Intersectional solidarity, empathy, or pity? Exploring representations of migrant women in German and British newspapers during the pandemic

ABSTRACT

This research examines migrant women’s representation in British and German news during the COVID-19 pandemic, paying particular attention to expressions of solidarity. While previous research has focused on the representation of women and migrants, studies have largely treated these groups as distinct entities. Combining topic modelling with informed grounded theory, this research compares the representation of migrant women, including women from ethnic minorities, across a diverse corpus covering 2020–21. This two-step methodology facilitates intersectional analysis at the macro and micro level, revealing the salience, topical concentration and context of migrant women’s coverage. Findings show that solidarity emerged through the platforming of migrant women’s voices, and connected to discourses of class, poverty and precarity. Representations containing empathy and pity – less likely drivers of social change – were more prevalent. This research contributes to our understanding of migrant women’s coverage during the pandemic and provides a granular and intersectional lens for analysing minoritized group representation.

KEYWORDS:

• Intersectionality
• topic modelling
• informed grounded theory
• solidarity
• empathy
• representation

Introduction

The pandemic was, and continues to be a period of great social, economic, political and cultural upheaval. COVID-19’s devastating and multifaceted effects on migrant women across the world include negative impacts on health, job security, legal status and living conditions (Kantamneni 2020; Paton et al. 2020). Hegemonic and institutionalized systems of sexism, racism and classism mean that the heterogeneous group of migrant women were more strongly affected by the pandemic; this statement holds across many nations (Crenshaw 2020). Despite the extent and varied dimensions of discrimination during COVID-19, little attention has been paid to whether or how migrant women were represented during this period in the news media: were their voices, narratives and experiences platformed? This matters because we know that representation, regarding both the extent and nature of coverage, can define audiences’ perception of and attitudes towards social groups via parasocial contacts (e.g. Appel and Weber 2021; Tukachinsky, Walter, and Saucier 2020).

Historically, pandemics have “forced humans to break with the past and imagine their world anew”, serving as a “gateway between one world and the next” (Roy 2020, 3). Following Van De Wiele and Papacharissi (2021), drawing upon Turner’s (1969, 1144) notion of “liminality”, a pandemic serves as a “long moment of in between”, where hegemonic social structures can be “disrupted and reimagined”, and where “new possibilities for agency and resistance are generated” (Van De Wiele and Papacharissi 2021, 1143). In such moments, traditionally less-visible actors such as migrants also have opportunities to be heard in public discourses, particularly through the news media.

We know that during the COVID-19 pandemic there was a shift in news representations. Front-line workers garnered more positive news coverage, and moments such as “clap for carers” were (temporary) disruptions in these hegemonic patterns of visibility. For these reasons, despite the traditional lack of voice assigned to migrant women in the news, we anticipate that during the pandemic, a shift towards more positive representations, including expressions of solidarity, might occur. It is important to understand the nature of migrant women’s coverage because minoritized group visibility is deeply connected to matters of social cohesion and the potential promotion of empathy and solidarity among groups (Walter and Glas 2024). Therefore, this study asks:

RQ1: In what ways were news media expressions of solidarity articulated towards migrant women during the COVID-19 pandemic?

RQ2: How can computational and qualitative methods help us understand expressions of intersectional solidarity with migrant women?

Speaking to the broader theme of this special issue on minoritized group representation during the pandemic, this paper combines computational and qualitative methods to conduct an intersectional analysis at the micro and macro level, in a multilingual corpus of British and German news. Doing so, this study contributes to the limited literature studying representations of migrant women from an intersectional perspective, expands knowledge regarding articulations of solidarity, empathy and pity in “positive” media coverage, and ultimately complicates our previous understandings of inclusive representation.

Migrant women in the news

Broadly speaking, studies of news media representation typically target “women” or “migrants” as distinct groups (Eberl et al. 2018; Van der Pas and Aaldering 2020). Across many countries, studies have shown that migrants tend to be underrepresented in the news, while being overrepresented in negative news, often framed as a threat, or in relation to criminality (see Eberl et al. 2018). There is variation across groups, with migrants from Muslim backgrounds receiving the worst coverage (Bleich and van der Veen 2021). Comparably, despite improvements in the proportion and nature of women being covered in political stories (Humprecht and Esser 2017), recent findings reveal sustained, gendered imbalances in terms of age and quoting women experts (Ross et al. 2018). Moreover, high-profile women are typically framed in relation to their status as a mother, or other caregiving role (Vandenberge 2019; Garcia-Blanco and Wahl-Jorgensen 2012).

Our understanding of representation for those who are represented as both “migrant” and “women” is limited, as several overlapping and interlocking systems of exclusion amalgamate around these identities. Focusing on “individual” labels leaves blind spots regarding identities and experiences; there is a need to pay attention to nuances which further compound exclusion. Regarding the German context, research has identified variation between outlets with left-leaning media being the most likely to cover migrant women (Lünenborg, Fritsche, and Bach 2014). Furthermore, while men are mentioned in nearly every article on migration, women appeared in only 12–26 per cent of these news stories (Lind and Meltzer 2021). Extant research in the United Kingdom (UK), on the other hand, found that TV and legacy news, rather than underrepresenting migrant women, covered them extensively in homogenizing, dehumanizing and stereotypical ways (Silva and Mendes 2009).

Our understanding of the nature of migrant women’s news coverage is generally limited. Given the disproportionate negative impacts of the pandemic on this diverse group, this is concerning (Kassova 2020; Laster Pirtle and Wright 2021). Nevertheless, as an unprecedented event, there is reason to believe that the COVID-19 pandemic could have profound implications, and a growing number of studies address women’s representation in this context. We have, for example, seen a low level of female experts in pandemic news (Jones 2020), whilst other intersectional identity markers such as age have been foregrounded (Adlung and Backes 2023). In representing issues impacted by gender, class and race, news media “cover these issues as much as they conceal them” (Lünenborg, Reißmann, and Siemon 2023, 86). Furthermore, the frequent use of masculine metaphors around war and conflict in pandemic coverage seems to limit women’s visibility (Williams and Greer 2023).

Intersectional solidarity: a lens for studying representations

Studying the coverage of migrant women in a meaningful way requires a nuanced, critical and contextually informed approach, accounting for overlapping lines of exclusion and inclusion. Following Crenshaw (1991), an intersectional perspective sheds light on how certain groups and identities come to have their exclusion compounded within a given social structure. Crenshaw identifies three types of intersectionality, political, structural and representational. The latter is most useful here.

