Threats, victims, or heroes? Media frames about migration in the United Kingdom and Brazil Isabella Gonçalves Department of Communication, Johannes Gutenberg University Mainz, Mainz, Germany Yossi David Department of Communication Studies, Ben-Gurion University of the Negev, Beer Sheva, Israel Abstract Conservative media often frames migrants negatively, echoing discourses of othering driven by populist leaders. Previous studies have examined the intersections between media framing, migration, and populism, but comparative studies on migration framing including Brazil remain scarce. This study uses quantitative content analysis to explore media frames of news items published in the United Kingdom and Brazil. It aims to contribute new insights regarding migration media framing by comparing conservative news coverage of two countries that experienced a rise in populism wave in the last years. We found that new items in both countries tended to frame migration in similar ways, suggesting a pattern in two different contexts. Findings show the prevalence of negative framing over positive framing, with signi fi cant differences for victim frames, and non-signi fi cant differences for hero and threat frames. This study contributes by providing new insights into the intersections between media framing, migration, and populism. Keywords Migration, media, news, framing, populism, content analysis, Brazil, United Kingdom Corresponding author: Isabella Gonçalves, Jakob-Welder-Weg 12, 55128 Mainz, Germany. Email: idesousa@uni-mainz.de Original Manuscript the International Communication Gazette 1 – 21 © The Author(s) 2024 Article reuse guidelines: sagepub.com/journals-permissions DOI: 10.1177/17480485241249007 journals.sagepub.com/home/gaz Migration is a global social phenomenon that has existed throughout human history and still changes nations and cultures (Bauman, 2011; Castles et al., 2014; Urry, 2012). Despite the long history of human movement and migration worldwide, public, political, and media discourses frame migration as a negative challenge to society that needs to be regulated (Bakewell, 2008; Milioni et al., 2015). The idea of regulating migration is often associated with media frames of migrants as threatening the local communities and as bringing risk to various social institutions, e.g., the economy, security, culture, and health systems of the receiving country (Froio and Ganesh, 2019; Lawlor, 2015; Morin, 2016; Pruitt, 2019). By focusing on the threats posed by migrants to the host society, news media often assembles populist rhetoric (Balch and Balabanova 2017; Bosilkov and Drakaki 2018; Dolea, Ingenhoff and Beju 2021). For example, several studies show that media tends to focus on the economic problems caused by migration, and similar discourses were used during the Brexit campaign (Balch and Balabanova 2017; Okoye 2020). Moreover, both populist movements and media frames promote ‘ us versus them ’ dis- courses, creating shared enemies (Araújo and Prior, 2021; Hameleers et al., 2017; Wettstein et al., 2019) and the otherization of outgroups (Matthes and Schmuck, 2017; Shehaj et al., 2021). Considering the similarities in migration frames between populist political communi- cation and news media, previous studies have been attempting to uncover the intersec- tions between populism rhetoric and media frames on the migration context (Bos et al., 2011; Diehl, Vonbun-Feldbauer and Barnidge 2021, Mazzoleni, 2008). This study aims to further contribute to studies uncovering such intersections, and it examines migration media frames in two countries that experienced a rise in populism in the last decade. We focus on UK and Brazil for that intent because both countries share similar- ities in the experience of populism. However, they also have differences in how migration issues were framed in political campaigns. In this regard, this paper aims to uncover the following research question: To what extent do news items in Brazil and the UK capture and portray the contextual variations in populist discourse surrounding migration? To answer this question, this paper is grounded on the framework of migration media frames and its intersections with populist rhetoric. This study uses quantitative content analysis to examine news items published between 2016 and 2018, during which both the Brexit referendum and the election of Bolsonaro took place. It aims to provide a further understanding of the way media frames migration in different contexts, contrib- uting to the literature examining the intersections between media, migration, and populism. Theoretical background Migration media frames Framing refers to the ways in which an author, e.g., journalist, assembles a narrative by using a particular interpretation (Entman, 1993, 2007; Kuypers, 2009; Van Gorp, 2007). To highlight a certain perspective of an