How British, German, and US Media Cover Crimea’s Annexation and the Donbass War

This article investigates how British newspapers covered Crimea’s annexation and the Donbass war. It shows that initially an “international conflict frame” was used as the dominant way to interpret Crimea’s annexation, which in about three months was succeeded by a “local conflict frame,” then used as the most prevalent way to make sense of the enduring Donbass war. Moreover, it shows that as the war on Ukrainian territory evolved, it changed from an issue for “us” (the West) to an issue for “them” (those parties involved in the conflict). This shifting understanding appears to stem from ambiguity about the nature of the Donbass conflict, which, in turn, is caused by Russia’s consistent efforts to deny their involvement herein. A follow-up article that compares these findings to how the conflicts’ are covered in the German and US media will be available soon.

Introduction

The relationship between Ukraine and Europe is currently as significant as it is delicate. In University College London professor Andrew Wilson’s words: “Ukraine is the largest, most important, and most contested” Eastern Partnership (EaP) state of the EU. Russia’s annexation of Crimea and the continuing war in the Donbass region are only increasing its significance, and thus also the challenge with which the EU, and at times the US, have to deal. The EU undoubtedly has a strong interest in the region’s stability and made several efforts to achieve a conflict resolution, with the Minsk protocol and the Normandy Format talks (both included leaders from Germany, France, Ukraine and Russia) as the most concrete steps in this direction.

However, a solution for what is sometimes called a “forgotten” conflict is not yet in sight. Moreover, there is also the—potentially dangerous—tendency within the EU to isolate itself from the EaP countries Armenia, Azerbaijan, Belarus, Georgia, Moldova and, crucially, Ukraine. This would turn them rather undesirably into eastern “buffer states,” rather than EU allies. 

Apart from populist/nationalist tendencies in several EU countries, which result in a more inward orientation, the EU may also be careful with interventions due to economic interdependencies between Russia and some of its membership states (e.g. Germany’s dependence on gas from Russia). Political and economic changes outside the EU borders may also affect how states and other stakeholders relate to Crimea’s annexation and the Donbass war. For example, the change from the Obama to the Trump administration in the US importantly altered its foreign policy, including its relationship with Russia, and thus also its stance to Ukraine.

This publication is the first of a two-piece article set on how Crimea’s annexation and the Donbass war are interpreted by those with a voice in public discourses in the UK, Germany, or the US. Rather than asking “what the true nature of Crimea’s annexation and the Donbass war is,” this study asks “how do people make sense of them.” The general research question is: How are Crimea’s annexation and the Donbass war perceived in British, German, and American public discourses, 2013-2019. 

This question will be addressed by studying newspaper articles from the countries under consideration using both machine learning and qualitative research methods. In the article that follows I will focus exclusively on British media coverage. A follow-up article providing a comparative analysis focusing also on German and US media will be available soon. 

This study betters our understanding of how dominant narratives of the war shifted over time, and it provides support for the need for western parties to re-evaluate their self-understood role in a conflict resolution, as recently advocated by prominent stakeholders from the international community. 

Why study media in general, and why study media on Crimea and Donbass in particular?

Before turning to the findings, it is helpful to consider the value of media research. Media outlets, most notably the established, daily newspapers, reflect what elites and other prominent people think and do, and research has demonstrated their “agenda-setting” power, for example, on other media and government policy. Newspaper articles contain various “voices,” of course from the author/journalist herself, but also from—both foreign and domestic—ministers, presidents, members of parliaments, representatives from businesses or NGOs, eye witnesses, various kinds of “experts,” and so on. Newspaper analysis allows us to tap into public discourses, the ideas they harbor, and the conversations, or conflicts, between such ideas [1].

Studying media from several countries can help us to better understand how those with a voice in public discourses make sense of Crimea’s annexation and the Donbass war, how they see their own role regarding this matter, and whether there are cross-national similarities and/or differences. More specifically, it allows us to ask how and when do Western media perceive Crimea’s annexation and the Donbass war as Western issues (i.e. issues pertaining to the EU and/or the US)? 

Data and methods

A more extensive description of the data and methods can be found in the method appendix in the bottom of this article. A short summary is provided here.

To understand coverage in the UK on Crimea’s annexation and the Donbass war, newspaper articles were retrieved from the Sun and Daily Mirror, two important daily, national (and paid) newspapers. Articles were collected if they contained the term Ukrain* and at least one of the following terms: Donbas*, Donetsk, Lugansk, or Crimea.

