ensure inter- and intra-coder reliability and the interpretative validity of our codebook, we first held several rounds of coder training, followed by two test coding sessions on the same, randomly chosen, articles both times. There were 10 articles and 329 sentences for the first test, 5 articles and 192 sentences for the second text. This training and discussion-intensive process resulted in high a Krippendorff’s Alpha scores, such as a score of 0.777 for the variable on diagnostic dimension.

In the end, we coded 18,566 sentences in 570 (42%) articles, written by 146 (58%) journalists from all nine outlets. The decision to not code the entire dataset⁷ of articles was taken after we evaluated the results of about 10–15% of all sentences coded across the three countries, whereby it became clear that there was very little variation in the variables assigned across and within country data. We subsequently agreed to randomly select a sub-sample per country. Setting the confidence level at 98% and error margin at 3%, we rounded up the number of sentences that needed to be coded at 1500 per country. After randomly ordering the sampled articles, we instructed our coders to code until there are at least 1,500 sentences coded as related to differentiation, dominance, and/or democracy. We reached this threshold at sentences numbered 3,035 in Germany, 9,474 in Denmark, and 6,009 in the UK.

Results

The diagnostic dimension

Overall, approximately 25% of all sentences across all news reports under analysis referred to differentiation in the EU – about four times the total share of sentences related to democracy or dominance.⁸ Figure 1 plots these results by country, showing that such thematic references are significantly higher in German newspapers for all three categories: 55% of randomly-selected sentences from outlets in Germany are concerned with differentiation in the EU. This compares with only 16% in Denmark and 24% in the UK.

Figure 1. Share of sentences with statements related to differentiation, dominance, or democracy in the EU.

Further descriptive analyses reveal important differences with regard to the different forms of differentiation highlighted in the news. In Germany, the main emphasis is on competence-based differentiation while territorial differentiation features most frequently in outlets from Denmark and the UK (Figure 2). In other words, in the latter two countries differentiation is discussed more often in relation to states and territories than to law and its impact on empowering or disempowering individuals. Given that the timeframe of analysis covers the UK’s withdrawal from the EU, an example of territory-based differentiation, it is not surprising that media coverage in the UK, and to a similar extent in Denmark, places a stronger emphasis on territory-based differentiation. In contrast, it is striking that law-making differentiation is covered substantively only in Germany, whereas this form of differentiation almost completely overlooked by journalists in Denmark and the UK. Equally, questions of differentiation with regard to civil and political rights, freedom of movement, and citizenship are rarely taken up by journalists in any of the countries in the dataset.

Figure 2. Share of differentiation types in news articles. Note: As a single sentence can contain statements about more than one type of differentiation, the sums of shares do not equal to 100%.

In terms of the sources of differentiation-related statements in news articles, we find a similar country-level variation that sets apart the way differentiation is covered in Germany from the coverage in Denmark and the UK. This is demonstrated in Figure 3. Specifically, we find that a large majority of differentiation-related statements published in the German media originates from journalist themselves, sometimes alongside other actors within the same sentence. This suggests that journalists in Germany speak more in the role of critical watchdogs, a crucial function that refers to the responsibility of the media itself to monitor, investigate, and report on the complex workings of the EU. Meanwhile, in Denmark and the UK, similarly large majorities of differentiation-related statements are instead attributed to public actors, such as ministers or members of parliament. This points towards a contrasting journalistic approach in these countries, where journalists primarily serve as intermediaries, translating complex EU processes – and the perspectives of public actors on these processes – for the public.

Figure 3. Share of sources of differentiation-related statements in news articles. Note: As a single sentence can contain statements from more than one source, the sums of shares do not equal to 100%.