The processes through which interlocking oppression is sustained can be messy, and many of our current methodological approaches fail to capture this complexity. Instead, two strands of intersectional studies have emerged; those tracing macro-patterns (quantitative), and those focused on micro-contexts (qualitative). Through adopting McCall’s (2005) inter- and intra-categorical approaches to intersectionality, we combine both levels of analysis. Intercategorical intersectionality sees existing social categories and labels like “gender” and “migrant” as (at least partially) useful; this approach lends itself well to quantitative studies. The more critical and qualitatively inclined approach of intra-categorical intersectionality encourages analysis of precisely how such categories are distinguished within and beyond language, acknowledging the ever-changing nature of social categories and their relationship to one another.

Within social movement literature, solidarity is defined as promoting and working towards collective aims in the political sphere, established by emphasizing “shared fate”, or “shared values” with different social groups and identities (Einwohner et al. 2021). Recent work has pushed beyond defining solidarity as “action”, instead moving towards a discourse perspective (Alharbi and Rucker 2023). Within journalism studies, solidarity is understood as an enduring but scarcely recognized news feature, or value (Varma 2022). This “logic of representation” aims to expose structural factors that contribute to the marginalization of certain groups, highlighting potential solutions and encouraging audience action (Varma 2022). There are different degrees of solidarity expressed within the news, ranging from short-term, single issue-focused “instrumental” solidarity, in contrast to the broader, longer-term matter of transformative solidarity. “Empathy” is a connected but distinct concept. Varma (2022) separates solidarity’s promotion of radical inclusion, from empathetic representation, which offers a narrative of individual exceptionalism and shared traits. Representations adopting this approach might make a marginalized individual “relatable” to their audience, or foreground outstanding achievements.

To study expressions of solidarity and empathy towards migrant women, Ciccia and Roggeband’s (2021) framework for intersectional solidarity offers a starting point (c.f. Beazer et al. 2023). This draws a distinction between solidarity based on “Common Denominators” (e.g. identities, experiences, etc.) and the “Recognition of Difference” (foregrounding the subjectivities and unique experiences, identities and issues). This study draws upon an analytical toolset to examine the composition of marginalized group representations in the media (Table 1). Dimensions of representation – e.g. scale of representation, constructed relationships, and level of context – have been adopted from Ciccia and Roggeband’s (2021) and Varma’s (2022) conceptualizations.

Table 1. Intersectional solidarity in the news framework.

	→ POTENTIAL FOR SOCIAL CHANGE →
Logic of representation	Empathy	Instrumental solidarity	Transformative solidarity
Scale of representation	Individual	Individual or collective	Collective
Constructed relationship to audience	Common denominators in identity OR role models/outstanding achievements	Marginalized subjectivities
Conditions of minoritized groups inclusion	Shared identity OR outstanding achievement	No conditions
Context	Personal life Mainly short-term Single issue	Mainly focus on individual/single group Short term Single issue	Broad context Connects to other issues/groups Mainly long-term, multi-issue
Tone	Emotive	Explanatory and analytical
Aim of representation	To be relatable to the audience or provoke an emotive response in the audience	To promote action To promote social change To elucidate structural factors To promote radical inclusion
Types of article	Stories of: Personal resilience Overcoming challenges Relatable stories	Calls for: Political Action Recognition of a single group/individual regarding a single issue	Calls for: Political Action Recognition of minoritized group(s) regarding a shared, often enduring issue

Methodology

This study examines representations of migrant women in the UK and Germany, including news outlets from two European countries with different migratory histories, but similar experiences of gender equality, current migratory groups, and COVID-19 impact. Indeed, in both countries, migrant women were among the most negatively impacted groups (Germain and Yong 2020; Willers and Barglowski 2023). By analysing a diverse corpus of minority and mainstream news outlets, including quality and popular newspapers from across the political spectrum, we address existing questions of variation between mainstream and minority news sources (Freudenthaler and Wessler 2022). Drawing upon classifications established in previous research, our full sample is displayed in Table 2 (Berry, Garcia-Blanco, and Moore 2016). This further enables exploring the nature of positive representations comparatively, identifying the contexts under which different articulations of solidarity and empathy emerge.

Table 2. Classification of news outlets included within the sample.

	DE	UK
Mainstream Right	Die Welt and Welt Online Bild, Bild Online^a	The Daily Telegraph and Telegraph Online The Sun and The Sun Online
Mainstream Left	Süddeutsche Zeitung	The Guardian and Guardian Online
Minority News	Islamische Zeitung Neuland Zeitung^b	Asian Image Muslim News

^a Including Bild Plus, i.e. online content behind paywall.

^b Unlike the other minority news outlets which cater to different religious and ethnic groups respectively, the Neuland Zeitung is written by and is for migrants within Germany from all national backgrounds.

For the sampling of relevant news stories, native and bilingual English-German speakers created a migrant-related keyword search string, semantically similar in English and German (Appendix 1) to retrieve 18,978 migrant-related news articles published 2020–21. Articles were accessed using LexisNexis and web scraping. 18,527 news articles included in the corpus were from mainstream news outlets, while 451 were from minority news sources. In terms of national balance, 9,716 articles were sampled from the UK and 9,262 articles came from German outlets.

Mixed-methods approach

Given this study’s ultimate aim to explore articulations of solidarity in coverage of migrant women across a diverse corpus, the analysis is composed of two main phases at the macro (quantitative) and micro (qualitative) levels. Firstly, informed by Nelson’s (2020) computational grounded theory approach, we adopt a broad, thematic analysis using the computational approach of topic modelling. Then, we conduct a more fine-grained analysis using the qualitative approach, Informed Grounded Theory (IGT) (Nelson 2021; Thornberg 2012). An overview of the approaches can be found in Figure 1. Importantly, due to the differences in sample size, topic modelling is only used for the mainstream news sample. We chose a manual classification approach for the minority news sample, explained in greater depth below.

Figure 1. Summary of methods and sampling.

Topic modelling and topic classification

Topic models are widely applied across social science to automatically identify latent themes (or topics) in large text collections (Chen et al. 2023). To improve the classification of topic models, scholars have called for joining computational and qualitative methods (Nelson 2020), as we do here. Specifically, we use topic models to highlight the broader themes present in the coverage of migrant women during the pandemic across the entire corpus, and therefore guide us towards a theoretically relevant sample.