The final dataset is comprised of 538 newspaper articles and analyzed with topic models, a method that, like other machine learning techniques, has “the ability to learn without being explicitly programmed”. More specifically, topic models inductively identify clusters of words based on word co-occurrences across a collection of texts.[2] These groups of associated words can be conceived as particular “interpretative frames” within discourses that bring certain aspects of the social world into focus, while also moving the attention away from other aspects. This study used a 10-topic model which retrieved 10 distinct word clusters (see Table 1). The topic models were also used to identify newspaper articles that strongly associate with a particular topic, and these article subsets were then qualitatively examined.

Most relevant for the present study’s research question are topic 1, 2 and 4 (highlighted in bold in Table 1). Topic 1 (including words such as “world” “West” “Europe” and “NATO”) and topic 4 (including words such as “military,” “Yanukovych,” “Crimean,” and “Moscow”) are mostly associated with Crimea’s Annexation, as also can be seen in Figure 1 below. Topic 2 consists of terms such as “rebels,” “troops,” “fighting,” and “region,” and mostly relates to the Donbass war. It is important to note that topics represent patterns in word usage. Topics do not necessarily align with particular events, but of course they can, since certain terms may appear more frequently together in relation to certain events.

Figure 1 shows that topic 1 and 4 were mostly prevalent in the beginning of 2014, whilst the presence of Topic 2 increases from approximately May 2014 until early-2015. From late-2016, a small number of articles associated with topic 1 occasionally resurfaces. Figure 1 also shows that topics 1 and 4 are hard to separate, as they have a near-identical distribution over time (with the most important exception that topic 4 does not resurfaces by late-2016).

The analysis below will zoom into these topics, as well as the newspaper articles behind them. 

Main results

Topic 4 Crimea’s annexation—As the words “military,” “Yanukovych,” “Crimean,” and “Moscow” suggest, Topic 4 is linked to newspaper articles that discuss military action during Crimea’s annexation. For example, on March 1, 2014, a journalist from the Sun describes the following scene:

Earlier yesterday a fleet of Russian helicopters had been seen flying into the troubled Ukraine hotspot and convoys of military trucks appeared on roads – some setting up checkpoints to halt civilian traffic. Meanwhile squads of heavily-armed men wearing camouflaged uniforms openly began patrolling key areas. The soldiers wore scarves and balaclavas over their faces and carried no identification markings – but their uniforms and weapons were of Russian type.[3]

The article contains a rather clear description of the events, and, importantly, the involvement of Russian troops is taken as a given, even though soldiers “carried no identification markings”. Other articles, too, give unambiguous descriptions of what is described as the “Russian invasion” of Crimea. Similarly, on March 5, 2014, a Daily Mirror journalists reports that: “Russian warships from the Baltic Fleet blockaded two Ukrainian warships in their Black Sea berths in Sevastopol”.[4] 

At this point I won’t go deeper into the Crimean events itself, but instead will turn to topic 1 to discuss the invasion’s implications, as understood by international key stakeholders.

Topic 1 the international conflict frame—Being comprised of the terms “world” “West” “Europe” and “NATO,” topic 1 is most importantly associated with news that discusses the foreign significance of the events in Ukraine. Russia’s annexation of Crimea in particular evoked strong responses from representatives of EU countries (which included the UK, during the period under consideration) and the US. By March 11, 2014, a headline in the Sun reads: “New Cold War in battle for Crimea; Putin and co face financial and travel restrictions”.[5] References to a new “Cold War” are abundant and the calls for sanctions are widespread. Especially, Donald Tusk, then Poland’s Prime Minister, urged to take a clear position against Russia (a line with which he continued as the president of the European Council between December 2014 and November 2019):

Donald Tusk warned the EU will find it hard to take a stand against Russia’s leader Vladimir Putin because of Germany’s economic dependence on Russian natural gas. He urged German Chancellor Angela Merkel to make urgent decisions “to avoid the paralysis of Europe at a time when it must act quickly and adopt a clear stance”. Mr Tusk added: “The Ukraine issue is a matter of the future security of the European Union.”[6]