The evaluative dimension

From the literature on the normative aspects of differentiation in the EU (Bellamy, Kröger, and Lorimer 2022; Fossum 2019), we derive three critical questions that underpin the public debate: (1) whether differentiation is acceptable, (2) whether it leads to domination, and (3) whether it is democratic. To assess how these controversial questions are addressed in news reports, we have coded the differentiation-related sentences in the reports as −1 for negative (unacceptable; leads to dominance; undemocratic), 0 for balanced (e.g. both unacceptable and acceptable statements in the same sentence), and as 1 for positive (e.g. acceptable) sentences. This allows us to analyse the evaluative dimension in news reports at the levels of reports (Figures 4 and A2) and sentences (Table 3).

Figure 4. Average evaluations in news articles. Note: error bars represent standard deviations.

Table 3. Ordinal logistic regression models.

	(1) Acceptability	(2) Dominance	(3) Democracy
Differentiation Type
Law-making	1.09**	0.46	1.27***
	(0.35)	(0.97)	(0.30)
Competence	0.44***	0.31	0.63
	(0.06)	(0.31)	(0.44)
Territorial	0.35	−0.65	1.40
	(0.21)	(0.37)	(0.93)
Rights	−0.09	−3.57	−2.19***
	(0.36)	(1.93)	(0.10)
Statement Source
Journalists	−0.19	0.53	−0.75
	(0.13)	(0.48)	(0.51)
Public Sector	−0.02	0.36*	0.07
	(0.18)	(0.14)	(0.32)
Private Sector	0.20	−12.87***	−15.62***
	(0.48)	(0.00)	(0.00)
Civil Sector	−0.43	−13.85***	−1.60**
	(0.34)	(0.00)	(0.57)
Citizens	0.28	−14.89***
	(0.36)	(0.00)
Media Origin
Denmark	0.99***	0.19	−0.89
	(0.30)	(0.62)	(0.90)
UK	0.71*	−0.68	−0.44
	(0.28)	(0.94)	(0.80)
Policy Area
Immigration	−0.11	1.83	−0.63
	(0.33)	(2.39)	(1.11)
Num.Obs.	3778	534	465

Standard errors, clustered at the outlet level, are in parentheses. Germany is the reference country. * p < 0.05, ** p < 0.01, *** p < 0.001.

Starting with the aggregate-level evaluations of differentiation, one important question is whether news articles are overall balanced in covering affirmative and critical aspects of EU differentiation or whether they are biased towards either positive or negative assessments. The results suggest that news reports on EU affairs are remarkably balanced, especially when the data is pooled across countries (Figure A2). The largest imbalance in evaluations is about dominance; with a mean score of −0.7 (standard deviation: 4.4), we find that news reports tend to include fewer positive statements about differentiation (that it prevents dominance), but the imbalance is less than one statement per article on average. Any imbalance is even less noticeable for the evaluations of differentiation in terms of acceptability (mean: 0.1, standard deviation: 7.1) and democracy (mean: 0.004, standard deviation: 3.0). Zooming in on the countries in the dataset, we find similarly balanced evaluations in articles published in media outlets from Denmark and the UK. However, as Figure 4 shows, news outlets from Germany stand out again as notable exceptions. Although the articles from Germany are on average balanced in terms of the relationship between differentiation and democracy, their mean scores are −2.8 for acceptability (standard deviation: 13.5) and −3.0 (standard deviation: 9.5) for dominance. In other words, in a typical article published in German outlets, there are approximately three more negative than positive sentences concerning the aspects of acceptability and dominance in relation to differentiation in the EU. These results suggest that the coverage of differentiation is rather critical in German newspapers.

Moving to the sentence-level analysis, we use ordinal logistic regression models to explore which factors correlate with evaluations of acceptability, dominance, and democracy in sentences related to differentiation in news articles. The results are presented in Table 3, with the number of observations ranging from approximately 500 for the models on the latter two aspects to 3,778 for the model concerning the acceptability of differentiation. These numbers reveal that a significant majority of statements (83%) imply an assessment of whether differentiation is acceptable, yet only a minority of these evaluations (19%) relate to aspects of dominance or democracy.