As an accurate and cost-effective approach for bag-of-words text analysis, we used machine translation using the DeepL API to translate German news to English (see Lind et al. 2022). Analysis used the quanteda (Benoit et al. 2018), Deepl and tidyverse R packages (Wickham et al. 2019). Having identified sixty as the optimal number of topics (see Appendix 2), we reviewed the results and identified forty-three interpretable topics. After labelling the topics, the sampled texts were grouped into six larger themes using hierarchical cluster analysis (Puschmann and Scheffler 2016). Finally, regarding the minority news texts included in the sample, using the topic clusters from the mainstream news media analysis as a starting point, two researchers manually coded and categorized the articles’, whilst remaining open to the emergence of unique, new themes.

Qualitative analysis

Of the six clusters identified through the topic model analysis, the cluster labelled “Management and Unequal Impacts of Covid-19” (N mainstream media = 6121; N minority media = 57) was most relevant to the research question. Besides containing the most relevant articles of migrant women during the pandemic, the cluster featured cases of solidarity and empathy, which we examined by through closer analysis. We narrowed these articles down by employing a second multilingual keyword search using gender-related keywords and terms uniquely relevant to female migrants (see Appendix 1). The two-step sampling process enabled us to reach a more relevant sample and gain a comprehensive overview of migrant women’s salience in articles on migration, broadly speaking.

Within the cluster, “Management and Unequal Impacts of Covid-19”, after the keyword search, articles were randomly sampled when they scored above a set threshold for relevance.¹ Out of the resulting 173 mainstream articles, 30 articles were excluded because migrant women were mentioned as fictional characters. Hence, the final qualitative sample consists of 143 articles from mainstream and 57 articles from minority media outlets.

The qualitative sample was analysed using IGT. Grounded Theory can generate new theories grounded in data by moving between analysis (both open and focused coding) and data collection, code creating and memo writing (Glaser and Strauss 1967). IGT builds on this by utilizing a previously defined theoretical framework – in our case, intersectional solidarity (c.f., Table 1) – to inform analysis, while allowing for modifications to the coding scheme in subsequent iterations of analysis (Charmaz 2006; see Appendix 3). Open codes were formed from the intersectional solidarity framework, with a focus on actor descriptions, the use – and tone – of direct (or indirect speech) and the context. Articles were analysed iteratively until saturation was reached.

Word embedding

Following this, we use word embeddings to demonstrate pattern confirmation from the qualitative findings, by analysing news articles from the topic cluster “the Management and Unequal Impacts of COVID-19” (N = 6,121). Simply put, word embeddings are numerical representations created using co-occurrences of words in a corpus, which can be useful for illustrating relationships between different words (Mikolov, Yih, and Zweig 2013). In this paper, word embeddings allowed us to identify differences in the language used to represent female vs. male migrants, thus adding an additional layer of intersectional analysis.

To discover words that were related to female migrants in our corpus, we created Word2Vec embeddings using the GloVe algorithm with 100 dimensions and a window of 5 for words that occurred more than 10 times. We created a list of words for females (e.g. woman, mother, girl), males (e.g. man, father, boy), and migrants (e.g. immigrant, migrant, foreign). The category vectors were the mean average of the list of vectorized words for each category (see Appendix 1). Pairwise combinations of the vectors for each category were then averaged to create the intersectional identities (females + migrants), and compared across different identities.

Results

Quantitative analysis

To analyse expressions of solidarity and empathy in migrant women’s representations, we focused on a corpus comprised of migrant-related articles from mainstream and minority news. The reason we focused on migrant, as opposed to migrant-women-related articles from the offset is because we wanted to explore intersectional identity dimensions within a broader context. Using topic modelling followed by a cluster analysis, our findings revealed six general themes of mainstream news coverage of migrants in the corpus. Table 3 shows a brief description of each of these themes.

Table 3. Results from topic modelling and cluster analysis of mainstream news (N = 18,527).

Label topic cluster	N Topics	Description
Border crossing	17	Describing perilous journeys and refugee camp conditions
Management and unequal impacts of Covid-19	11	Describing impacts on e.g. work, health, treatment, vaccination
Politics	8	Covering politics e.g. regarding social/economic issues, EU policy, international court decisions
Refugee and migrant integration	4	Covering e.g. migrant/refugee experiences of integration/inclusion
Risks and danger in crossing borders	2	Covering mistreatment/detainment/asylum
Arts and culture	1	Covering art/cultural content and events

Figure 2 is a scatter plot that includes a point for each topic cluster in the mainstream news analysis. The x-axis shows the total number of distinct articles that were classified into each of the clusters. The most occurring topic cluster is “Border Crossing” (N = 11,176 articles), followed by “Management and Unequal Impacts of Covid-19” (N = 6,121 articles) and “Politics” (N = 5,873 articles). The y-axis shows the percent of distinct articles in the topic cluster that included female migrant keywords. The “Border Crossing” and “Management and Unequal Impacts of Covid-19” had a similarly high proportion of articles that included both female and migrant keywords (38.4 and 36.5 per cent, respectively).

Figure 2. Topic clusters and share of articles with female migrant keywords.

Due to the much smaller minority news sample, rather than running topic models, articles were manually coded across the clusters identified in the mainstream analysis. Table 4 shows the number of articles across each cluster, where we see a very different distribution of articles in comparison to the mainstream sample. Strikingly, “Politics” is the dominant cluster, containing 266 articles, whereas articles relating to border crossing are very low. This reflects, potentially, the different audience addressed, different ownership and agendas.

Table 4. Results from topic classification of minority news and share of female migrants (N = 448).

Cluster name	Percent of articles (N)
Border crossing	3.1% (14)
Management and unequal impacts of Covid-19	12.7% (57)
Politics	59.4% (266)
Refugee and migrant integration	15.2% (68)
Risks and danger in crossing borders	4.9% (22)
Arts and culture	4.7% (21)

Qualitative findings and discussion

Focusing on articles from the “Management and Unequal Impacts of Covid-19” cluster enabled us to conduct more in-depth, micro-level analyses. Guided by our central research questions, we begin by discussing examples of the most inclusive forms of representation (“transformative and instrumental solidarity”) and move towards discussing empathetic representations, before outlining the emergent representations of pity. This intersectional analysis pays special attention to context and power dimensions in each representation.