Russia’s annexation of Crimea was, similarly, firmly condemned in the UK by David Cameron, then member of parliament, and William Hague, the Foreign Secretary of that time, who said: “[t]here would be far-reaching trade, economic and financial consequences. It would bring the great danger of a real shooting conflict. There is no doubt about that”.[7] At times there were also breaches in the EU’s collective “hard stand” against Russia. For example, when “French president Francois Hollande defied the US and Britain to go ahead with the sale of two state-of the-art warships to Moscow in a £1 billion deal”.[8] There are only a few references to the Donbass conflict in newspaper articles strongly associated with topic 1 the international conflict frame, like in this 2015 article, where the two appear together:

Philip Hammond [then Foreign Secretary] … said: “President Putin’s actions – illegally annexing Crimea and using Russian troops to destabilise eastern Ukraine – undermine the security of sovereign nations of Eastern Europe.” His warning came as ex-Army chief General Sir Peter Wall claimed the increased threat from the Kremlin and Islamic State had caught the West napping.[9] 

Such unambiguous accusations and convictions of Russian involvement in the Donbass are rather rare, and mostly reserved for the annexation of Crimea. This conflict, as the key terms in topic 1 indicate, was lifted outside its local context, rendered relevant for international foreign affairs, and perceived as an issue that may even erupt into a new “Cold War”. Figure 1 also shows how this way of speaking occurs over time. It was most prevalent by March 2014, but it reappears late-2016, early-2017, and occasionally in 2018. This resurfacing of topic 1 will be discussed below. But let us first turn to topic 2.

Topic 2: the local conflict frame —The shift in discourse and use of language can be illustrated by the following sentence, published by the Sun in the summer of 2014: “Russia, which annexed Crimea from Ukraine after a disputed referendum in March, denies arming the pro-Russian rebels.”[10] The claim, whether (or not) pro-Russian rebels were armed by Russia, is not further investigated, and left unchecked. As the fights in the Donbass region continue, British journalists become ambivalent about Russia’s role in the conflict. Take the following text fragment as an example: “Russia, and Vladimir Putin, which the Ukrainian government and Western countries allege is supporting the rebels”.[11] The key word here is “allege,” defined in the Cambridge Dictionary as “to say that someone has done something illegal or wrong without giving proof”. In a similar vein, another articles states, “Mr Lysenko [then Ukraine’s security spokesman] also alleged a Russian military train with self-propelled mortars was in neighbouring Belarus,” and, the journalists adds, “[t]here was no independent confirmation of the claims”.[12] When separatist leaders publicly claim that they will receive (or have received) Russian support, this is also contradicted by Moscow:

Alexander Zakharchenko, leader of the self-proclaimed Donetsk People’s Republic, said the rebels were in the process of receiving 150 armoured vehicles, including 30 tanks, and 1,200 fighters who he said had spent four months training in Russia. Russia denies the accusations.[13]

Without further background checks by the journalists in question, the reader is left puzzled about the verifiability of the claims. Given that experts currently have established that Russian forces have been and are involved in the Donbass war (as will be developed below), we can understand Russia’s denial as the public spreading of disinformation, which effectively kept the ambivalence about their role in the Donbass war intact.

By 2015, especially the Daily Mirror still brings updates on developments in the Donbass region conflict, discussing interactions such as truces (and often violations of those) between “pro-Russian rebels” and the “Ukrainian army.” However, the question of Russia’s involvement is almost completely ignored and appears to be bracketed by those writing about the issue. Throughout the period 2014-2019, there are only a few clear declarations in both the Sun and Daily Mirror that the “rebels” or “separatists” unmistakably received Russian support, like in the example above where Philip Hammond in plain language said “Russian troops … destabilise eastern Ukraine”.       

The temporal resurfacing of topic 1—As mentioned, topic 1 is importantly but not exclusively linked to Crimea. Figure 1 shows that topic 1 resurfaces again during the 2016 US elections. The Daily Mirror summarizes the stakes of the competition between Trump and Clinton as follows:

WHEN the world’s only superpower goes to the polls, the rest of the world pays attention.

The result of the presidential election tomorrow has huge implications for global security, international trade and relations and climate change.

A victory for Donald Trump will result in a major recasting of the world order with America retreating from its role as the world’s policeman.