Regarding types of differentiation, our findings reveal statistically significant variations in the evaluations of acceptability and democracy within the EU. Specifically, we find that statements concerning law-making and competence-based differentiations tend to receive positive evaluations in terms of acceptability. For instance, if differentiation pertains to changes in the way legislative decisions are made in the EU, the results indicate a 1.1 increase in the expected value of acceptability on the log odds scale, after controlling for the other variables in Model 1. Moreover, such changes are also associated with improvements in democracy, as indicated by the statistically significant, positive coefficient for law-making differentiation in Model 3. In stark contrast, right-based differentiation elicits negative evaluations. According to the results, the statements on the changes in the EU citizens’ rights are linked to a decrease of 2.2 in the expected value of democracy on the log odds scale.

As regards the sources of statements, all but one statistically significant coefficient is negative. That one exception comes from the group of public-sector actors, such as politicians and bureaucrats. Whenever these actors are cited in news reports on EU affairs, their statements are associated with an increase of 0.36 in the expected value of dominance on the log odds scale. Given how this variable is coded, this suggests that public sector actors tend to support the idea that differentiation prevents dominance. In contrast, the results from statements by citizens and actors in the private and civil sectors point significantly to negative evaluations of differentiation in terms of dominance and democracy.⁹ For example, we find that statements by actors in the private sector are associated with a 12.9 decrease in the expected value of dominance and a 15.6 decrease in the expected value of democracy on the log odds scale. Notably, the statements by journalists themselves are not associated with any statistically significant change in the expected value of evaluations – echoing the results from article-wise averages (Figures 4 and A2) that news reports are rather balanced in terms of how differentiation is evaluated.

Confirming what is already evident in Figure 4, the regression analyses show that news reports from Denmark and the UK tend to be more positive regarding the acceptability of differentiation in the EU when compared to those from Germany. However, after accounting for the other variables in the regression models, no other statistically significant differences emerge between Germany and Denmark or the UK regarding dominance or democracy evaluations. Likewise, the results indicate that differentiation in the area of immigration is evaluated similarly to differentiation in other policy areas – an encouraging sign for the generalisability of our results. Even if we analysed news articles published around important events related to migration, it appears that our findings about how EU differentiation is presented and assessed in the could be applicable to other policy domains.

Conclusion

For citizens to effectively contribute to the debate on the future of the Europe (Oleart 2023), it is imperative that they gain a comprehensive understanding of the differentiated system of governance that characterises the EU. At this critical junction for the EU, and in the face of its increasingly differentiated and complex governance, the news media can have a crucial role to play. Our study has been motivated by the idea that the media coverage of differentiation can affect not only what the public knows but also what it thinks about the EU. While differentiation is as complex and controversial as it is central to European integration, journalists are best placed to function as translators and critical watchdogs of EU governance. They combine insights into the Brussels-based policy-making mechanisms with first-hand understanding of national politics and interests.

We therefore investigated whether and how differentiation is presented and assessed in newspaper articles on EU affairs. Our analysis focused on coverage in three ‘most likely’ case countries, where differentiation might be covered in the news: Germany, Denmark, and the UK. Sampling news reports from those published between 2015 and 2020 in nine media outlets, our qualitative coding was designed to identify the coverage of four types of differentiation, sources of statements, and evaluations of differentiation in terms of acceptability, dominance, and democracy. This comprehensive approach allowed us to explore the nuances of whether and how differentiation is communicated to the public, highlighting its relationship with democracy and balance of power within the EU.

We found that approximately a quarter of all sentences in news reports on EU affairs pertain to differentiation in the EU, four times the share related to democracy or dominance. The diagnostic dimension of our analyses suggests that there are noteworthy country-specific differences as to whether and which form of differentiation is covered in the news in different countries. German newspapers, in particular, have a denser thematic focus on the various forms of EU differentiation, where journalists themselves are the main source of differentiation-related statements. Additionally, the evaluative dimension of our analysis demonstrates that news reports are overall remarkably balanced in terms of the relationship between differentiation, democracy, and dominance. However, across news articles, certain nuanced variations become evident. For example, we also found that law-making and competence-based differentiations receive more positive evaluations, while right-based differentiation elicits negative ones. Public-sector actors are an exception, supporting differentiation as a prevention of dominance. The media’s emphasis on various forms of differentiation thus mirrors the complex and controversial nature of differentiation in the EU. At the same time, EU news coverage fulfils in this sense a critical watchdog function raising attention to the implications of EU differentiation and its possible impact on democracy, yet expressing plural opinion without clear biases in support or opposition of differentiation.