event, journalists introduce and raise 2 the International Communication Gazette salience to speci fi c ideas, increasing the chances that the audiences will evaluate an event in a particular way (Entman, 2007). Establishing a dominant frame depends on the interactions between politicians, societal elites, media, and the general public. In a top-down network, the political administration holds more power than the other actors, while the public holds the least (Entman, 2003). However, establish- ing a particular frame is sometimes con fl icted, and different frames can compete with one another (Guggenheim et al., 2015; Knüpfer, 2018). In other cases, the public also plays a role in determining the frames, although they have less power than other actors (Entman, 2003). Media usually frames migration through three main frames, and migrants are either represented as threats, victims, or heroes (Benson, 2013; Cogo and Silva, 2015; Gonçalves and David, 2022; Greussing and Boomgaarden, 2017). Threat frames mainly focus on economic, security, health, and symbolic issues. For example, migrants are framed as taking over jobs, abusing social services, increasing the crime rate, spread- ing diseases, and threatening moral values (Benson, 2013). The victim frame usually represents migrants as a helpless mass of victims in need of help (Bosilkov and Drakaki, 2018; Greenwood and Thomson, 2020; Mizga and Trovão, 2018). Another frame, albeit less frequently used, is the hero frame, which involves positive stories about migrants that feature integration, cultural enrichment, and overcoming of a complex background (Ryan and Reicher, 2019; Zanforlin and Cogo, 2019). An example of the hero frame is stories focusing on migrants who became entrepreneurs in their new country of residency or successfully integrated into the local culture, such as by learning the host country language. However, positive frames are less frequent, and, overall, conservative media in various countries adopt mainly a negative tone to represent migration, contributing to their social exclusion (Quinsaat, 2014; Gonçalves, 2024; Sarah Liu, 2019). Media and populism A growing body of research shows congruences between populist rhetoric and conserva- tive media framing (Bos et al., 2011; Dolea et al., 2021; Hameleers et al., 2019; Mazzoleni, 2008). News framing shares similarities with the populist style, such as high personalization, dramatization, and anti-establishment attitude (Araújo and Prior, 2021). Additionally, conservative news media are more likely to adopt a populist style, and populist supporters with anti-establishment attitudes are more likely to inform them- selves through them (de Rooij et al., 2022). Previous studies have shown that populist leaders received increased media attention, contributing to the rise of populist movements worldwide (Krämer, 2014; Mazzoleni, 2008). For example, fi ndings show signi fi cant effects of media framing on the public image of populist leaders, and increased media attention contributes to building their image of being strongly in fl uential and legitimate (Bos et al., 2011). Besides raising high attention to populist leaders, journalists also actively construct news framing with similar moral oppositions as the populist style, such as opposing the people with the elite, which creates us vs. them divisions (Hameleers et al., 2019). Gonçalves and David 3 Framing may also contribute to agenda-setting by emphasizing speci fi c topics (Chyi and McCombs, 2004; Entman, 2007). For example, media also promote the rise of popu- lism movements because of the issues they cover (Kieslich and Marcinkowski, 2020). News media are particularly interested in topics with higher emotional appeal, such as crime and migration (Burscher et al., 2015; Damstra et al., 2021; Tuttle and Harris, 2019). By focusing on migration, public opinion concerns over the issue are more likely to grow (Gottlob and Boomgaarden, 2022; McLaren et al., 2017). With increased concern over migration, parties with anti-migration agendas are more likely to receive support because they focus on the same issues highlighted in the media and are perceived as more competent for solving the issue (Boomgaarden and Vliegenthart, 2007; Diehl et al., 2021; Kieslich and Marcinkowski, 2020). To this extent, populist rhetoric also shares similarities to migration media frames, such as focusing on threats posed by migrants (Cap, 2019; Dammert and Erlandsen, 2017; Dolea et al., 2021). Populism and migration in Brazil and the UK The evolution towards populism and the increase on popularity of populist fi gures often emerges from societal fears and discontent surrounding issues of migration and identity (Béland, 2020; Hogan & Haltinner, 2015). This process usually evolves in stages, begin- ning with increased public anxieties over economic insecurity, job