A win for Hillary Clinton will lead the White House to adopt a more hawkish stance on Syria and the Ukraine.[14]

Prior to being elected, Trump has vowed to change the US-Russia relationship, and by November 2018, shortly after Trump’s electoral victory, “Obama tried to ease fears about the future of the transatlantic partnership, especially on how the US would now deal with Vladimir Putin. He urged Donald Trump to ‘stand up to Russia’ in a stark warning.[15] However, with the accusations of Russia’s cyberattack on the US elections, commenters fear again for tense international relations.[16]

 Figure 1 shows that newspaper articles associated with topic 1 occasionally appear during 2017 and 2018. This is notably the case when in March 2017 the NATO is present “in Estonia, Lithuania, Latvia and Poland to strengthen 700 miles of Eastern Europe’s borders with Russia,” which was “to ensure there is no repeat of Russia’s overnight takeover of Crimea and parts of East Ukraine”.[17] The Crimea and Donbass conflicts reached again the news as they, for example, were linked to contention over Russia as the 2018 World Cup host,[18] or the assassination attempt of Sergei Skripal in the UK.[19] Hence, now and then the events in Crimea and Donbass reappear in the media landscape as a context to discuss new Russian-related controversies.

Despite some slight variations, the analysis did not indicate that there were marked differences in how the Sun and Daily Mirror covered the events under consideration (for more information, see method appendix below). 

EU and US issues?

So far this study zoomed into two dominant interpretative frames used to speak about the Ukrainian Donbass and Crimea conflicts. The first frame may be referred to as an international conflict frame (topic 1, which in turn is associated with topic 4), whereas the second is best understood as a local conflict frame (topic 2). Based on the topic models and qualitative data analysis of the newspaper articles, we may expect that the EU and the US are frequently connected with the first “international conflict” frame, but substantially less with the second “local conflict” frame.

A correlation analysis between a count of the terms “European Union” (including its abbreviations) and the “United States” (including its abbreviations), on the one hand, and, the count of newspaper articles that strongly associate to topic 1, 4 and 2, on the other hand, can confirm whether this is the case. To this extent, a table with daily time series data was computed, tracking the prevalence of these terms and the number of articles that strongly associate with these topics. Table 2 below shows the correlation coefficients which provide quantitative support for the findings so far: topic 1 and 4 are more strongly associated with both the EU and US, but, as they are replaced by topic 2, this association largely disappears.

Discussion and conclusion

This study started with the question how Crimea’s annexation and the Donbass war in Ukraine are perceived in British public discourses and, more specifically, how and when are they perceived as forming Western issues. The analysis demonstrated that an international conflict frame, as the initial dominant interpretative frame, was in about three months succeeded by a local conflict frame, then used as the prevailing way to make sense of the enduring war in the Donbass region. Eventually, news reports on the two conflicts largely disappeared. 

So, what can we make of this? It is important to note that the shift from an international to a local conflict frame can be regarded as rather troublesome. Maria Altshuller warns in the Harvard International Review that, as the Ukrainian conflicts disappear from the public’s consciousness, this obstructs a resolution and road to peace: “Given the West’s current indifference to the chaos in the Donbass region, eastern Ukraine may be doomed to become one of the long-term ‘frozen zones’ too often left in the wake of Russian interventions in its neighboring nations”.[20] 

Altshuller also points at comparable cases in which Russia sought to increase its influence in post-Soviet states by allying with local separatists, such as in Moldova’s Transnistria or Georgia’s regions of South Ossetia and Abkhazia. The newspaper articles on Crimea and Donbass make a number of references to these conflicts (i.e., 26 references to Moldova/Transnistria, and 24 to Georgia/Abkhazia). Nearly all of them followed directly after Crimea’s annexation in March 2014 due to fears about the continuation of Russian military force in these regions. However, as coverage on the Crimea and Donbass conflicts waned, the comparisons to similar Russian-related conflicts dissipated, too.

During the recent Ukrainian Lunch at the Munich Security Conference (February 16, 2020) Ukraine’s president Zelensky, too, emphasized that a conflict resolution required international attention: “[w]e realize how important it is for Ukraine not to disappear from the international radars, especially, with regard to security”. Wolfgang Ischinger, Chairman of Munich Security Conference, was also present at the lunch and observed that the Ukrainian conflicts appear to have fleeted from the minds of European audiences: “[h]ere in the European Union, we often claim that we have created a continent of peace and we do not conduct wars anymore. It is very important that you, Ukrainians, remind us and that we remind ourselves that it is actually not true. There is still a war going on, and we have to make a much greater, more intense collective effort to bring it to an end”.

As noted in the introduction, the goal of this study is to examine how people make sense of events, rather than investigating what really happened, which is not to say that the latter question is not important. Wilson provides a well-informed discussion of Russia’s role in the Donbass conflict, and there is the well-publicized and influential Bellingcat report about the source of the Buk missile launcher which took down flight MH17.