Far from a mundane observation, the result of our diagnostic analysis is both reassuring, as far as journalistic standards are concerned, and instructive for how future research and media policies may approach the role of journalists in communicating the process of EU integration to citizens. Combined with our evaluative analysis, our findings highlight the critical role of journalists as translators of EU integration, making the complexity of EU governance accessible, and providing the basis for a knowledgeable discussion of the implications for EU citizens. The challenge of translating intricate EU matters while contending with selection and framing filters underscores the continuing need for specialization and adaptability within the media landscape. Such specialisation is not needed to achieve broad public consensus towards either more or less EU differentiation, but to ensure that as broad a range of relevant evaluations are available and accessible in the public sphere. Our study reiterates the significance of media in forming public understanding and opinion on EU differentiation, emphasizing the necessity for balanced journalism that accounts for the complicated and often conflicting aspects of EU differentiation.

Acknowledgments

This work was supported by the EU’s Horizon 2020 programme (Societal Challenges 6: Europe in a changing world – Inclusive, innovative and reflective societies) under grant agreement NO 822419 (EU3D).

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1. Andorra, Monaco, San Marino, and the Vatican City are the states using the euro with the EU’s agreement. In addition, Kosovo and Montenegro have adopted the euro unilaterally.

2. Note that this timeframe coincides with the UK, one of our case countries, leaving the EU. Following the Brexit referendum on 23 June 2016, the UK entered a transition period on 31 January 2020, and finally left the EU single market and customs union at the end of that year.

3. The statistics on differentiation in this section originate exclusively from territory-based differentiation, a focus dictated by the existing knowledge in the literature, which primarily addresses this form of differentiation. Although our sampling decisions are therefore centred on a single form, we anticipate a degree of correlation between the occurrences of different forms of differentiation.

4. About 30% of these articles are co-authored by journalists.

5. The coder for the Danish language also helped with coding the data from the UK.

6. Categories for this or other variables were not exclusive; coders were instructed to select multiple categories for any given variable when applicable.

7. The entire dataset included 11,327 sentences in 351 articles from Germany 11,509 sentences in 313 articles from Denmark, and 19,434 sentences in 699 articles from the UK.

8. About 13% of all these sentences were also related to migration, despite the fact that the time frame for analysis has been constructed around five important events related to the topic of migration through internal and external borders of the EU. See Figure A1 for further details. All tables and figures pre-fixed with the letter A are in the online Appendix.

9. Note that we are unable to estimate a coefficient for citizens in the third model. This is due to the absence of any statements in the dataset from citizens that evaluate differentiation in terms of democracy.

Appendix A

— Further Analyses

Immigration

Figure A1 plots the share of migration-related sentences, among the sentences that are coded as including a statement on differentiation, democracy, or dominance in the EU. The time frame for analysis has been constructed around five important events related to the topic of migration through internal and external borders of the EU. Nevertheless, Figure A1 shows that only a minority of sentences were about this topic in all countries under analysis.

Figure A1. Share of migration-related sentences, among those coded as related to differentiation, democracy, or dominance in the first place.

Average Evaluations

Figure A2. Average evaluations per news article. Note: error bars represent standard deviations.

Appendix B

— Codebook

Method

Qualitative content analysis of text in newspaper articles.

Unit of analysis

A sentence.

Instructions

For each sentence within every news article, begin by coding the first variable. Proceed to code the subsequent variables for a particular sentence only if it contains a statement related to differentiation. If a sentence falls into multiple categories of a variable, indicate all applicable categories.