Transformative and instrumental solidarity

Transformative solidarity was found in a range of articles across topics. This mode of representation focuses on political and structural context, and is based upon the logic of radical inclusion, with its ultimate aim of inciting social action (Varma 2022). Across the corpus, radical inclusion was extended to a range of groups, including one frequently overlooked group disproportionately impacted by the pandemic: students. This is apparent in one Guardian article (14th December 2020), where the author, a female migrant, connects her experiences as an international student with other overlooked groups including “temporary visa holders”, “backpackers, refugees and asylum seekers”. She refers to these groups using collective pronouns: “they are now unemployed, in danger of homelessness and lining up at charities for food vouchers just to survive” (emphasis added). The discursive construction of solidarity is made not only with whom, but in relation to whom: “They wanted us to come here, but in a crisis, we feel abandoned. It is deeply hypocritical” (emphasis added). Though the author recognizes marginalized subjectivities and differences in experience between groups, the issues common to multiple marginalized groups is emphasized through collective pronouns that establish a cleavage between those who are exploited and lack access to services, and those who hold power; in this case the government.

Similar patterns can be seen in one Muslim News article (13th May 2020), discussing the complex causes of higher levels of COVID-19 related deaths among migrant women in the UK. Here, the case of one migrant woman, Belly Mujinga; “the ticket officer who died after being spat on by an infected passenger” is highlighted as just one example of the increased risks facing minoritized groups. The article reports, “[t]here is a concern that they will find reasons to blame BAME communities for their own deaths, with these deaths being pinned on cultural and genetic factors” (emphasis added). Solidarity appears in the construction of connections across different ethnically minoritized groups, who all suffer from structural inequality and the racism which perpetuates this.

We also see solidarity emerge within representations of migrant women when journalists include migrant women’s voices and analytical perspectives. Migrant women’s voices are included, for example, in articles relating to the so-called “refugee crisis” in articles discussing how the pandemic worsened conditions for refugee women and their communities. In Die Bild article (4th February 2021), Ms. Jihan Khodr, a Lebanese-born German resident and volunteer in refugee aid, is quoted in her address of the former Chancellor Angela Merkel. She is quoted as follows: “I have a lot of contact with refugee women and mothers. There were always problems, but it was never as visible as in the Corona times. We want to help our children more!”. Although the journalist does represent her in emotional terms (“she burst into tears”) Khodr’s own voice dominates her representation within the article.

In the Munich-based minority news outlet Neuland Zeitung, solidarity towards migrant women emerges not through quotations, but by providing a platform for women to write about their personal perspectives on the integration process. As such, solidarity emerges not only through representations within articles, but through the outlet’s design, giving voice to women to describe their experiences without paraphrasing, or journalistic interpretation. These are found in articles discussing linguistic and cultural-related challenges (Article One, Issue 1, 2021) and reflections on the experience of Lockdown (Article Two, Issue 1, 2021). Whilst many focus on structural challenges, there are also hopeful pieces (Article Two, Issue 1, 2021). The deeply personal reflections in Neuland Zeitung shine a spotlight on personal narratives, assigning space for critique, analysis and deliberation to the migrant women writing the articles, rather than treating them as news subjects to be covered in a more distant voice.

Combined modes of representation: solidarity and empathy

Previous work on representations of marginalized groups has distinguished between empathetic representations at the level of the individual, versus politicized, collective representations which “enact solidarity through a technique of radical inclusion” (Varma 2020). This study finds the binary distinction between individual (empathetic) and collective (politicized) representations is not clear cut. Many articles use the stories of individual women as a means to communicate a message of radical inclusion and solidarity. One Guardian article (1st September 2021) focuses on an organization for hotel cleaners in Spain, founded as a union-style organization, developing an app to support these women. One migrant woman involved in this organization is quoted, describing the extreme circumstances many women face:

The women were forced to survive on food banks and charity from community groups and the church. […] “I couldn’t claim furlough because my husband was receiving it, €900 a month, and I only say that because I’m one of the lucky ones”, Atana said.

Drawing on her own experiences, anecdotes of others, and historical context, Atana describes a complex experience of simultaneous intersectional advantage and disadvantage, with the key underlying point: that hotel profit underpins this exploitation. Atana also describes the nature of the group: “the majority are immigrants from Latin America, eastern Europe and Africa” She later adds, “they [business owners] prefer single mothers because they’re easier to exploit” (emphasis added). As intersectional markers, these offer indications as to who the most vulnerable women are, tied to place of birth, and gendered family roles. At the end of the article, Atana calls for collective solidarity in a direct mode of audience address: “if you’re looking for a hotel, look for one where there are humane working conditions”. This provides an example of a migrant woman’s voice being included both actively, as she calls for change, and critically, when represented in mainstream news.

Coverage of the exploitation of female migrant workers in a variety of precarious sectors extends across the sample. For example, an article from the Guardian (12th May 2020) highlights the case brought against one migrant woman, Rosa, who was held liable for spreading COVID-19. The article begins by describing her as “a legal permanent resident from the Dominican Republic whose children and husband live in New Jersey”. These intersecting identity markers make her relatable to the audience (her role as a mother, her background) and legitimize her inclusion (being “legal”); as such, the representation could evoke empathy in the audience. Rosa’s voice is included within the article, and she is quoted in predominantly emotive terms (“all my children were suffering watching me cry at night […] I saw everything gray”). Here there are elements of the empathetic mode of representation in the usage of personal, emotive quotes. However, the representation also includes context regarding the history of disease criminalization in the US. For example, it notes, “you can go all the way to ‘Typhoid Mary’, an Irish immigrant woman who was detained for many years after being accused of spreading the disease”. This socio-historical context, displaying the discriminative roots of the accusations made against Rosa, allows us to see this representation as employing elements of transformative solidarity.

Empathy

Throughout the corpus, migrant women appear frequently within individual profiles. A Guardian article (1st July 2020) focusing on Ernesta Nat Cote and her experiences working as a cleaner in Lewisham Hospital during the pandemic describes her first in reference to her profession (“a cleaner at Lewisham Hospital”), then via her experience and family background (“At 57, she is older than many of her colleagues, a grandmother”), her hardworking nature (“an old timer in a workforce where people tend to come and go”, “she applies an obsessive rigor”), her popularity (“everyone knows Ernesta”), her personality (“buoyant with confidence”), and finally her migrant background (“born in Equatorial New Guinea”). This reflects the representational logic of empathy in two ways: firstly, by outlining the many points of a possible connection with readers, and, secondly, by constantly foregrounding Ernesta’s exceptional and outstanding work.