loss, or demographic changes, which are then ampli fi ed by both media outlets and political fi gures (Benesch et al., 2019; D ’ Ancona and Angeles, 2015). In this climate, populist leaders gain attention by channeling these anxieties, such as by focusing on the dangerous others supposedly threatening the host society (Matthes and Schmuck, 2017; Shehaj et al., 2021). The aggressive rhetoric and the controversial claims often gain media attention, further increasing societal polarization and rising popularity of fi gures creating such divisions (Krämer, 2014; Mazzoleni, 2008). Considering the evolving dynamics of populist rhetoric and the growing support of populist leaders in context of perceived crisis, the present study delves into two distin- guish cases in which migration, populism and media were interconnected. Particularly, we examine the cases of Brazil and the UK. We focus on those countries because during the timeframe between 2016 and 2018, there were important contextual similar- ities in both countries. For instance, the incoming migration movements from Venezuela to Latin America and from Syria to Europe increased the salience of migration media coverage, setting the public agenda for the issue (Hobolt, 2016; Okoye, 2020; Ranincheski and Uebel, 2018). Venezuelan migration movements towards other Latin American countries gained prominence particularly in 2017 and 2018 years, when oil prices reduction, rising authori- tarian measures and society ’ s polarization led to economic and political crisis in the country (Pantoulas and McCoy, 2019; McCoy et al., 2018). In the other side of the ocean, Europe also witnessed incoming migration movements from Syria. The increased migration was primarily driven by the civil war in the country, and the years of 2015 and 2016 fi gured with the highest quantitative of migrants seeking safer conditions abroad, many of them in Europe (Eurostat, 2017). The arrival of Syrian refugees has signi fi cantly 4 the International Communication Gazette shaken the European Union ́ s diplomatic relations, as member states faced the challenge of managing internal political pressures against migration (Blanchard et al., 2020; Karacuka, 2021). In this sense, the period between 2016 and 2018 was characterized by intense debate over migration themes in both countries. The increased numbers of migrants were aligned with the increasing media coverage of migration issues in both regions as well as with an increase negative attitudes and public concerns towards migration issues (Blinder and Jeannet, 2018; Giacomozzi et al., 2023; McLaren et al., 2017). Previous results revealed that media coverage of migration issues in the UK tends to frame it from a Eurosceptical perspective and focuses mainly on migration in and to European countries (Evans and Mellon, 2019; Hobolt, 2016). In the meantime, similar discourses against the European Union and the threats caused by free movements, such as a rise in migration numbers, were used by populist discourses before the Brexit referendum (Cap, 2019; Schweyher et al., 2019; Walter, 2019). The topic of migration was an important intersection between popu- list movements in different contexts, and migrants were usually framed as dangerous others to the people, in discourses of us vs. them (Alamillo et al., 2019; Dammert and Erlandsen, 2017; Merry, 2022). Moments of unstable democracy and authoritarianism were associated with increased othering of migrants in Brazil, such as during the populist dictatorship of Vargas (1930 – 1945) and the recent government of Bolsonaro (Filomeno and Vicino, 2021). Likewise in other countries with increased populism, the election of Jair Bolsonaro was preceded by intense political crisis. Particularly, the Operation Lava Jato (Car Wash) contributed to rising political instability while multiple political fi gures were prosecuted for corruption, culminating in increased presidential dissatisfaction and the impeachment of President Dilma Rousseff in 2016 (see Damgaard, 2018; Lagunes & Svejnar, 2020). Brazilian presidency was later conducted by Michel Temer, who mainly focused on unpopular reforms during his mandate (Milhorance, 2022). In the meanwhile, Jair Bolsonaro gained political and media attention while blaming political others for the political and economic crisis Brazil was facing (Araújo and Prior, 2021; Lugo-Ocando, 2020). Brazil shares similarities with the UK in how populist politicians frame migration, but there are also important differences. While Bolsonaro targeted migrants to some extent (Filomeno and Vicino, 2021), his discourse of othering focused mostly on corruption and communism (Casarões and Barros Leal Farias, 2022; Novoselova, 2020). He accused various elite groups of posing a symbolic and economic threat to the country, such as establishing a communist regime or a gender ideology, degrading