Variables

1. eu3d: Does this sentence include a statement about differentiation, dominance, and/or democracy in the EU?

• [0] No

• [1] Differentiation

• [2] Dominance

• [3] Democracy

Definitions:

Differentiation covers both differentiated integration and differentiated disintegration, while ‘differentiated integration refers to specific features of the EU integration process, such as multiple speeds, exemptions, opt-outs and opt-ins, and questions of variable geometry’ (Fossum 2019, 7). For detailed definitions of the types of differentiation, see variable 3.

Dominance is defined as ‘relationship or a circumstance wherein an actor (be that a person, an organization, or a collective) can be arbitrary interfered with and/or manipulated’ (Fossum 2019, 2). There are several possible types of dominance referring to actors’ ‘formal legal status; limits to or constrains on the actor’s choice options; vulnerability or susceptibility to external influences; deprivation (material and emotional such as sense of self-worth); lack of or denial or recognition; undue impositions; and forms of exclusion’ (Fossum 2019, 3).

Examples:

• • a powerful member state takes decisions informally without notifying those affected

• • ECB acts beyond its bounds through undertaking a monetary policy that effectively trumps national fiscal policy

• • Eurogroup side-lines parliaments

• • European Stability Mechanism is regulated by international (not Community) law and its decisions are not accountable to the European Parliament

• • there is a breakdown of coordination and governing no longer proceeds according to predictable rules but is the result of caprice and circumstances

• • an actor knows that their interests and concerns will be affected but they don’t know by whom, when and how

• • an asylum seeker is denied legal standing or access to legal recourse

• • a state is no longer recognised on a par with other states

• • a rights-holder is stripped of rights or their rights are ‘less worth’

• • an actor experiences material loss or negative distributive effects that can be traced back to a wilful act or structural-institutional arrangement and not some natural disaster

• • an actor is actively held down and controlled by another that it has not authorised

Democracy is an arrangement whereby those subject to the law are able to understand themselves as the authors of the law. A key requirement for that is that citizens have private and public autonomy. Autonomy in turn enables democracy to combine a principle of justification with an organisational form for the handling of common affairs (Fossum 2019, 12).

• 2. immigration: Is this statement about immigration as well?

  • [0] No

  • [1] Yes

  • [2] Unclear

• 3. differentiation: What type of differentiation is the statement about?

  • [1] Law-making

  • [2] Competence-based Functional

  • [3] Territorial

  • [4] Citizenship & Rights

  • [5] Unclear/Other

Definitions:

Law-making differentiation refers to the proposals that argue for change in the relation between law-making arrangements and the relations between the executive, legislature and courts at a given level of government (EU level, member state level, regional level). The claims mostly refer to make-up of political system. This is about how power is functionally organised at a given level of governing (horizontally). This is about the democratic nature and quality of the EU, as understood in the structure of the system of governing and how accountability is structured: who is accountable to whom?

Examples:

• • The EP (or, the Commission, Council) becomes more (or, less) powerful in the law-making process in the EU.

• • The law-making process involves more (or, fewer) actors at the EU level.

Competence-based Functional differentiation refers to the political system’s scope of competence and the degree of functional specialization: which issues and how many a governing system at a given level is in charge of, what kind and range of expertise it possesses, how that is organised, and how specialised this political system is. It focuses on the nature, range and scope of functions that are undertaken at a given level of governing (EU, national or subnational). It seeks to capture the role of expertise; the extent to which the EU is technocratic; the scope of expertise and possible built-in biases in the type of expertise that is available at the EU level. It focuses on the type of expertise and policy specialisation: how many agencies, what type of agencies and the relationship between EU agencies and EU directorates. This dimension includes focus on the type and range of policy instruments: regulatory, fiscal, and monetary, shedding light on the EU’s biases in terms of monetary union without a fiscal union; and the EU’s strong regulatory imprint and its weak redistributive ability.

