Risk Narratives on Immigration During the COVID-19 Crisis in Italy: A Comparative Analysis of Facebook Posts Published by Politicians and by News media

Abstract

This work explores how the narrative on immigration changes when society is threatened by “real” risks, i.e., during the COVID-19 health crisis. We compared the frequency and engagement of over 348,684 posts published on Facebook between December 2019 and November 2020 by Italian politicians and news media. We identified two waves of  “tangible crisis” assuming that in these two periods the risk stemming from COVID-19 was strongly perceived by the Italian population, contrasting our observations to the periods preceding the first wave and between both waves. Our findings suggest that the political discourse and risk narratives on immigration decreased during times of “tangible crises” for right-wing populist parties and news media. This happened at a time when key policies regarding immigration and regularization of migrants were highly discussed by the Italian government, receiving unexpectedly low reactions. This leads us to theorize that anti-immigrant communication decreases during times of “tangible crises.” 

Introduction

In public, political, and mediatized discourse, the topic of immigration is multidimensional and the relevance of its different aspects differs among countries and along time (Lecheler et al., 2019). Despite these contextual dependencies, a vast body of literature shows that (1) some common key frames and metaphors on immigration exist across countries and time, wherein immigrants are often portrayed as a risk, that is, they are described as a “common enemy” to the economy, culture, and security of the receiving country; (2) the way in which immigrants are portrayed in mediatized discourse (i.e., sentiment and framing) affect public attitudes negatively; (3) once anti-immigration metaphors become accepted or internalized,  this influences the public to take action or support repressive policies toward immigrants (Lecheler et al., 2019) or even to hate crimes due to xenophobia (Padovani, 2018).

Previous empirical evidence suggests that immigration is often framed as a risk and threat to the host society particularly in right-wing media outlets and by right-wing populist parties and politicians: the “risk” stemming from immigration to a host society is often blown out of proportion via the use of metaphors, frames and paradigms that portray migrants as a dangerous “out-group” (Boeynaems et al., 2021). Both the immigration and refugee situation (as was observable since 2015) and the COVID-19 pandemic of year 2020 are crises—defined as “exceptional events that may interrupt journalistic routines and create opportunities for newly emerging aspects and interpretation of an issue” (Greussing & Boomgaarden, 2017, p.1750; Horsti, 2008). Moreover, the discursive construction of a crisis is highly linked to the concept of risk (Hier & Greenberg, 2002).

In this work, we consider risks associated to immigration as a “constructed” or perceived risk, that is, a risk often not associated with real and concrete evidence but only rhetorically constructed for political purposes by some parties and media outlets, particularly those of the right. In contrast, despite misleading information and fake news surrounding the COVID-19 crisis (Eberl, Huber & Greussing, 2021), the risk stemming from the pandemic is mediatized in more concrete terms, as statistics of daily new hospitalizations and deaths are available and visible, and media report evidence that the risk is tangible.

In this context, we aim at understanding how the topic of immigration is mediatized in political and mainstream media communication during times of “tangible” crisis. We refer to “tangible crisis” as periods in which the COVID-19 risk is most observable (i.e., tangible) for the majority of the Italian population. For our analysis, we identify two waves of tangible crisis in Italy, the “first wave” corresponds to the period between February 21st to May 3rd, 2020 (the first national lockdown), the “second wave” corresponds to the period from September 1st to November 15th, 2020, when the rate of confirmed cases started to increase rapidly again following a period of steady rate of confirmed cases (WHO, 2020). The aim of the analysis is to understand the effect that a tangible crisis such as that of COVID-19 has on a “constructed” crisis such as that of immigration. In this context, right-wing media outlets and right-wing populist (RWP) communication by parties and politicians on the right spectrum constitute a privileged object of analysis, since these media outlets, parties, and politicians usually frame immigrants as a dangerous out-group and use a general rhetoric of risk detached from statistical accuracy.

So far, a growing body of research explores media and immigration as well as the immigration discourse on social media (Lecheler, Matthes & Boomgaarden, 2019; Padovani, 2018; Heidenreich et al., 2019). However, to our knowledge, not many studies focus both on a comparison between media and political communication. Furthermore, we contribute to the analysis of immigration discourses in the context of crisis events, in the framework of sociological theories of risk. This study thus offers an interdisciplinary approach, utilizing theories from sociology and communication science and social analytical methods.

Theoretical Framework 

Social media, Right-Wing Populism, and the Topic of Immigration

One essential characteristic of populism is the idea of direct expression of citizens without institutional mediation (Tarchi, 2015). The core concept of populism has been defined as an antagonistic relationship between “the people” and “the elites,” which can be easily linked with other ideologies such as nationalism, socialism, communism, or ecologism (Aslanidis, 2018; Mudde, 2004) and which argues that politics should be “an expression of the volonté générale (general will) of the people” (Mudde, 2004, p. 543).

However, the concept of populism is so broad that it varies greatly depending on the context within which it is analyzed. In the United States, for example, the rise of populism is generally traced to the presidency of Andrew Jackson in the 1830s and the subsequent formation of the People's Party in the 1890s. The latter was an agrarian political party which aimed to represent the interests of farmers and emerging industrial wage workers, opposing the monopolistic concentration of capital by industrialists and banks (Goodwyn, 1976). The expression “anti-establishment,” in that context, mainly refers to the financial establishment, perceived as parasitic (Lowndes, 2017).

In the contemporary Italian context, instead, populism has a more strongly anti-political connotation. In 1992, in fact, an investigation called Mani Pulite (Clean Hands) revealed a widespread corruption system implemented by some members of the main Italian parties of the time (the Socialist Party and the Christian Democrats). This caused the collapse of these same parties and the end of the so-called “First Republic”; it was at the origin of a widespread anti-political sentiment, also known as “anti-caste” (the term “caste” refers to the Italian political class), which characterized Italian populism in the following decades (Tarchi, 2015).