moral values, or committing corruption crimes (Araújo and Prior, 2021; Lugo-Ocando, 2020). Some of his favorite themes were the risk of Brazil becoming Venezuela, while Venezuelans were described as victims of Maduro ’ s dictatorship. This focus on Venezuela and Maduro paradoxically led Brazil, under a right-wing government, to adopt a progressive refugee policy and to grant a 2-year residence permit to refugees from Venezuela (Brumat and Geddes, 2023). To some extent, this policy was linked to the will of the Brazilian government to establish a closer partnership with the United States of America, which in the long run had in its best interest the weakening of Venezuela ’ s government (Brumat and Geddes, 2023). Given the congruence between Gonçalves and David 5 populist rhetoric and conservative news framing (Hameleers et al., 2019; Kieslich and Marcinkowski, 2020), and that in Brazil, populism rhetoric was less likely to focus on the threat frame compared to the UK, we hypothesize as follows: Hypothesis 1 (H1): News items published in conservative newspapers in the UK use more threat frames in comparison to conservative newspapers in Brazil. Hypothesis 2 (H2): News items published in conservative newspapers in the UK use fewer victim frames in comparison to conservative newspapers in Brazil. Hypothesis 3 (H3): News items published in conservative newspapers in the UK use fewer hero frames in comparison to conservative newspapers in Brazil. Method Sampling The study data were collected from the LexisNexis Academic database for all the news items covering migrants that were published in two major conservative newspapers in the UK ( The Sun and The Times ) and two in Brazil ( Estado de S. Paulo and O Globo ) between January 1, 2016, and December 31, 2018. In LexisNexis, the only Brazilian newspapers available were Estado de S. Paulo and O Globo – conservative newspapers with high circulation rate in Brazil. To maintain parity, we also selected conservative newspapers from the UK. In this sense, the selection criteria of the British newspapers were based on their political leaning, daily circulation rate and their availability through LexisNexis Database. Among the British newspapers available, The Times and The Sun were the conservative newspapers with the highest circulation rate. The analysis focuses on the period 2016 until 2018 to encompass two major political events in Brazil and the UK: the Brexit Referendum in June 2016 and the election of Bolsonaro in October 2018. Moreover, this timeframe also contemplates the increased migration movements to both countries. We searched for the keywords ‘ immigrant ’ or ‘ migrant ’ or ‘ refugee ’ or ‘ migratory crisis ’ or ‘ migration ’ in English and Portuguese to identify relevant news items. The search led to a total of 28,911 news items. More speci fi cally, we found the following number of news items for each newspaper in 2016: Estado de S. Paulo (n = 597), O Globo (n = 403), The Times (n = 5579), and The Sun (n = 3703); in 2017: Estado de S. Paulo (n = 634), O Globo (n = 216), The Times (n = 5392), and The Sun (n = 2093); and in 2018: Estado de S. Paulo (n = 382), O Globo (n = 197), The Times (n = 7402), The Sun (n = 2313). To control for the large differences in the number of news items between Brazilian and British newspapers, we randomly selected a sample of the media coverage of 720 news items regarding migrants in each country — a total of 1440 items in both countries. We opted for balanced sample sizes across news outlets to have a similar power in the stat- istical analysis, in line of other studies using content analysis for comparing media cover- age in different countries (see Gonçalves, 2024; Lawlor, 2015; Thorbjørnsrud and Figenschou, 2016). We fi rst conducted a Chi-square Test and then a Binary Logistic 6 the International Communication Gazette Regression. Chi-square Test is an ideal approach to analyze categorical variables with cell frequencies (McHugh, 2013). Following the Chi-square test, we used Binary logistic regression in order to control for additional central factors (such as regional focus, news article source, and length on the news item). Binary logistic regression is the stand- ard regression method when the dependent variable is categorical and dichotomous (Maroof, 2012). For the sampling collection, we aimed at selecting a random sample covering this three-years ’ time frame, to be able to quantify the use of such frames in the British and Brazilian news items. Thus, we did not select speci fi c events which could lead to a non-representative sample of news items. Hence, the random sample gave us a repre- sentative sample of these four newspapers in both countries. The sample included 120 news items each year ( ∗ 3 years) in each newspaper ( ∗ 2 newspapers) in each country ( ∗ 2 countries), a sum of a total of 