Examples:

• • A competence, the authority to control the borders, is transferred to (or, from) the EU, from (or, to) member states

• • The EU establishes a new agency to accomplish a specific task.

• • The EU army is created.

Territorial differentiation refers to set-ups in which not all EU member states take part in a common policy or institution. It also refers to the selective participation of third countries in EU policies.

Examples:

• • Ireland is not being the party of Schengen Area.

• • The UK exits from the EU.

• • Sweden opts out of using the euro as its currency.

• • The EP passes a legislation that applies to only a subset of EU member states.

Citizenship & Rights differentiation refers to proposals that argue for issues connected with the nature and range of rights to persons, such as civil and political rights, freedom of movement, citizenship, etc.

Examples:

• • EU citizens can vote (and, stand for election) in local/municipal or EP elections only if they are resident in a different member state than their own, while the citizens of the recipient country can also vote in national elections.

• • EU citizens resident in another member state gain (or, lose) certain rights and/or obligations.

• • The EU creates new regulations for asylum seekers.

• • An EU member state creates a special category of rights and obligations for British citizens resident in their country after Brexit.

Unclear/Other category is for statements about types of differentiation that does not fit in any of the categories above. However, please note that the following examples are not considered differentiation:

• • An EU member state has a different culture, history, or identity than another member state.

• • EU member states disagreeing on how to address a problem. For example, Germany would like to open the borders to refugees while Hungary would like to close them.

• • The EU decides to sanction Belarus.

4. actor: Who does the statement originate from?

• [1] The reporter(s)

• [2] Public-Sector Actors

• [3] Private-Sector Actors

• [4] Civil-Sector Actors

• [5] Citizens

• [6] Unclear/Other

Definitions:

Public-Sector Actors include, for example, commissioners/ministers, politicians, MEPs/MPs, bureaucrats, and their spokespersons.

Private-Sector Actors refer to those who own, or work for, for-profit businesses and industries that are not controlled by governments.

Civil-Sector Actors work for, or represent, non-governmental and non-business organisations such as universities (including private ones), trade unions, and charities.

5. acceptability: Does the statement include a suggestion that differentiation is politically acceptable or sustainable?

• [0] No

• [1] Acceptable/Sustainable

• [2] Unacceptable/Unsustainable

Definitions:

An acceptable differentiation is one that is framed as satisfactory. It is at least good enough to be initiated or to be allowed to continue existing, given the conditions – even if not great or otherwise ultimately desirable.

An unacceptable differentiation is one that must be rejected; altered, or completely removed if already in existence.

A sustainable differentiation is one that would continue to exist (if initiated) at least over a period of considerable time, irrespective of the framer’s opinion as to whether it is (or, would be) acceptable or not.

An unsustainable differentiation is one that cannot continue to exist (even if initiated), irrespective of the framer’s opinion as to whether it is (or, would be) acceptable or not.

6. dominance: Does the statement include a suggestion that there is a positive or negative relationship between differentiation and dominance?

• [0] No

• [1] Positive relationship

• [2] Negative relationship

Definitions:

No refers to sentences that do not relate differentiation to democracy in any way.

Positive relationship occurs when differentiation leads to more or emergence of dominance. In other words, the more we have differentiation, the more we have dominance.

Negative relationship occurs when differentiation leads to less dominance or removes it completely. In other words, the more we have differentiation, the less we have dominance.

Examples:

• • [Positive relationship] The UK has a weaker hand in the Brexit negotiations due to its small size compared to the EU

• • [Negative relationship] A small EU member state is outside the eurozone, and therefore avoids being dominated by large eurozone states

7. democracy: Does the statement include a suggestion that there is a positive or negative relationship between differentiation and democracy?

• [0] No

• [1] Positive relationship

• [2] Negative relationship

Examples:

• • [Negative relationship] Due to the EEA Agreement, Norway is subject to rules decided by other states, but it cannot affect these rules because it is not on the table.

• • [Positive relationship] The UK leaves the EU because the majority of the electorate voted for Brexit in the referendum.