The breadth of the concept of populism also implies different characterizations according to different political ideologies. We can therefore have a left-wing populism, as well as a right-wing populism. Right-wing populism is a populism that incorporates exclusionary ideas (Heinisch et al., 2019). Moreover, while the exclusion of groups is common to all populists, the exclusion of specific groups such as immigrants is characteristic of radical right wing populist parties (Gründl, 2020). Previous research also shows the portrayal of migrants as dangerous “cultural others” is intrinsically linked to the characteristics of far right and populist parties that follow a nationalist agenda and try to appeal to voters through fear mongering (Heisbourg, 2015). Importantly, as Gründl and Aichholzer (2020) note, the preference for a homogenous “ingroup” as well as exclusive social identity (a.k.a. inclusive-exclusive intergroup attitudes) has been shown to be motivated by uncertainty reduction (see Hogg, 2000).

Moreover, as Gerbaudo (2014) points out, social media platforms such as Twitter or Facebook allow populist communication to thrive, as they offer an arena for political actors to interact directly with potential voters, bypassing the mediation of traditional gatekeeping by journalistic values (Wirth et al., 2016). The setting of “structurally disintermediate” communications (Bracciale & Martella, 2017) provides a perfect ground for populist communication styles (Gerbaudo, 2014), and allows political actors to actively share and propagate unverified information directly to their followers.

Immigration has been found to be a highly divisive topic on social media as well as the most distinguishing factor of Right-Wing Populist (RWP) communication in Austria and Germany (Heiss & Matthes, 2020). This is problematic, as anti-immigration attitudes move citizens to follow RWP actors on Facebook, which further fuels anti-immigration attitudes (Heiss & Matthes, 2020).

Right-Wing Populism in Italy and the Topic of Immigration

Italy, due to its geographical position, has experienced significant migratory flows for decades, especially from North African countries. In recent years the issue of immigration has gained salience in the political and media debate. This occurred primarily as a consequence of the 2015 refugee crisis, which significantly increased the level of polarization of the public debate between pro-immigrants and anti-immigrants (Di Mauro & Verzichelli, 2019). The issue of immigration has been progressively monopolized by the right-wing and populist parties that have electorally grown in recent years, in particular the Lega, Fratelli d’Italia and the Five Star Movement, mainly through the communication they conveyed on social media (Combei & Giannetti, 2020).

In contemporary Italian politics, the origin of populism, as already noticed, can be traced back to the Mani Pulite judicial scandal of 1992 and to the consequent anti-establishment political sentiment that has spread among large sections of the population. Since the 2000s, this anti-political sentiment has found expression in the affirmation of the Five Star Movement. The Movement was founded in 2009 by the comedian Beppe Grillo, who for some years had been commenting on the events of Italian politics on his blog, railing against the corruption of “professional politicians,” exalting the honesty of citizens and invoking the advent of a bottom-up democracy in which all citizens could express themselves (through the web and digital platforms) on issues of public interest (Lanzone & Woods, 2015). In 2013, in the political elections, the Five Star Movement became the most important party in Italy, obtaining 25 percent of the votes and electing 163 deputies.

Although the populism of the Five Star Movement has been defined as “eclectic” (Mosca & Tronconi, 2019), that is, beyond the classic political categories of right and left, the elements of disintermediation (connected to the web and social media) and anti-establishment, in Italy, have been absorbed in particular by right-wing parties such as Lega and Fratelli d’Italia. These two parties can therefore be defined as right-wing populists: to the typical elements of populism, they have combined some specific right-wing issues, such as neo-nationalism and social conservatism. Specifically, Lega and Fratelli d’Italia have distinguished themselves over the years for their positions in defense of Italian culture and national identity against perceived attacks by outsiders, thus developing a solid anti-immigrant rhetoric.

Therefore, in Italian politics, a substantial opposition has arisen between a populist right, strongly focused on the idea of protecting the Italian population from a series of external “threats” (financial capital, power of the European institutions, migrants), and a center-left (represented mostly by the Partito Democratico) centered on a globalist, pro-European, cosmopolitan agenda, and thus more open to the inclusion of migrants on Italian territory. The Five Star Movement, connoting itself as a “post-ideological,” neither right-wing nor left-wing party, has formed alliances with both coalitions. In the last elections, held in 2022, the populist right-wing parties Lega and Fratelli d'Italia, together with the center-right party Forza Italia, obtained a large majority in the Italian parliament, relegating the Democratic Party and the Five Star Movement to the opposition.

Risk and the Topic of Immigration

In this work, we consider threat perception to be one element included in the overarching notion of risk perception. In seminal literature on social approaches to risk, risk perception is understood to some extent as socially constructed (e.g., Beck, 1992; Douglas, 2013; Slovic, 1999). In this sense, it is understood as “an idea in its own right” as the perception generated by both social processes (representation and definition) as well as the actual experience of harm (Burgess, Wardman & Mythen, 2018). Sniderman, Hagendoorn & Prior (2004) have highlighted that strong public concern of the issue of immigration is measured even by citizens who have little or no direct experience of immigration (Boomgaarden, McLaren & Vliegenthart, 2017). Both, the risk from COVID-19 as well as the perceived risk from what is often termed an immigration “crisis” or simply from the perceived threat of immigration, present as “common” and global threat, possibly affecting anyone in society. Furthermore, both immigration and health threats such as COVID-19 are often framed as risks in the media and on social media (Liu, Zhang & Huang, 2020; Eberl et al., 2018; Gottlob & Boomgaarden, 2020). Several studies show that the way in which risks are framed in the media on a particular subject can affect people's risk perception as well as individual and public attitudes toward these issues (Scheufele, 2000; Gottlob & Boomgaarden, 2022). Because of this perception/relatability, both issues can therefore be instrumentalized by parties and politicians to mobilize their voters. Media coverage and political communication on immigration can be seen as reflective of this particular “uncertainty and uncontrollability” element of risk society (Ungar, 2001) by feeding into the narrative of immigration representing a risk, threat, or crisis. Research shows that in general, people are more prone to accept risks if they are voluntary, natural, familiar, and/or controllable. On the contrary, risks surrounded by uncertainty and which are perceived as being less controllable are less likely to be accepted (Fischhoff et al., 1978; Slovic et al., 1979) and have been characterized as more “dreadful” (Morgan, 1993) ¹ .