The contexts in which representations emerge must also be considered. The intersectional positioning and use of identity markers in covering Ernesta implicitly reference class-based distinctions that shape perceptions of inclusion and exclusion in the UK. Ernesta is quoted: “I’m a cleaner, but I don’t put myself like a cleaner. The way you put yourself is the way they’re going to take it”. There are, nevertheless, points of contrast within the article that suggest dynamics of exclusion. Ernesta is referred to by her first name, while a second migrant woman represented in the article – a white, Irish migrant woman – is referred to by her surname. While no conclusions can be drawn on this based on one example, the contrast raises questions regarding the representation of migrant women of different backgrounds in the same piece. It brings to mind racialized hierarchies in the representation of migrant women which have been found in previous studies (Slakoff and Brennan 2023). Regarding the inclusion of Ernesta’s voice throughout the article, she is quoted describing the broader situation, with most direct quotes reflecting her feelings, attitude, and experiences in emotive and descriptive terms (“I’m very blessed. God is good for me”). Key, analytical perspectives, however, are articulated by the journalist. One of the defining factors of representations containing empathy is the inclusion of groups or individuals based upon certain conditions, such as outstanding achievements or shared identity markers. Here, Ernesta’s inclusion is largely based upon the notion of individual exceptionalism, rather than radical inclusion.

Within the corpus, conditions of inclusion in the representation of migrant women emerge along temporal and spatial axes. Simply put, the “acceptability” of certain groups of migrant women was conditioned in some outlets by historical trajectories (women who migrated further back in history, in contrast to contemporary migrants; those who are “legal”, versus undocumented) and spatial patterns (migrant women being mistreated in other countries being framed differently than domestic cases). For example, the historical conditions of inclusivity can be seen in the obituaries of migrant women. In the UK, this is apparent in coverage of migrant women from the “Windrush Generation” (see: “Paulette Wilson”, The Daily Telegraph, 31st July 2020). In German outlets, this can be seen in the coverage of Nükhet Kivran, a former chairwoman of the Munich Advisory Council for Foreigners (8th September 2020, Süeddeutsche Zeitung). Whilst Kivran’s migrant background is highlighted (“she came to Germany from Turkey at the age of 15”), her outstanding qualities and work on behalf of others are emphasized: “she campaigned tirelessly and successfully for the integration of people” (emphasis added). Elaborate descriptions of the work of important figures in obituaries are what defines this genre; however, there remains a striking, and consistent, difference in tone and the nature of representation between coverage of migrant women who migrated in earlier decades, versus those going through the same experiences now, enduring comparable hardships in the present day.

Geographical cleavages in migrant women’s inclusion are notable in coverage of domestic violence. Domestic violence towards migrant women worsened during the pandemic globally; this is reflected across the corpus in coverage in both minority (see: In Namen der Frau, Neuland Zeitung, 10th October 2021) and mainstream outlets (see: Refugees Refuse: One in four Abused, The Sun, 16th June 2020). Many articles include representations of solidarity towards these women, often through the inclusion of rich contextual detail. In The Sun (19th September 2021), context is established between violence and the disproportionate impact that government cuts had on migrant women during the pandemic: “migrant women are locked out of services and unable to access support altogether because they have no recourse to public funds” (emphasis added). However, in furtherance of the dynamic noted above, UK news coverage of domestic violence within the corpus primarily addressed international cases, rather than the rise in domestic violence against migrant women in the UK. In The Daily Telegraph article (22nd December 2020), journalists focus on one migrant woman in the US, described as “an undocumented migrant who only speaks Spanish”. This establishes a hierarchical distinction drawn between the domestic and international framing of these stories, and the legitimacy of their status (“undocumented”) with the stories of migrant women abroad being consistently foregrounded above domestic cases.

Speaking for migrant women

One key element determining the inclusiveness of representations is the use of direct quotes from migrant women in articles. As has been emphasized throughout the findings, many articles include emotive and descriptive quotes from women, whilst the provision of analysis comes through the voice of the journalist, or often, non-migrant spokespeople. This means that migrant women’s issues are often presented through an emotional frame, rather than allowing migrant women to have a say on the broader structural context or causes of their situation, or to speak on behalf of a collective; e.g. other migrant women, other workers, or citizens. This results in a hierarchy of voices between the personal (migrant) and professional (expert). In a Daily Telegraph article, the experiences of migrant women in the US at an Immigration and Customs Enforcement detention centre (15th September 2020) quotes a whistleblower, Dawn Wooten, and migrant women at the detention centre. However, the overarching analytical perspective is limited to the whistleblower and the journalist, whilst migrant women (described as “detainees”) are quoted in emotive terms, or paraphrased.

Another Daily Telegraph article (20th October 2020) addresses the increased precarity of migrant women working as domestic maids in oil-producing states around the Persian Gulf during the pandemic but confines migrant women’s voices to emotional accounts. One migrant woman’s voice is included, as she states, “I cry so much because my employers are always angry”. Contrastingly, the analytical perspective comes from the journalist and spokespeople cited in the piece. The exclusion of migrant women’s analytical perspective can also be seen in articles within minority news sources, for example, covering domestic violence (“In Namen der Frau”, Neuland Zeitung, 10th October 2021), and the challenges of the integration of female refugees during COVID-19 (“Scotland welcomes Afghan family after immigration battle defies the odds”, Asian Image, 22nd September 2021). In one Asian Image article, the husband of one migrant woman – Tabasum – is quoted extensively, whilst her representation remains in the third person. This compounds prevailing concerns around silencing marginalized voices and detracts from the potential of news coverage to provide a platform for promoting solidarity.

Lazy and pitying references

Whilst representations of solidarity typically focused on underlying points of connection to other groups and empathetic connections highlighted points of commonality, “pitying” representations identified in this analysis focused not only on stressing vulnerability but (intersectional) distance between the reader and the subject, through disempowering labels such as “women and children” or “especially women”.