1440 news items. From the sample we deleted duplicate items that the automatic exclusion performed by LexisNexis did not capture, news items that did not focus on migration, and news items from other newspapers, resulting in a total sample of 1312 news items — 653 in the UK and 659 in Brazil. We opted not to include further news items because the sample distribution was similar, which would not under- mine the data analysis. In this sense, each outlet had a sample size of at least 325 news articles and no more than 334 news items. Thus, the sample is balanced enough among years and outlets (see Table 1 for more details on our sample). Coding procedure We conducted a quantitative content analysis of news frames (see Aalberg & Beyer, 2015; Mancini et al., 2021; Sarah Liu, 2019; Thorbjørnsrud, 2015) across the sample of 1312 news items. The unit of analysis was news item. The coding scheme included identi fi cation questions, such as the country of the newspaper, the name of the newspaper, the year of publication, and the length of the news item (for more details, see the Online Appendix). Media frames To measure the presence of the main frame, we coded three main frames as (a) threat, (b) victim, and (c) hero. Each was coded as 1 (used in the news item) and 0 (not used in the Table 1. Number of sampled news items by country, newspaper, and year. Country Newspaper 2016 2017 2018 Total Brazil Estado de S. Paulo 115 104 106 325 O Globo 115 110 109 334 UK The Times 114 105 102 321 The Sun 119 108 105 332 Total 463 427 422 1312 Gonçalves and David 7 news item). In addition, following Benson ’ s (2013) classi fi cation, each of those frames included a list of corresponding subframes, which will be detailed below. Threat subframes included the economic, security, and symbolic (each coded as 1 [used in the news item] or 0 [not used in the news item]). Economic threat focuses on fi nancial dif fi culties caused by migrants in government service programs and competition over jobs. The security threat subframe emphasizes increased crime rates, associating migrants with breaking the law. Finally, news items focusing on symbolic threat approach the dif fi culty of integration because of different values and belief systems and their threat to national cohesion. The victim subframes included inequalities in the economy, humanitarian, and racism/ xenophobia (each coded as 1 [used in the news item] or 0 [not used in the news item]). Inequalities in economy items approach migration as an outcome of a liberal economy and unjust distribution of resources. The humanitarian subframe highlights migrants ’ fra- gilities and their suffering from violations of human rights, business practices, and legal processes. Stories related to racism and xenophobia represent migrants as victims of dis- crimination and hate crimes. Hero subframes included cultural diversity, integration, and good worker (each coded as 1 [used in the news item] or 0 [not used in the news item]). Stories focused on cultural diversity show the bene fi ts of migrants in enriching the host country ’ s culture. Subsequently, the integration subframe focuses on success stories of migrants ’ integra- tion. Finally, the good worker news items center on hard-working migrants, either by cre- ating new opportunities through entrepreneurship or by taking jobs refused by citizens. Control variables To statistically control for additional factors, we added as control variables the year of publication, the country of the newspaper, and the different newspapers. The year of pub- lication was coded as: 2016 (1); 2017 (2); 2018 (3). The country of the newspaper was coded as the UK (0) and Brazil (1). Each of the four newspapers was coded as dummy variables. Estado de S. Paulo, O Globo, The Times , and The Sun were coded as 1 (news- paper analyzed), and other newspapers were coded as 0. In the regression models, The Sun was used as a reference point because it is the only tabloid in the sample. The other three newspapers were measured relative to The Sun . To control the impact of prox- imity of the news event, we included the variable Region , where news items covering international events were coded as 1, and news items covering national events were coded as 0. Moreover, we controlled if the news item was covered by the news outlet or if it reproduced outside sources or wired content. In this sense, Source was coded as 0 if it was original content and as 1 if it reproduced external content. This content analysis allowed us to identify more than one frame per news item because our goal was to identify all the frames in each news item. Our content analysis inter-coder reliability test was performed by two Brazilian graduate students of 120 news items. First, they were trained based on articles not included in the sample. After the train- ing, the reliability coding was