The overlap between the manufactured immigration crisis and the tangible Covid crisis could be explained as a “displacement effect”, defined by media studies (Bryant & Fondren, 2009). The Displacement Media Theory proposes that people's limited time and attention in one communication activity takes away from participation in others. Journalism studies highlight that, when certain conditions change, a “goal displacement” can be generated in the media outlets (Harlow, 2015), and then the newsworthy topics can radically change. Transferred in the context of risk perception, these displacement effects could explain why the (unexpected) Covid crisis, bursting into both people's lives and among the topics covered by the newspapers, did not leave more room for the migration crisis. However, the right-wing anti-immigration posts immediately spread again at the end of COVID emergency and in the presence of many other potential risks (e.g., post-COVID economic crisis, limitations in social life, lack of attention to other life-threatening diseases) that were not aligned with the right-wing agenda. Hence, the displacement theory by itself cannot explain the special attention to immigration in the right-wing agenda.

In light of these premises, we therefore ask the question:

RQ: Does the typical right-wing populist communication style of “building a common enemy,” through the portrayal of immigration as a risk, decrease during the times of a “tangible crisis”?  And if so, how does this vary between political party posts, media posts, and in terms of user engagement?

To address this question, we perform a quantitative and qualitative analysis to compare: 1) the frequency and engagement, and 2) the type of content related to the immigration topic in Italian media and political narratives on Facebook before, during and after the first and second waves of COVID-19 in Italy (December 1st, 2019, to the 15th of November 2020). Specifically, frequency corresponds to the number of posts related to the immigration topic and engagement is defined as the sum of interactions (comments, shares, reactions) received by these posts on Facebook as measured by Crowdtangle.

We formulate the following two hypotheses:

H1: The frequency of immigration related posts significantly decreases during times of tangible crisis for far right and right-wing populist parties and media in comparison to mainstream and left-wing parties and media. 

H2: The engagement of online users for immigration related posts significantly decreases during times of tangible crisis for far right and right-wing populist parties and media in comparison to mainstream and left-wing parties and media.

Methodology

For our quantitative analysis, we defined a dictionary related to immigration in general to analyze the variance in frequency on immigration-related posts. The data allowed us to identify/measure the number of posts related to immigration posted by/within and across different political groups (political parties and politicians) as well as “media groups” (media outlets from left to far-right), during/across the periods of interest. We collected a total of 276809 posts, out of which 71875 posts were associated with the 7 parties that we have considered in our study (see below).

This first step allowed us to characterize the differences in the number of posts (frequency) which include the topic of immigration/terms related to immigration (from the dictionary, see appendix table 1.). In order to also understand differences in content of these posts between the analyzed groups, in particular between the left and the right, we used a sub-sample of the most popular/liked posts, and performed a second, in depth-qualitative analysis. In particular, we were interested in understanding whether the posts on immigration contained risk and threat related terms and if so, in which context across our groups of analysis. We therefore collected the 10 most popular posts from each page associated with the seven groups, i.e., a total of 70 posts; the accurate qualitative reading of these posts allowed us to understand deeper meanings, possible audience interpretations, and context (Macnamara, 2005).

Table 1.

Statistics on Groups of News Media in our Dataset.

Media	No. followers	No. migration posts	No. risk posts	No. total posts
FAR-RIGHT	588,334	1,797	4S5	20,629
RIGHT	735,504	3,181	564	42,471
MAINSTREAM	18,572,399	5,516	708	202,386
LEFT	521,922	1,052	314	11,323

Quantitative Analysis

Data Collection

To collect posts related to the topic of immigration, we created a list of keywords related to immigration in Italian (such as “migranti” and “profughi”), using an approach that was effective for studying the Italian disinformation by Pierri, Artoni & Ceri (2020). On top of the keywords included in this dictionary, we also added terms related to events of relevance and policies concerning immigration in Italy, such as “Lampedusa” or “Decreti Sicurezza” (for more details, see Table 1 on the immigration dictionary in the appendix). The keywords were therefore both latent when relating to specific immigration events or locations in Italy (e.g., “Lampedusa”), as well as manifest (example “immigrati”).

We captured different phases of the COVID-19 crisis, aka. periods of interest (see the delimitation of the periods of interest below) by collecting data posts shared by Facebook pages of Italian politicians and newspapers in a period of 11 and a half months from December 1st, 2019 to November 15th, 2020. Specifically, we used the Crowdtangle API (Crowdtangle Team 2020) for collecting posts containing specific keywords within dictionaries (see below). Crowdtangle is a public insights tool owned and operated by Meta, that allows academics, journalists, and practitioners to perform keyword-based searches of posts shared by public Facebook pages and groups that either meet a requirement in terms of number of followers or that have been selectively added by users of the platform. More details on the coverage of pages and groups are available on the official documentation page (https://help.crowdtangle.com/en/articles/4201940-about-us).