This disempowering mode of representation emerged in offhand mentions in longer articles, for example, in one Süddeutsche Zeitung article, (29th June 2020): “Women, who are more likely to take on the now urgent childcare and are more often exposed to violence by partners and people with a migration background” are often affected (emphasis added). When patterns of intersectional exclusion are highlighted, such representations do little to meaningfully elevate the experiences and perspectives of this group of women. Similar representations are found in The Sun, where migrant women were represented solely through the description “forgotten migrant women”. Other examples can be seen in minority news outlets; in one Muslim News article (13th May 2020), the causes of gross disparities in health outcomes of COVID-19 between white and ethnic minority and especially migrant and refugee communities are speculated on. The journalist states “with men at higher risk, women and children may be left more isolated”. This also emerges in other coverage areas; reporting on the Calais-Dover crossing of the English Channel, The Sun uses the labels “vulnerable women and children” (19th December 2020). Such representations not only persistently collocate women and children (melding their concerns, and equating their levels of vulnerability), and overlook the compound challenges these groups are facing, but also disempower migrant women by suggesting that they are passive actors.

At the same time, representations that highlight differences and infantilize women to exclude them are also used as a justification to “help” or “save” them. Put simply, it is through victimizing migrant women, their inclusion is somehow deemed more legitimate in some outlets. The discursive construction of migrant women as a lesser group, often in parallel with children, permits their being helped. This logic of representation is predominantly found in the right-wing newspapers within the sample, where certain identity markers belonging to the majority group (being male, economically active, middle class, educated) are repurposed as grounds for exclusion and a lack of empathy or sympathy. For example, in an article from The Sun (25th March 2021) examining the UK’s rising immigration during the pandemic, migrant women’s vulnerable status is used to justify their humanity, as Priti Patel stated, “87 per cent of arrivals last year were men […] where are the vulnerable women and children that this system should exist to protect?”. Rather than channelling empathy, here, explicitly gendered identity markers associated with vulnerability are used, reinforcing gendered and socially embedded power imbalances.

Word embeddings: pattern confirmation

Given this study’s concern with the presence of solidarity and empathy in migrant women’s representations during the pandemic period, all articles from the mainstream news cluster, “The Management and Unequal Impacts of Covid-19” were analysed using word embeddings. In essence, word embeddings can highlight patterns, allowing us to tentatively confirm whether qualitative results are reflected more broadly.

Firstly, we compared how adjectives² most commonly associated with the group “migrants” compared with adjectives most closely associated with migrant women, or migrant men (see Table 5). As Table 5 shows, the adjectives more closely associated with “migrant women” are discursively differentiated from the adjectives associated with migrants, and migrant men. Namely, there are more emotive adjectives with disempowering connotations, with “vulnerable” being not only the top adjective associated with migrant women in the corpus but the top word overall. This connects to our qualitative analysis of pity in migrant women’s representations, whereby the word was used frequently to group women and children. Interestingly, whilst the adjectives closest to the gender-neutral “migrant” term do not mention age, “migrant women” are associated with the word “elderly”, an adjective with connotations of frailty, as opposed to the more neutral terms “old” and “young” discursively associated with the representations of migrant men.

Table 5. Top 10 adjectives most associated with migrants, migrant women and migrant men using word embeddings.

Migrant	Migrant women	Migrant men
Undocumented	Vulnerable	Old
Small	Undocumented	Young
Young	Unable	First
Alone	Elderly	Good
Vulnerable	Female	British
Poor	First	Black
Numb	Numb	Italian
British	Black	Foreign
Unable	Indian	Alone
Foreign	Foreign	Unable

Next, we analysed the nouns (referring to individual people or groups) most commonly associated with migrant women and migrants.

Table 6 shows the nouns most closely associated with migrants as a general group, in comparison to words which were more closely associated with “migrant women”. Whilst we see some overlap, there are clear differences. As the qualitative analysis touched upon, migrant women’s coverage during the pandemic did focus prominently on migrant women as workers, in articles related to job loss and increased precarity, environmental exposure and exploitation. We see “worker”, “staff”, “nurse” and “employee” within the list of top nouns associated with migrant women, whilst employment-related labels are less associated with “migrants”. Also, as a potential indication of solidarity within the articles on migrant women, the word “community” is much more associated with migrant women than it is with migrants as a neutral term. Furthermore, as evidence of the pitying frame of representation here we see that “victim” features within the top nouns associated with migrant women uniquely, in line with our qualitative findings. Overall, whilst it is impossible to conduct informed grounded theory on all articles in the selected cluster, word embeddings allow us to tentatively explore some of the qualitative findings, by illustrating the most discursively associated words across a large corpus.

Table 6. Top nouns (person labels) most associated with migrants and migrant women, using word embedding.

Migrant	Migrant women
Migrant	Girl
Refugee	Worker
Immigrant	Community
Worker	Migrant
Child	Woman
People	Child
Woman	Student
Family	Refugee
Man	Staff
Parent	Nurse
Group	Partner
Community	Group
Citizen	Immigrant
Us	Victim
Girl	Employee

Conclusion

The COVID-19 pandemic served as an opportunity to shift the politics of visibility in the news media and to promote solidarity and inclusion towards intersectionally minoritized groups. Yet, extant research has not engaged with the dynamics of news media coverage during this period, and explored how minoritized groups such as migrant women were represented (Walter and Glas 2024). This mixed-method study finds evidence of transformative and instrumental solidarity in representations of migrant women. Addressing our first research question regarding the articulations of these representational lenses, we find that they differ in scale (collective vs individual representation), their inclusion of context (broad political and socio-historical vs close, individual context), their issue focus (long-term, multi-issue vs short-term, single issue) and their platforming of voice (emotive or analytical). Somewhat unsurprisingly, such representations were generally more concentrated in left leaning and minority news, perhaps due to these outlets’ concerns with advocating for minoritized groups. Representations demonstrating solidarity towards different groups of migrant women generally were found in longer features; this suggests that such representations, which aim to promote inclusion and social action lend themselves to a particular style of journalism.