performed using 120 news items that were part of the sample, and they were incorporated in the fi nal dataset. After the reliability test, Coder 8 the International Communication Gazette 1 coded the remaining sample. We have followed recommendations in the literature that indicate that coding more than 50 items is adequate for reliable samples (see for example, Haßler at al., 2023; Lacy and Riffe, 1996; Lombard et al., 2006; Neuendorf, 2010). After coder training, we conducted the intercoder reliability for the framing variables in the coding scheme. Intercoder reliability assessment yielded strong agreement scores, result- ing in a Krippendorff ’ s Alpha score 0.81 or higher (for the Krippendorff ’ s Alpha for each variable, see the online Appendix). Results Descriptive statistics Table 2 presents the main issues, the region of the event (international or national) and the source of the outlet (produced by the outlet or external news agency) in the analyzed British and Brazilian news items. News items covering migration as the main issue repre- sented 58.4% of the British distribution, compared to 52.6% from the Brazilian distribu- tion. The second most prevalent theme was intragovernmental relationships, capturing 14.9% of the thematic space in the UK and 19.4% in Brazil, indicating a substantial inter- est in governmental internal affairs. Foreign relations constitute 11.6% and 13.5% of the thematic distribution in the UK and Brazil, respectively, suggesting a moderate level of importance attributed to this sector. The coverage on security and law occupies a smaller segment, encompassing 6.3% of British and 6.9% of Brazilian journalistic content, while economic issues feature even less prominently, constituting 3.5% and 3.7% of news items in the respective countries. Human rights issues appear to be less prioritized, with representation of only 2.3% and 2.0% in British and Brazilian news. Other political topics occupy marginal space, accounting for 3.1% of the content in the UK and 1.9% Table 2. Main issues, regional focus, and source percentages. UK Brazil Issues Migration 58.4% (354) 52.6% (312) Foreign relations 11.6% (70) 13.5% (80) Security and Law 6.3% (38) 6.9% (41) Economic Issues 3.5% (21) 3.7% (22) Human Rights 2.3% (14) 2.0% (12) Intragovermental Relationship 14.9% (90) 19.40% (115) Other political topic 3.1% (19) 1.9% (11) Region International 17.7% (118) 64.6% (416) National 82.3% (550) 35.4% (228) Source Original Content 100% (668) 98.1% (632) External Source 0.0% (0) 1.9% (12) Gonçalves and David 9 in Brazil. Regarding regional focus, the data reveal a predominant national focus in British media at 82.3%, contrasting with the Brazilian media ’ s inclination towards inter- national coverage at 64.6%. In terms of content originality, the fi ndings indicate a near- universal prevalence of original content produced by outlets in both countries, evidenced by a 100% originality rate in the UK and 98.1% in Brazil. Table 3 presents the different uses of the main frames in the British and Brazilian news items. The frequency of the negative frame of threat was 28.9% in the UK and 23.3% in Brazil. The Chi-square test yielded signi fi cant differences between the UK and Brazil when considering the frequency of the threat frames, (df = 1) = 5,025, p < .05 (see Table 3). Among the threat subframes (see Table 4), the frequency of the security threat subframe was high in both countries (16.8% in the UK and 16.5% in Brazil), fol- lowed by the economic threat subframe (11.1% and 7.1% respectively), and the symbolic threat subframe (4.0% and 3.9%, respectively). The Chi-square test yielded a signi fi cant difference between the UK and Brazil when considering the frequency of economic threat (df = 1) = 6.110, p < .05. However, no signi fi cant differences were found for the security, and symbolic subframes (see Table 4). To summarize, the frequency of the frames and subframes in the news items shows mixed results regarding the differences in migration media frames in both countries, as well as more negative frames and fewer positive frames. The frequency of the negative frame of victim was higher in both countries, 28.6% in the UK and 30.6% in Brazil (see Table 3). No signi fi cant differences were found in the victim frame between the countries. Among the victim subframes (see Table 4), the fre- quency of the use of the humanitarian subframe was higher in both countries (22.6% in the UK and 23.9% in Brazil), followed by the racism subframe (6.1% and 6.7%, respect- ively) and the economy inequality subframe (1.5% and 5.1%, respectively). The Chi-square test yielded signi fi cant differences between the UK and Brazil when consid- ering the frequency of inequalities in the economy (df = 1) = 13.609, p < .01. However, no signi fi cant differences were found for humanitarian and racism subframes (see Table 4). Finally, the frequency of the positive frame, measured as the hero frame, was low in both countries — only 13.8% in the UK and 12.5% in Brazil (see Table 3). No sig- ni fi cant differences were found for the hero frame. However, among the hero sub- frames (see Table 4), the frequency of the use of the good worker subframe was higher in the UK (10.6%) than in Brazil (6.1%). The Chi-square test yielded signi fi cant differences between the UK and Brazil when considering the frequency of good workers Table 3. Threat, victim, and hero frames in the UK and Brazil. Frame UK Brazil Chi-Square Threat 28.9% (193) 23.4% (151) 5.025 ∗ Victim 28.7% (192) 30.6% (197) .537 Hero 13.8% (92) 12.1% (78) .802 Note. ∗ p < 0.05. 