We considered accounts of main politicians and associated political parties covering the entire political spectrum from left to right. The nominal category of Left can be defined by “ideas such as freedom, equality, fraternity, rights, progress, reform and internationalism,” whereas the nominal category of Right by “notions such as authority, hierarchy, order, duty, tradition, reaction and nationalism” (Heywood, 2015). Finally, we followed the classification made by Rooduijn et al. (2019) on populist, far right, and far left parties, which includes the Five Stars Movement as a populist party, Lega and Fratelli d’Italia as far right parties. For the reasons explained previously in the paper, the characterization of populist parties was also added to Lega and Fratelli d'Italia. Therefore, the classification of politicians and parties whose accounts were monitored is as follows:

1. Matteo Salvini and “Lega per Salvini Premier” (LEGA, far-right, right-wing populist), 2) Giorgia Meloni and “Fratelli d’Italia” (FDI, far-right, right-wing populist), 3) Nicola Zingaretti and “Partito Democratico” (PD, center-left), 4) Luigi di Maio and “Movimento 5 Stelle” (M5S, populist). Similarly, we collected data for principal Italian newspapers aggregating them as follows: 1) La Repubblica, Corriere della Sera, La Stampa, Huffington Post and Fanpage (MAINSTREAM), 2) Il Manifesto, Avvenire (LEFT), 3) IL Giornale, Il Tempo (RIGHT), 4) Libero, La Verità (FAR-RIGHT).

For each post, we collected the number of public interactions such as shares, likes, reactions, comments, etc. We consider the number of posts within pages as an indication of the page's activity, and we compute the associated engagement of online users as the sum of daily interactions. Tables 1 and Table 2 below show descriptive statistics on the different groups. Note that for what concerns political leaders/parties, as shown in Table 1, the total amount of immigration-related posts by LEGA pages is much higher than the rest; for what concerns newspapers (Table 2), MAINSTREAM newspapers are most active.

Table 2.

Statistics on Groups of Political Parties/Leaders in our Dataset.

Politicians	No. followers	No. migration posts	No. risk posts	No. total posts
LEGA	4,808,330	6,677	2.620	41,696
FDI	1,719,076	396	183	4,107
M5S	3,744,775	253	68	10,063
PD	646,161	349	176	6.911
IV	1,237,755	163	S3	5.883
FI	1,282,399	92	41	2551
CONTE	2,030,400	14	2	664

Because our data is not normally distributed, we employed a non-parametric Kruskall Wallis test, to statistically analyze differences (in medians) across various groups (posts by different political parties and media) during our periods of interest.

Periods of Interest

As described in the introductory section, we specify four periods of interest to investigate our research questions. Specifically:

1. Pre-virus (01/12/2019–21/02/2020)

2. 1st wave (22/02/2020–04/05/2020)

3. Summer (05/05/2020–31/08/2020)

4. 2nd wave (01/09/2020–15/11/2020)

Qualitative Analysis

To collect the posts that we needed for the qualitative analysis, we defined a second dictionary containing keywords connected to risk and threat relative to immigration, based on typical (mostly right-wing) lexica connoted to risk and threat on the immigration topic. In this way, we could identify posts containing “risk narratives,” i.e., terms and keywords connoted to risk and threat, within the immigration post sample used in the quantitative analysis. We then extracted a sample of the ten most popular (with the most engagement) posts in this category manually. For each political group, we used the following procedure: we computed the engagement of each post, i.e., the amount of interactions received; then we listed these posts in a decreasing order by their engagement; finally, we selected the top 10 of such a list. We used the content analytical method of critical discourse analysis (Fairclough, 1989; Van Leeuwen, 1996) to understand more on the content within the posts. We based our dictionary construction on existing content analytical research in this field (Maltone-Bonnenfant, 2011; Milkowska-Samul, 2018, see appendix.)

Risk and Threat Dictionary Construction

To construct the risk and threat dictionary, we used a list of terms related and synonymous to risk in the Italian language such as “rischio, pericolo, incertezza, insicurezza, minaccia, imprudenza, danno, azzardo” as well as a list related to risk and immigration in Italian media and on social media platforms in Italy. For this latter step, we considered prior content-analytical findings, regarding commonly used words that portray immigration as a risk to the host society (in this case, Italy), which are mostly related to lexica of danger, insecurity, and criminality (explained in the theory section). Taking all this into consideration, the dictionary related to risk and threat is formed using general risk terms as well as risk terms related to migration (see Tables 2 and 3 in appendix).

Results

Frequency of Immigration Topic

We first use the dictionary related to immigration to investigate hypotheses H1 and H2, which are related to the frequency and engagement of immigration related posts by Facebook pages of political leaders/parties and newspapers. For both metrics, we consider the percentage value with respect to the value computed when considering all posts shared by pages in the period (i.e., including also posts not related to immigration). We employ a Kruskal-Wallis test, which is non-parametric as it does not make any assumption on the distribution of the data, to assess significant changes in these metrics across different periods.

H1: The frequency of immigration related posts significantly decreases during times of tangible crisis for far right and right-wing populist parties and media in comparison to mainstream and left-wing parties and media. 

For politicians, Figure 1 provides the time series of the daily percentage of posts related to immigration (we highlighted different periods with different background colors), whereas Figure 2 summarizes the results of our statistical tests in a heatmap where each cell describes the difference between median values of the daily % of posts related to immigration two different periods. Rows indicate different political groups and columns indicate pairs of periods which are compared. Colors are mapped to indicate positive variations in red and negative variations in blue. White cells indicate non-significant changes. We observe that after the COVID-19 outbreak LEGA and M5S exhibit a significant decrease in the frequency of immigration-related posts, whereas FDI and PD do not exhibit significant changes. Exiting the lockdown, LEGA and M5S exhibit a very large increase in their activity (respectively + 18 and + 10), followed by PD (only + 5.9). M5S does not exhibit significant changes. When the 2nd wave approaches, LEGA consistently decreases again its activity (−6.8%), followed by PD which however shows only a slight decrease (−1.7). M5S and FDI instead do not exhibit any significant change.

Figure 1.

Time series of the daily percentage of posts related to immigration for each political group. Thick lines indicate a 7-day smoothed average. Different periods are highlighted with different background colors.

Figure 2.

Heatmap that shows the difference between median values of the distribution of daily % of immigration related posts across periods. Each row represents a political group, and each column compares two different periods. White cells correspond to non-significant differences.