Differently to Varma (2020), who finds empathy and solidarity to be largely mutually exclusive logics of representation, this study finds representations to appear on a spectrum, often mixed within the same article. From analysing articulations of these different representational modes, we propose understanding empathy and solidarity as lenses temporarily employed by journalists to evoke specific reactions in the audience. The temporary and interchangeable employment of these different lenses of representation relates to conceptualizations of solidarity as “thin” and “thick”, resonating with Wright et al.’s (2022) notion of “drive by solidarity”. In addressing our first research question on articulations of solidarity and empathy, studying news representations of migrant women revealed a further representative logic: pity. Here, migrant women’s inclusion – predominantly found in right-leaning outlets – becomes palatable with increased (intersectional) distance to the subject. Temporal and geographical distinctions are employed to emphasize how different, desperate and exceptional these women’s cases are, thus how these migrant women are “victims”, worthy of help. These findings connect to de Carvalho, Santia, and Ramasubramanian (2024), who identified the framing of Indigenous people as victims within pandemic news coverage. Beyond news coverage, Liu and Wan (2024) also showed how policymakers in Taiwan employed a similar, gendered, victim frame in reference to migrant spouses, thus compounding patterns of exclusion.

In answering our second, methodological research question, combining qualitative and computational approaches enabled us to understand how intersectional identity markers function in different ways and on different scales across representational modes. Whilst the initial analysis revealed topical patterns in the coverage of migrants and migrant women, the qualitative analysis allowed us to explore precisely how representations were articulated. Stories of migrant women from EU and Western countries were largely absent across the corpus. Instead, news coverage of “migrant women” constructs a frequently racialized figure, associated with certain contexts (poverty, precariousness, the family, foreignness). Intersectional identity markers collocated with migrant women are often connected to discourses of class and socioeconomic status, poverty, and precarity. Ethnic backgrounds and nationalities of women emphasized in representations revealed who the media define as “migrant women”. Specifically, the backgrounds of migrant women foregrounded were those from Syria, from South-east Asia, Latin America and Central America.

In line with the intersectional focus of this special issue, the qualitative analysis revealed that the news media’s construction of the migrant woman is a constellation of minoritized identities, underpinned by notions of pity assigned to those who “deserve” help; in other words, through a form of intersectional stereotyping. Similar to visual analysis of representations of the “refugee crisis” (Maneri 2021), this study finds that the figure of the “migrant woman” emerges as an intersectional assemblage of identity markers and contexts – being a low-wage worker, from a poorer country, often being a person of colour, and a mother. This reproduces hegemonic power structures and racialized hierarchies. However, different constellations of identity markers lend themselves to different types of representation. Whilst gendered identity markers related to family roles typically emerge in empathetic representations, markers of socioeconomic status appear in representations expressing solidarity.

Overall, and despite limits in the extent of comparability, we locate patterns emerging regarding patterns in migrant women’s representation across different outlets in Germany and the UK. Topic modelling guided us towards contexts in which migrant women are visible, word embeddings highlighted key discursive dimensions of representation, and qualitative analysis allowed us to deepen our understanding of the representational modes of solidarity, empathy and pity. As such, from a methodological perspective, this study shows how computational and qualitative methods can be combined in order to study complex, nuanced topics, such as representation. Focusing on news media coverage from the pandemic period has shown that during crisis periods, whilst the news media can promote solidarity towards groups, the presence of a crisis does not guarantee that inclusive representations would be used. Indeed, despite the crisis context of the pandemic and the fact that migrant women are among the worst impacted groups, findings show that even in “positive” news coverage, a range of representational lenses are employed that can reinforce patterns of exclusion in subtle ways. In considering the possible role the news media can play more broadly in supporting inclusion or enacting solidarity, this study complicates and enriches our understanding of different representational lenses, and how they are employed in relation to minoritized groups, working towards an identification of problematic patterns in exclusionary representation.

Our research has not been without challenges; whilst we included a broad corpus including outlets from the UK and Germany, the topic modelling results led us to a set of particularly relevant articles, which were potentially skewed towards the English language. This article was not intended to include a systematic, cross-national, cross-outlet comparison; however, the authors did not anticipate the extent to which the quantitative results would focus on the UK section of the sample. We used machine translation to cross the language barrier, however automated translations are imperfect and mistranslations can impact further analysis results (Lind et al. 2022). On one hand, topic modelling as an approach to reach smaller, more concentrated qualitative samples for further analysis works very effectively, facilitating both a comprehensive overview at the macro level and the ability to identify a small, highly relevant sample. On the other hand, state-of-the-art multilingual text analysis methods still imply a tradeoff between scalability and human-level accuracy (Baden et al. 2022).

Ultimately, in studying the potential for social change embedded in the news representations, this study enriches discussions of journalism's possible role in promoting inclusion and solidarity in society. Representations based upon both solidarity and empathy challenge traditional objective, monitorial journalistic norms (Broersma 2017). Unpacking the intersectional dimensions of migrant women’s representations helps us identify who is visible (and who is not), who is defined as worthy of empathy, and in which contexts (in which moments and places/situations is solidarity expressed with certain migrants). Combining computational and qualitative methods allowed us to use the intersectional solidarity framework on a larger but more focused set of articles, which dominant computational approaches alone (e.g. keyword searches) simply could not reach. Looking to the future, scholars could examine the way in which different intersectional identity markers are employed in different national contexts; how the legitimacy criteria for inclusion vary by outlet, and by country, and how we might explain these different “conditions” for inclusion. Future studies could explore the potentially varied impacts of these different modes of representation on audiences, and begin to include more diverse news formats.

Disclosure statement

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

Additional information

Funding

This work was funded by Emmy Noether Programme from the DFG [grant number WA 4161/1-1].

Notes

1 This was calculated using the mean Gamma (γ) score. Gamma, in LDA topic modeling, represents the measure of how relevant a document is to a topic. The mean gamma score for “management and unequal impacts of COVID-19” was used as a threshold when sampling articles.

2 We excluded adjectives making reference to time, speed, order and volume.

Appendices

Appendix 1. Keyword searches

A multilingual keyword search was created and matched across German and English. Differences in grammar between German and English created an increased challenge for gendered search terms because, unlike English, German uses gendered nouns. This resulted in search strings that were structurally different, but semantically equivalent.

Migrant

EN = Migrant.*, Foreign.*, Immigrant.*, Student.

DE = Migrantin.*, Einwanderin.*, Zuwanderin.*, Ausländerin.*, Gastarbeiterin.*, Immigrantin.* OR Migrations.*, eingewandert.*, einwandert.*, zuwandern.*, zugewandert.*, zuwandert.*, ausländisch.*, “aus dem Ausland”, “immigriert.*”

Women

EN = Women.*, woman.*, girl.*, female.*, Mother.*, Sister.*, Daughter.*

DE = Frau.*, Mädchen, Mutter, Mütter, Schwester.*, Tochter, Töchter, Bürgerin.*, Staatsbürgerin.*, Studentin.*, Schülerin.*

Appendix 2. Pre-processing topic models

Before translation, minimal pre-processing steps included cleaning up duplicates, splitting documents into individual words (tokenization), and removing punctuation. To reduce noise for topic modelling, we followed Maier et al. (2018) suggestions of lowercasing, lemmatization, stop-word removal, and pruning of outlier occurrences (<5 and >99 per cent), using the default quanteda dictionary and lexicon R packages (Rinker 2018).