10 the International Communication Gazette (df = 1) = 8.926, p < .01. In addition, the frequency of the cultural diversity subframe was higher in Brazil (6.5%) in comparison to the UK (2.8%). The Chi-square test yielded sig- ni fi cant differences between the UK and Brazil when considering the frequency of cultural diversity (df = 1) = 10.002, p < .01. Lastly, the frequency of the integration subframe was 2.8% in the UK and 3.3% in Brazil, yet no signi fi cant differences were found for the inte- gration subframe (see Table 4). Hypothesis testing The fi rst hypothesis stated that news items published in conservative newspapers in the UK use more threat frames in comparison to conservative newspapers in Brazil. We con- ducted a binary logistic regression to estimate the likelihood of Brazilian newspapers using the threat frame. Table 5 demonstrates that the country of the newspaper is not sig- ni fi cant in estimating the likelihood of news item from the UK using the threat frame. Classi fi cation accuracy was 73.7% (see Table 5, Model 1). We controlled for each of the newspapers, year of publication, length of the item, region, and source. The fi ndings show that relative to The Sun , The Times is less likely to use threat frames, yet no signi fi - cant differences were found for O Globo and Estado de S. Paulo . The regression model also shows signi fi cant differences in the year of publication and that newspapers use fewer threat frames over time. The fi ndings suggest that the country alone has a non- signi fi cant association with the likelihood of the newspapers in a given country to use a threat frame. To test our second hypothesis (H2) that news items published in conservative news- papers in the UK use fewer victim frames in comparison to conservative newspapers in Brazil, we conducted a binary logistic regression to estimate the likelihood of Brazilian newspapers using the victim frame . Model 2 demonstrates that the country of the Table 4. Subframes in the UK and Brazil; percentage (n). Subframes UK Brazil Chi-Square Threat Economic 11.1% (74) 7.1% (46) 6.110 ∗ Security 16.8% (112) 16.5% (106) .022 Symbolic 4.0% (27) 3.9% (25) .022 Victim Inequalities in the Economy 1.5% (10) 5.1% (33) 13.609 ∗∗∗ Humanitarian 22.6% (151) 23.9% (154) .315 Racism 6.1% (41) 6.7% (43) .159 Hero Cultural Diversity 2.8% (19) 6.5% (42) 10.002 ∗∗ Integration 2.8% (19) 3.3% (21) .192 Good Worker 10.6% (71) 6.1% (39) 8.926 ∗∗ Note. ∗ p < 0.05; ∗∗ p < 0.01; ∗∗∗ p < 0.001. Gonçalves and David 11 Table 5. Binary logistic regressions. Variables Model 1 (Threat) Model 2 (Victim) Model 3 (Hero) B(SE) Exp(B) B(SE) Exp(B) B(SE) Exp(B) Country (Brazil = 1) .721(.775) 2.056 1.400(.547) .247 ∗ − .603(.794) .547 Newspaper (reference category is The Sun ) Estado de S. Paulo − 1.405(.778) .245 1.748(.554) 5.742 ∗∗∗ .877(.800) 2.405 O Globo − 1.551(.800) .212 2.039(.580) 7.683 ∗∗∗ .914(.838) 2.495 The Times − .871(.185) .419 ∗∗∗ .649(.183) 1.914 ∗∗∗ .386(.242) 1.471 Year of publication − .219(.079) .803 ∗∗ .181(.076) 1.199 ∗ − .087(.104) .917 Size of the item .000(.000) 1.000 .000(.000) 1.000 .001(.000) 1.001 ∗∗∗ Region (International = 1) .106(.155) 1.112 − .077(.145) .926 − .789(.211) .454 ∗∗∗ Source (External = 1) 20.095(11579.906) .000 .144(.621) 1.555 1.811(.614) 6.115 ∗∗ Constant − .273(.201) .761 − 1.575(.213) .207 ∗∗∗ − 2.234(.286) .107 ∗∗∗ Nagelkerke ’ s R 2 .052 .032 .067 N 1312 1312 1312 Classi fi cation Accuracy 73.7% 70.6% 87.0% Note. ∗ p < 0.05; ∗∗ p < 0.01; ∗∗∗ p < 0.001. 