Concerning the media, the time series of the daily percentage of posts related to immigration in Figure 3, and a breakdown of the statistical results in Figure 4, shows a heatmap with the same rationale of Figure 2. We observe that all media groups significantly decrease their activity when the COVID-19 pandemic hits, with the largest drop in far-right activity (−5.4) and the smaller change in Mainstream newspapers (−1.5).

Figure 3.

Time series of the daily percentage of posts related to immigration for each news media group. Thick lines indicate a 7-day smoothed average. Different periods are highlighted with different background colors.

Figure 4.

Heatmap that shows the difference between median values of the distribution of daily % of immigration related posts across periods. Each row represents a news media group, and each column compares two different periods. White cells correspond to non-significant differences.

Towards summer, all groups of pages increase their activity significantly, with the largest change in Far Right and Right leaning newspapers (respectively + 7.6 and + 4.6) and only a slight change in Left and Mainstream newspapers (less than 2). During the 2nd wave, all groups decrease significantly except Left leaning newspapers.

H2: The engagement of online users for immigration related posts significantly decreases during times of tangible crisis for far right and right-wing populist parties and media in comparison to mainstream and left-wing parties and media.

Concerning politicians, as shown in Figure 5 and Figure 6, we note that engagement follows a pattern similar to the frequency of posts related to immigration. All groups exhibit a decrease when the lockdown is applied, except for the far-right FDI party, and they all increase when the lockdown is released; in particular, LEGA and FDI have the largest increase (respectively + 11 and + 6.8) whereas PD has only a slight increase of 2.7. Finally, LEGA only exhibits a pronounced drop when the second wave approaches (−5.5).

Figure 5.

Time series of the daily percentage of engagement of posts related to immigration for each political group. Thick lines indicate a 7-day smoothed average. Different periods are highlighted with different background colors.

Figure 6.

Heatmap that shows the difference between median values of the distribution of daily % of engagement of immigration related posts across periods. Each row represents a political group, and each column compares two different periods. White cells correspond to non-significant differences.

Concerning media, we observe similar results as in H1, as shown in Figures 7 and 8: all groups decrease their engagement during the 1st wave, whereas during summer the largest increase is observed for far right and right-wing newspapers (respectively + 6.6 and + 9.6), consistently with our hypothesis. Likewise, these groups have a large decrease during the 2nd wave (−4.9 and −7.5).

Figure 7.

Time series of the daily percentage of engagement of posts related to immigration for each news media group. Thick lines indicate a 7-day smoothed average. Different periods are highlighted with different background colors.

Figure 8.

Heatmap that shows the difference between median values of the distribution of daily % of engagement of immigration related posts across periods. Each row represents a news media group, and each column compares two different periods. White cells correspond to non-significant differences.

Qualitative Analysis of Risk and Threat Terms

As a second step, we perform a qualitative analysis using critical discourse analysis (Fairclough, 1989; Van Leeuwen, 1996). We remark that posts by left wing politicians tend to be used in a different/opposing context than the ones by right-wing politicians. Based on prior literature, we assumed that left-wing politicians use risk-related terms with a positive stance toward immigration, in a context of “criticism of the right,” wherein anti-immigration claims, and opinions of far-right politicians are referenced/paraphrased or cited in a negative/critiquing or condemning way. In this sense, we can treat the risk keywords as a double negative in terms of valence/meaning for the known pro-immigration politician. 

This assumption was confirmed by an in-depth, qualitative evaluation of the most popular posts in our sample: we found that risk-terms were used in opposing contexts by right-wing versus left and remaining parties and pages. Posts from right media and political pages used a negative value of risk-related narratives, while the posts on the left pages cited anti-immigration rhetoric in a critical and condemning way.

Among the 10 most popular posts from the center-left party Partito Democratico and its leader Nicola Zingaretti, there is mainly explicit criticism of the association between the arrival of migrants and the spread of the virus made by right-wing parties in the period following the first wave of Covid, that is, in the summer of 2020. On August 25, for example, this post appeared on the Facebook page of Partito Democratico, which quotes a statement by Salvini: “A Lampedusa i migranti passeggiano tra i turisti che poi portano il covid nelle loro regioni’. L'ennesima menzogna di Matteo Salvini” (“In Lampedusa, migrants walk among tourists, who then bring covid to their regions. Yet another lie by Matteo Salvini”). Additionally, among the most popular posts from the center-left party, there are arguments that recall the Covid emergency to highlight how it should serve as a lesson to understand the suffering of other people, such as migrants, who are constantly in a situation of difficulty and risk. These are therefore narratives about risk and immigration based on a rhetoric of welcoming and inclusion of migrants.

As for the top ten posts from the Five Star Movement and its leader Luigi Di Maio, there is on one hand a narrative of risk that highlights how it is important for everyone to respect the rules, both Italian citizens and immigrants, to avoid the spread of the virus. However, this narrative never comes to consider immigrants as an explicit risk to the Italian population. There are also explicit criticisms of right-wing parties and their narratives that consider migrants a risk and threat to Italy.

Regarding right-wing party Lega and its leader Matteo Salvini's most popular posts, we found posts that explicitly associate the risk of COVID-19 infection with the arrival of migrants, especially in the period between the first and second wave, when the emergency in Italy seemed to be over. On August 16, for example, Salvini wrote on his Facebook page: “Incredibile. Il governo cambia un’altra volta idea e torna a chiudere negozi e locali, mettendo a rischio migliaia di posti di lavoro, mentre il virus viene importato dall’estero. Un governo duro con gli italiani e morbido con i clandestini”. (“Unbelievable. Government changes its mind once again and closes shops and bars, jeopardizing thousands of jobs, while the virus is imported from abroad. A harsh government with the Italians and soft with the illegal immigrants”).