LDA topic models use statistical co-occurrence of words to group words and documents into a pre-selected number of topics. We followed the methodology laid out by Maier et al. (2018). We trained and optimized the topic model on our mainstream news sample using the LDA model in topicmodels (Grün and Hornik 2011) and ldatuning (Murzintcev and Chaney 2020) R packages. We found the best number of topics (k) by comparing LDA scoring algorithms from ldatuning across models that varied in their number of topics (10 to 100, by 5). The researchers then compared the interpretability of each model; sixty topics was the best performing model.

Appendix 3. Provisional coding scheme

Scale:

– is this story about one woman/multiple women (personal/collective)?

3rd person labels

“women and children”

“especially women”

By occupation

Through their relationship to men: as wives and girlfriends

Family labels: Daughter/Mother/Sister/Wife

Type of citizen/spokespeople

• “ordinary” citizens

• Activists

• Elite source

• Politicians

• NGO

• Businesswomen

Use of direct voice/reported speech

• Emotive

• Descriptive

• Analytical perspective

Context

• Personal (one person profile, personal context)

• local (discusses local context, similar women)

• Structural/broad (discusses story in context of broader events/issues/structure)

• Historical (connects to similar events in other periods)

• Horizontal/connection to other minoritized groups

Appendix 4

Table A1. Articles cited in text.

Outlet	Date	Title	Summary
The Guardian	14th December 2020	Australia wants international students to study here but abandoned them during the Covid crisis	Reflections and call for action from Colombian female migrant and international student – in reaction to lack of support during the COVID-19 pandemic
Muslim News	13th May 2020	How Govt failed Black, Pakistani and Bangladeshi communities during Covid-19 pandemic	Description and analysis of the unequal impacts of COVID-19 on the population of the UK
Die Bild	4th February 2021	Verzweifelte Eltern reden über den Lockdown; Tränen bei Merkels Bürger-Talk	Reflects on the points raised during Angela Merkel’s annual discussion with citizens – one migrant women’s question is highlighted in the article
The Guardian	1st September 2021	Spanish hotel booking app to show working conditions of staff	Reports on Union-like organization and app to monitor and report potential mistreatment of hotel workers in Spain – in a sector dominated by migrant women
The Guardian	12th May 2020	He lived in the US for forty years. Then he became the first to die from Covid-19 in immigration jail	A man who had lived and worked in the US for over four decades died of COVID in immigration custody. The Guardian report focuses on his sister’s account of the event
The Guardian	1st July 2020	You have to take Action: One hospital cleaner’s journey through the pandemic	The profile of Ernesta, a hospital cleaner who played an integral role in the union and protecting the rights of workers
The Daily Telegraph	31st July 2020	Paulette Wilson	Obituary for a female migrant (from Windrush generation) who campaigned a lot for the rights of the Windrush generation
Sueddeutsche Zeitung	8th September 2020	Eine Frau für das Miteinander; Nükhet Kivran, ehemalige Vorsitzende des Ausländerbeirats, ist im Alter von 56 Jahren gestorben	One woman with a migration background, who was on the migration council in Munich, passed away
Neuland Zeitung	10th October 2021	In Nahmen der Frau	A piece on domestic violence worsening during the pandemic.
The Sun	16th June 2020	Refugees Refuse: One in four Abused	Report on the fact that one in four domestic abuse survivors are rejected from refuges because they have no housing benefit access
The Sun	19th September 2021	Credit Crunch: Universal Credit £20 cut could push domestic abuse victims into poverty, warns charity	Describes rising domestic abuse in pandemic – and the potential increased vulnerability, especially of migrant women – due to government cuts to welfare
The Daily Telegraph	22nd December 2020	Inside New York’s domestic violence shadow pandemic	Reports on rising domestic violence during the pandemic and its potential causes, highlighting the increased vulnerability of migrant women and drawing on several individual’s experiences and accounts
The Daily Telegraph	20th October 2020	“I want to leave but I’m scared”: Pandemic increases risk of violence for Gulf’s Domestic workers	Highlights how the precariousness of many domestic maids situations (working in Gulf state) has become much worse during the pandemic. Reports on working conditions, physical abuse, and human rights abuse
The Daily Telegraph	15th September 2020	Whistleblower nurse alleges “hysterectomies performed on immigrant women” in US	Detained immigrant women under ICE custody undergoing different medical procedures without consent. Reports on whistleblowers account
Neuland Zeitung	Winter/Spring 2021	Article One: Machmal fühle ich mich immer noch hilflos	Discusses a range of challenges on the experience of lockdown in Germany as a migrant
Neuland Zeitung	Winter/Spring 2021	Article Two: Lockdown: eine Zeit mit zwei Zeiten	Discusses the experience of lockdown as an immigrant with the focus on some positive aspects
Asian Image	22nd September 2021	Scotland welcomes Afghan family after immigration battle defies the odds	An Afghan-Scottish man is reunited with wife, children and family after Taliban takeover
Sueddeutsche Zeitung	29th June 2020	CORONA; Vor allem ein großes Unglück	The impact of COVID-19 pandemic on families. Includes focus on several families, the issues they are working through, and the work of Caritas Erding
The Sun	19th December 2020	CUT-PRICE CROSSINGS Small boat smugglers “slash price for migrants” to cross the Channel ahead of festive surge	Highlights the work of people smugglers, who are charging increasing amounts to bring families from Calais to Dover
The Sun	25th March 2021	Be Tough to be Kind on Asylum	Argues that healthy, economically active working men should not be coming to the UK and claiming asylum, but rather the system should be only aiding women and children
The Guardian	4th June 2021	“Outrageous”: why was a US health worker charged with spreading Covid?	Description of the case/attempt to hold a female migrant care worker criminally liable for the spread of Covid – in causing the death of her client. Josefina Brito-Fernandez lost her license to work and faced the possibility of deportation