12 the International Communication Gazette newspaper is signi fi cant in estimating the likelihood of Brazilian newspapers using the victim frame. Classi fi cation accuracy was 70.6% (see Table 5, Model 2). We used the same variables as the previous model. The fi ndings show that Estado de S., Paulo , O Globo , and The Times are more likely to use victim frames in comparison to The Sun . The fi ndings also show that, over time, newspapers use more victim frames. The variables region and source were non-signi fi cant to this model. The fi ndings presented in Table 5 provide support of H2 on the likelihood of Brazilian newspapers using the victim frame in comparison to the British newspapers. To test H3 — that news items published in conservative newspapers in the UK use fewer hero frames in comparison to the conservative newspapers in Brazil — we con- ducted a binary logistic regression to estimate the likelihood of Brazilian newspapers using the hero frame . Model 3 demonstrates that the country of the newspaper is not sig- ni fi cant in estimating the likelihood of Brazilian newspapers using the hero frame. Classi fi cation accuracy was 87.0% (see Table 5, Model 3). We also controlled for each of the newspapers, year of publication, length of the news item, region and source. The fi ndings show that no signi fi cant differences were found for newspapers and year of publication. However, it shows signi fi cant differences for the length of the news item, source, and region. Longer news items are more likely to use hero frames in com- parison to shorter items. News items from external sources and news items covering the national context are more likely to use hero frames. Against our third hypothesis, the results shown in Table 5 demonstrate that neither the country of newspapers nor news- papers are signi fi cant in estimating the likelihood of using hero frames in the news. Discussion This study has demonstrated how news items from conservative newspapers frame migration in the UK and Brazil. Our study focuses on the different uses of three central frames — threat, hero, and victim — between the two countries. Although Brazil and the UK have differences in the intensity populist leaders platformed on migration, our results show that newspapers in both countries frame migration negatively, suggest- ing similar representation patterns. Both countries have experienced a rise in populism, including the Brexit referendum in 2016 and Bolsonaro ’ s election in 2018. Therefore, the political context can be a possible explanation for the repetition of media framings comparing both, in line with past studies showing a connection between populism and nativism, with increased othering and separation between ingroups and outgroups (Hughes, 2019; Ross and Bhatia, 2021; Tong and Zuo, 2019). Our results show signi fi cant differences in using the threat economic subframe. Moreover, we found signi fi cant differences in the hero subframes of good worker and cultural diversity. Finally, signi fi cant differences were found for the victim frame and its subframe of inequalities in the economy. At the same time, we found non-signi fi cant differences for the other factors. Although our regression models did not show signi fi cant differences in the threat frame based on country, we did observe signi fi cant differences in the economic subframe. The higher number of threat economic subframe in the UK suggests an alignment Gonçalves and David 13 between populist rhetoric and media frames, in line with previous showing that UK media and populist rhetoric have ampli fi ed concerns over migrants ’ economic impact, often overshadowing migrants ’ individual stories or contributions (Evans and Mellon, 2019; Ross and Bhatia, 2021). This study ’ s fi ndings, therefore, underscore a critical implica- tion: the reinforcement of a singular narrative can narrow public discourse, limiting dis- cussions to threat-based contexts and potentially affecting public perception and policy making (David, 2022; McLaren et al., 2017; Merry, 2022). The fi ndings show that news items from conservative newspapers in the UK use fewer victim frames in comparison to conservative newspapers in Brazil. Signi fi cant differences were found for the subframe inequalities in the economy, and our regression model shows signi fi cant differences for the victim frame. The higher number of newspapers using the subframe inequalities in the economy in Brazil aligns with studies showing the framing of Venezuelan refugees as a mass of victims in precarious conditions (Morais and Santos, 2017; Ranincheski and Uebel, 2018). The emphasis on Venezuelan refugees ’ challenges re fl ects broader regional dynamics and Brazil ’ s political positioning. Such as news items from the UK, Brazilian coverage focus on economic issues, but from another point of view. Here, stories mostly enlighten the economic dif fi culties faced by migrants, positing a broader understanding of the complexities surrounding migration. Yet, the focus on victimization also suggests an alignment between populist rhetoric and media framing, as Bolsonaro focused on the dif fi culties faced by Venezuela and its citizens caused by a dictatorship regime of the left spectrum (Brumat and Geddes, 2023). The results con fi rm our hypothesis that news items from conservative newspapers in the UK are less likely to use victim frames in comparison to conservative newspapers in Brazil. Contrary to the hypotheses that British news it