This post can be analyzed using transitivity analysis (Halliday, 1994), which in the critical discourse analysis is defined as the study of “what people are depicted as doing and refers, broadly, to who does what to whom, and how. This allows us to reveal who plays an important role in a particular clause” (Machin & Mayr, 2012, p. 104). Here Salvini initially uses the rhetorical strategy of representing the action in abstraction (Fairclough, 1989) when he states that “the virus is imported from abroad” without specifying from whom it is imported. Immediately afterward, however, he refers to illegal immigrants. The latter, in the post, are not directly characterized as those who import the virus (it is not said: “the virus is imported by illegal immigrants”). Nevertheless, their agency as carriers of the virus is implied in the reference to the government's “softness” toward them, that is mentioned in the following sentence.

We also note that the most popular posts on the pages of Salvini, Meloni, Lega, and Fratelli d’Italia are almost all concentrated in the period between May and August, after the end of the first COVID-19 crisis and before the beginning of the second one. During that period, on RWP pages we found fearmongering posts that stressed single news stories regarding immigrants.

On June 30, Meloni emphasized the story of an immigrant who would have killed an “innocent” kitten by roasting him: “Immagini raccapriccianti! clandestino cuoce come se nulla fosse un povero gattino, tra urla e la (comprensibile) rabbia di una signora che lo riprende. per questo clandestino tolleranza zero: in italia non sono consentite queste barbarie” (“horrifying images! an immigrant cooks a poor kitten as if nothing happened, amidst screams and the (understandable) anger of a lady who filmed him. zero tolerance for this delinquent: these barbaric acts are not allowed in Italy”).

Using critical discourse analysis, we can notice how in this case, the protagonist of the story is represented not as a specific individual but as a generic type, that is, “an immigrant.” As highlighted by Van Leeuwen (1996), this is a strategy of “impersonalization” and “genericization”: the use of generic terms can be used to give the story a racist connotation, since it creates a frame in which immigrants are characterized as being dangerous for the host nation. In this post, we also notice a strategy of “overlexicalization,” which “results when a surfeit of repetitious, quasi-synonymous terms is woven into the fabric of discourse.” (Teo, 2000, p. 20). The image in the post is repeatedly described as “horrifying” and “barbaric”—quasi-synonymous terms that emphasize the gravity of the actions carried out by the immigrant.

This analysis indicates that our sample of ten most popular posts for each source was effective for enlightening the type of narrative on immigration conveyed by different parties and media in the different considered periods. This goal required a thorough reading of a sub-sample of the overall dataset containing keywords explicitly referring to risk and threat. We considered 10 posts because this number was sufficient for confirming, using a magnifying glass, the findings that emerged from the quantitative analysis; a detailed reading of a larger number of posts was not necessary to answer our qualitative research question and therefore was outside of our study, whose primary goal is to confirm the quantitative analysis findings.

Discussion

Our quantitative analysis findings indicate that on average during both the first and second wave, posts on immigration by right-wing populist party Lega Nord decrease by a remarkable amount compared to pre-wave periods. In the period between the two waves, however, the posts of the Lega Nord increased significantly. A partially different trend is visible for the other right-wing populist party, namely the Fratelli d’Italia: while mirroring the Lega Nord's trend concerning the increase of posts in the summer, it does not show significant drops during the two waves. Similar trends are visible for engagement levels of online users with the same pages. These different trends between Lega and Fratelli d'Italia partially confirm our two hypotheses, namely that the frequency of immigration-related posts by right-wing actors and the engagement of the followers decreases during times of tangible crises.

In contrast, the center-left party PD shows very limited variations during the two waves, both in the number of posts and in the level of engagement. We also note that while the overall number of posts of the PD is lower, variations during the summer (especially in the number of posts) are more significant and mirror those of right-wing parties. This last trend may be explained in light of the qualitative analysis of the top ten posts. Within these posts, center-left parties often use risk-related terms with a positive stance toward immigration, in a context of “criticism of the right,” therefore as a form of “reaction” to the posts of the right. Thus, we may argue that, although center-left parties exhibit much less variation during all phases, such variations mirror those of the right, whose narratives the center-left parties must oppose. The fact that these variations are overall much more limited, however, suggests that the narratives on risk and immigration of the center-left parties are less affected by COVID-19 crisis.

Furthermore, our findings concerning the online news media also reinforce our hypotheses to a certain extent. The trend of right and far-right newspapers mirrors that of right-wing party Lega Nord, with significant drops both in the number of posts and engagement during the first and second wave, and a strong increase between these two waves, when the number of COVID-19 cases in Italy had dropped. Mainstream and left-wing newspapers show similar trends concerning the number of posts, but with more limited variations, especially regarding the increase of posts during the summer. Therefore, overall, the findings support our hypotheses (H1 and H2) that the frequency of immigration-related posts by right-wing actors and the engagement of the followers decreases during times of tangible crises, even if the partial exception of Fratelli d’Italia must be noted. This right-wing populist party, in fact, does not always mirror the trends of the other right-wing actors (parties and media). Additionally, the qualitative analysis helps us to confirm that, in line with what prior research and literature have shown, only right-wing parties refer to immigration in terms of risk and threat.

Several additional observations corroborate the main quantitative and qualitative findings. Noticeably, during the first wave, the Italian government proposed a law for the regularization of 600,000 immigrants, which was then approved in May 2020. Under normal conditions, such a law would have raised a wave of protests from right-wing parties and would have produced a spike in immigration-related posts. However, both our analyses suggest that activity of right-wing parties was almost absent in March and April, when the law was proposed but the country was in the midst of the COVID-19 crisis. Then, toward the end of the first wave up until August 2020, the salience of the immigration topic increased by a considerable amount in RWP politicians.  During that period, in Italy the virus had almost disappeared, with an average of 100–200 daily infections (unlike many other European countries where infections were more numerous). Significantly, therefore, we postulate that when the virus was no longer an “internal” threat to the country, the narrative of the virus as an “external threat” brought by immigrants regained strength, similarly to what happened in the period preceding the first wave.