The present study draws attention to how ambiguity—which stems from Russia’s consistent denial of involvement in the Donbass war—may cause those people, i.e. those with a voice in public discourses, to disengage from an issue, and start losing a sense of responsibility. While the situation in Crimea was clearly covered by the press, the unclear nature of Donbass conflict is likely one of the key factors [21] that contributed to the discursive, and thus also cognitive, separation between “us” (the West) and “them” (those parties involved in the war). Contributing to this discursive separation are not only the possible hidden disinformation campaigns through (social) media, but also simply Russia’s public denial of involvement in the Donbass conflict, which can be understood as the open spread of disinformation.

Method appendix

This study focuses on news coverage from daily, national (and paid) newspapers. Two newspaper titles were selected in each country based on their outreach (those with the highest circulation figures) and political orientation (where the aim is to include one more “right-oriented” and one more “left-oriented” leaning title). This first research project, which focuses on the UK, is based on articles in the Sun (i.e. the largest British newspaper, and the largest right-center/populist title) and the Daily Mirror (i.e. the fourth largest British newspaper, and the largest left-center title).

An “open” search strategy was used by collecting every article that contained the term “Ukrain*” and one or more of the following terms: “Donbas*,” “Donetsk,” “Lugansk,” or “Crimea”. The asterisk is used to capture possible variations (e.g. Donbass or Donbas). After removing duplicates and articles that were very clearly about football (e.g. FC Shakhtar in Donetsk) but not about one of the two conflicts, this resulted in a final corpus of 538 newspaper articles (271 and 267 from the Sun and Daily Mirror, respectively, see also Figure 2). A small of amount of football articles still remained present in the final dataset, but these did not interfere with the analysis as they were clearly demarcated by topic 8. Also note that some of these football articles still contained a reference to the conflicts.

Figure 2.

The final dataset is analyzed using two methods. First, topic models, a machine learning technique, are used to identify different themes, or so called topics, in large collection of texts. Topic models are based on the LDA algorithm which inductively identifies themes based on patterns of word co-occurrences across multiple texts. The only parameter which is set by the researcher is the number of topics that the model yields. Given the relatively small size of the dataset, this study uses a 10-topic model, represented by a list of the top-30 words for each topic (as can be seen in Table 1 in the body of this article). The analyses are conducted in Python and draw, among others, on the Scikit-learn machine learning library. 

Apart from the topics themselves, the model also produces a so called “document-topic matrix,” showing for each document its probability of belonging to one of the topics. To assess how prevalent a topic is within in a corpus (i.e. a collection of texts), one can calculate the average probability of the entire corpus of belonging to that topic (see Table 3). Table 3 shows, for example, that topic 1 and 2 have a relatively strong presence in the corpus, whereas topic 5 and 6 are not very present. Those documents with a probability higher than documents’ average probability plus 1 standard deviation (SD) are considered as articles that “strongly associate” to a topic under consideration (for the cut-off points, see Table 3). These calculations are also used to plot topics’ prevalence over time, and to compile newspaper article subsets, which were analyzed qualitatively (see section below).

 Secondly, the newspaper articles were also analyzed qualitatively, that is, they were read and manually annotated, focusing on how language was used both to uncover and cover certain information about the conflicts, as well as other possible interesting patterns.

As for differences between the two sources used, a distinguishing feature of tabloid the Sun is, as could be expected, formed by its sensationalist and informal wording, most notably in its headlines, which for example spoke of “Vlad’s army” or “Putin’s lap dog” (a reference to Yanukovych). In addition, Figure 2 shows that the Sun ceased its publications on the issues by the end of 2014, whereas the Daily Mirror still published a few articles in the period 2015-2019. Apart from these differences, the analysis indicates that both the Sun and Daily Mirror covered the events in a similar way. For example, the contentious term “civil war,” was used 12 times by the Sun, and 11 times by the Daily Mirror. There were 111 references to the EU in the Sun, compared to 85 in the Daily Mirror. Some terms, however, appear to be somewhat more source-specific, such as the references to the “Cold War,” which appeared 13 times in the Sun versus 32 times in the Daily Mirror. Most importantly, articles from both sources were underlying the topics this study focused on. Illustrative of this is Table 4 below, showing the top 10 articles associated with topic 4.