Moreover, the results indicate that even after several months into the pandemic, when the numbers of COVID-19 infections started rising significantly again on a national level, far right and right-wing media as well as the right-wing party Lega Nord posted less content on the topic of immigration.

A limitation of our analysis is that we only monitored the political communication of parties and news media on a single social network, Facebook, which might not be fully representative of their communication agenda. Further, our data analysis was performed with a combination of manual choice of keywords, which might not completely capture relevant data, and automated content analysis on a very large sample which doesn’t allow for manual verification. Finally, we cannot discern the specific way in which parties and media outlets employ keywords related to immigration in general across the political spectrum, as this would require to manually analyze thousands of posts. However, we performed a qualitative study on a subsample of the seventy most popular posts to understand the different context and rhetorical strategies employed by different parties and media outlets from Left to Right and our analysis confirmed that our keywords were effective.

In closing, our findings provide indications on the effects of a tangible crisis in the context of the so-called post-traditional societies, in which the perception of risk is considered a structural dimension in the construction of identity and in the relationship between the individual and the social order to which he/she belongs. The COVID-19 crisis is a global problem that has an immediately concrete dimension (people's health, illness) and not an abstract one such as, for example, the climate crisis (which refers to the rights of future generations). It can therefore effectively stimulate a more interdependent and less nationalist awareness of risk. In Beck's terms, it can be conceived as a crisis that could present an opportunity to reach on a European level, a “cosmopolitan entente” that exceeds the nation state in a common mobilization against these risks and manufactured uncertainties (Beck, 1996; 2006).

Our findings support the thesis that the COVID-19 crisis, due to its characteristics, has oriented the perception of risk, which is structural to post-traditional social orders, toward a supranational and cosmopolitan response. However, we should note that this is only a short- term observation, and that we do not imply that the popularity for right-wing populist parties will therefore decrease in terms of voter turnout. In particular, considering the implications that the infodemic, misinformation, and conspiracy theories surrounding COVID-19 on social media (Boberg et al., 2020) has for trust in governing institutions, we cannot expect RWP rhetoric to lose strength in the long term only by investigating the topic of immigration. In a recent study, Eberl, Huber & Greussing (2021) have provided evidence for their argument that the COVID-19 pandemic provides fertile ground for populists’ opposition to scientific and political elites. Their results confirm a positive relationship between populist attitudes and conspiracy theories independent of political ideology. An interesting question for further research would thus be to investigate whether in times of tangible crises, anti-immigration narratives in RWP rhetoric observed in our results are replaced by antagonism to scientific and political elites.  

Appendix

• 1. Migration Dictionary

To construct the migration dictionary, we based our keywords on a prior list that was used in a study published a short time prior to our investigation (Pierri, Artoni & Ceri, 2020). There, keywords were obtained with a data-driven approach of the “Top-500 most frequent words appearing in the titles, and taking into account relevant events that occurred in the last months” (Pierri, Artoni & Ceri, 2020: 11). The authors provided top-20 keywords for immigration. We then added terms related to events to do with immigration in the particular time of analysis, such as “Musumeci” (keyword related to an event in August 2020) to make our list more extensive. The additional words are based on all major/newsworthy events that happened in that time-period and the keywords related to them as used in the mainstream media. The approach was suggested by the authors of the study, of which one was also working as a journalist (in Italy) at the time and thus had expert knowledge on the media landscape and terminologies used in relation to the topic of migration in Italy.

Table A1.

Migration Dictionary.

immigrazione, migranti, migrante, immigrato, immigrati, Decreti Sicurezza, decreto sicurezza, regolarizzazione, sanatoria, Ong, clandestini, clandestino, irregolari, barconi, barcone, barchini, Lampedusa, scafisti, scafista, profughi, profugo, Carola Rackete, africani, Africa, africano, islamici, tunisini, libici, Tunisia, Libia, musulmani, ius soli, naufragio, naufragi, diritto di asilo, guardia costiera, trafficante, trafficanti, sbarcare, sbarco, sbarchi, nomadi, rom, porti, permessi di soggiorno, centri di accoglienza, centro di accoglienza, blocco navale. lan Kurdi, Open Arms, Gregoretti, Ocean Viking, Sea Watch, Moby Zaza, Augusta, Mare Jonio, ricollocamenti, respingimenti, Mimmo Lucano, cpr, Porto Empedocle, rotta balcanica, Ventimiglia, Mondragone, Don Biancalani, Pozzallo, Treviso (luglio 2020), Musumeci (agosto 2020), Celio (agosto 2020). English translation of the immigration keywords; immigration, migrants, migrant, immigrant, immigrants, security decrees, security decree, regularisation, amnesty, NGOs, illegal immigrant, illegal immigrant, irregular immigrants, barges, barge, barges, Lampedusa, smugglers, smuggler, refugees, refugee, Carola Rackete, Africans, Africa, African, islamics, tunisians, libyans, tunisia, Libya, muslims, ius soli, shipwreck, shipwrecks, right of asylum, coast guard, smugglers, traffickers, disembarkation, landings, nomads, rom, ports, residence permits, reception centres, reception centre, naval blockade lan Kurdi, Open Arms, Gregoretti, Ocean Viking, Sea Watch, Moby Zaza, Augusta, Ionian Sea, relocations, rejections, Mimmo Lucano, cpr, Porto Empedocle, Balkan route, Ventimiglia, Mondragone, Don Biancalani, Pozzallo, Treviso (July 2020), Musumeci (August 2020), Celio (August 2020).

• 2. Risk and threat dictionary construction

To construct the risk and threat dictionary, we used a list of terms related and synonymous to risk in the Italian language such as “rischio, pericolo, incertezza, insicurezza, minaccia, imprudenza, danno, azzardo” as well as a list related to risk and immigration in Italian media and on social media platforms in Italy. For this latter step, we considered prior content-analytical findings, regarding commonly used words that portray immigration as a risk to the host society (in this case, Italy), which are mostly related to lexica of danger, insecurity and criminality. For instance, in a computer automated analysis of articles of the fifteen Italian newspapers, the Maltone-Bonnenfant (2011) find that there is a tendency to associate immigration with crime and deviance through terms such as “clandestini” (clandestines), “cotrolli” (controls), “controlli di polizia” (police controls), “insicurezza” (insecurity) . Importantly, they argue that the use of these linguistic registers contributes to the creation of the stereotype of immigration being equal to “danger” which as a consequence: “inevitably fuels distrust and raises even higher barriers between the ‘us’ versus ‘them’” perception (Maltone-Bonnenfant, 2011: 8). The authors further identify several relevant paradigms, such as the one the “immigration as social-danger paradigm” wherein the immigrant is portrayed as an aggressor to the Italian values Maltone, 2011: 10).

Table A2.

Risk and Threat Dictionary.

rischio, pericolo, incertezza, insicurezza, minaccia, imprudenza, danno, azzardo.

incognito, violazione di valori, incivile, corpo estraneo, nostri valori, vittime dell’immigrazione, veri discriminati, romeno, Albanese, problema, allarmante, invasione, controllo, eserciti, orde, “profughi,” “rifugiati,” invasione, irregolari, stranieri, criminali, criminale, delinquenti, delinquente, arresti, arresto, arrestare, clandestino, clandestini, clandestinamente, senza controllo, mercato nero, migrazione non autorizzata, segregazione, traffico di migranti, illegali, illegale, trafficanti, trafficante, reato, denuncia, denunciare, controllo, controlli, controllo di polizia, forze dell’ordine, riconduzione alle frontiera, espulsione, espulsioni, rimpatrio, rimpatriare, insicurezza, insicuro, insicura, rimandare, rispedire, rispediamo, rimandiamo, protezione, sicurezza del paese, sicurezza, misure di sicurezza, sicurezza economica, sorveglianza, sorvegliare, difesa, difendere, salvaguardia, salvaguardare, prudenza, accoglienza, accogliere. English translation risk, danger, uncertainty, insecurity, threat, unsafe, harm, hazard. Incognito, violation of values, uncivilized, foreign body, our values, victims of immigration, truly discriminated against, Romanian, Albanian, problem, alarming, invasion, control, armies, hordes, “refugee,” “refugees”, invasion, irregulars, foreigners, criminals, criminal, delinquent, arrests, arrest, illegal, clandestine, clandestinely, unchecked, black market, unauthorised migration, segregation, trafficking illegal, unlawful, traffickers, trafficking, crime, denounce, control, police control, law enforcement, back to the border, expulsion, deportation, repatriation, repatriate, insecure, insecure, send back, send again, send back, protection, country security, security, security measures, economic security, surveillance, guard, defence, defend, safeguard, safeguard, prudence, welcome, welcoming

This paradigm is particularly recurring in conservative and populist press. However, the authors also note that the metaphor of immigration as danger is also present in mainstream articles, although to a lesser extent, in articles reporting on the arrival of refugees-presenting flows of migration as an “alarming situation,” a “problem” or an “invasion.” Similarly, in a recent social media content-analysis, Milkowska-Samul found numerous occurences of terms that highlight the illegal status of migrants (“illegali,” “irregolari,” “clandestini”) and relevantly note that in this context “… it is not the issues of legality that are the prevailing negative element, but rather the migrants themselves portrayed as dangerous, lazy people, who want to take advantage of the goods and rights that should be reserved to Italians…”. (Milkowska-Samul, 2018: 68). Taking all this into consideration, the dictionary related to risk and threat is formed using general risk terms as well as risk terms related to migration in the following way:

Table A3.

Keywords of Risk Related to Migration in Line with Prior Content Analytical Research.

Categories	Specific keywords, examples:	English version of keywords
Examples of frames portraying immigrants/immigration as a risk.	Examples of Keywords and phrases specific to Italian media narratives and political communication and mediatized events as well as metaphors commonly/historically used in Italy.	Literal translation
Immigration and criminality/threat Immigrants bearing Illegal status and aggressivity	clandestini, irregolari, illegali, trafficanti, reato, denuncia, controlli di polizia, arresti, forze dell’ordine, riconduzione alla frontiera, espulsione, rimpatrio, insicurezza, delinquenti, salvaguardia, difesa, sorveglianza, prudenza, protezione, arresto, migrazione non autorizzata, reato, incognito, misure di sicurezza, controllo, entrare	illegal, irregular, illegal, traffickers, crime, reporting, police checks, arrests, law enforcement, return to the border, deportation, repatriation, insecurity, criminals, protection, defense, surveillance, prudence, protection, arrest, unauthorized migration, crime, incognito, security measures, control, entry
Stereotypes	romeno, albanese	Romanian, Albanian
Dehumanization	rimandare, rispedire	Return, send back (ship back)
Paradigm of immigration as a risk to society and values  +  Metaphors of immigration as a risk  +  Polarization/ divide: “us versus others” immigrants/"the cultural other”	violazione di valori, incivile, corpo estraneo, nostri valori, vittime dell’immigrazione, veri discriminati, cittadini Italiani, “profughi”, “rifugiati”, eserciti di clandestine, disturbo, disordine, degrado, invasione, rispetto, regole, violazione, nostra comunità, nostri cittadini, Italiani, allarmante, problema, segregazione, segregazione, sicurezza economica, economia italiana.	violation of values, uncivilized, foreign body, our values, victims of immigration, real discrimination, Italian citizens, “refugees,” “refugees,” armies of illegals, disturbance, disorder, degradation, invasion, respect, rules, violation, our community, our citizens, Italians, alarming, problem, segregation, economic security, Italian economy.