Portrayal of Women in Right-Wing
Populist Communication
By Alena Eichler
Student ID: 1284678
Master’s thesis in Journalism, Media and Globalisation
Graduate School of Communication
Supervisor: dr. Alyt Damstra
Date: May 29, 2020
2
ABSTRACT
This study investigates the portrayal of women in Tweets among the right-wing populist parties in Austria, Denmark, and Germany in a quantitative content analysis. While previous research suggests a link between anti-feminism and extreme right ideology, this work
examines in how far these concepts apply to right-wing populism. By studying Tweets, it was expected to find values representative of their applied politics and at the same time, being considerate that people increasingly obtain political information online from social media platforms. Contrary to the expectations, women were not portrayed frequently in traditional roles, but the most covered subject was female politicians. Besides, a difference between the portrayal of native and migrant women could be found. The country comparison reflects on the findings connected to the medium and the link of anti-immigration rhetoric and women being depicted in context with a “cultural threat”.
INTRODUCTION
Two political movements that seem at odds with each other have both been rising over the past decades, and both continue to do so. The first one, feminism: Women are still
outnumbered in positions, like political offices or Chief Executive Officers, although
European countries continue to work towards more equality for women. With the analyses of gender pay gaps, adjustments of maternity/paternity leave, and tackling issues of everyday sexism, increasing sensitivity for equal chances and can be observed. Also on the rise, right-wing populism: The average share of votes for populist right parties doubled within the last 50 years (de Vreese, 2017). Since the beginning of the 21st century, right-wing populist parties’ power is increasing in democratic countries as they have been elected into national
parliaments, governing coalitions, or the European Parliament (Köttig & Blum, 2017; Mudde, 2019). Their agenda revolves around anti-elitist and anti-immigration strategies. While there has been research on both movements, the study of their interconnection is scarce.
3
To draw on extreme examples, the right-wing terrorist in Norway who killed dozens of children in a political camp by the social democrats to hit the “elites”, was explicitly
misogynist in his manifesto (Keskinen, 2013; Walton, 2012). He was blaming women for increasing multiculturalism. Furthermore, the assassinators of the right extremist terroristic acts in Toronto in April 2018, Christchurch in March 2019, Halle in October 2019, and Hanau in February 2020 were all found to have felt hatred towards women and feminism (Bongen & Schiele, 2019; Flade & Mascolo, 2020).
Of course, there is a difference between right extremism and right-wing populism. Yet, research shows that there is also a link between right-wing populism and anti-feminism (e.g. Keskinen, 2013; Klammer & Goetz, 2017, Spring & Webster, 2019). Many right-wing populist parties, e.g. Alternative for Germany (AfD), openly support the image of nuclear families and are explicitly anti-feminist in their contents (Gilloz, Hairy, and Flemming, 2017; White, 2014). Klammer and Goetz (2017) even state that anti-feminism is a key element of the extreme right ideology, which is missing in most contemporary definitions.
While most extant research has focused on qualitative studies, like case studies of countries or movements (e.g. Keskinen, 2013; Klammer & Goetz, 2017), or studies about the role of gender in manifestos (e.g. de Lange & Mügge, 2015), quantitative comparative approaches are barely found in the field. This work aims to address two shortcomings of current literature. First, it introduces a quantitative method with a country comparison. Second, it analyzes data from social media as they are undisputed an increasingly big part of the media landscape. Twitter is playing an essential role for political parties to reach out and convey their messages to journalists and citizens. To analyze the stances on women of right-wing populists rather than extremists, the central research question of this thesis is:
How are women portrayed in public communication on Twitter of right-wing populist parties in European countries?
4
Investigating the role of women and stances on gender issues in politics continue to have high societal relevance as awareness is the only way to ignite changes. Despite progress in these areas, women still often enough have to fight for their fundamental rights in a male-oriented society and anti-feminist stances or behaviors are too broadly accepted in high political positions (e.g. Kantor & Twohey, 2019). And with right-wing populism gaining more and more popularity, it is important to scrutinize the emerging political force. THEORETICAL BACKGROUND
Right-wing populism, far right, radical right, and extreme right are terms that are often used in a confusing, interchangeable way. Mudde (2019) draws a line between the
“mainstream right”, that he assigns e.g. conservatives to, and the “far right”, which in his eyes are parties that are “anti-system” (Terminology, para. 6) and what other scientists would also call “anti-elite” or “anti-establishment” (e.g. de Vreese, 2017; Bakker, Schumacher, & Rooduijn, 2020). The “far right” Mudde (2019) divides into two kinds, the “extreme right”, which is “reject[ing] the essence of democracy” and the “radical right”, that “accepts the essence of democracy, but opposes fundamental elements of liberal democracy”
(Terminology, para. 6). As an example of these fundamental elements, he lists minority rights.
Populism is often defined around three distinguishing features: the reference to “the people”, a corrupt elite (in academia also called “anti-elite” or “anti-system”) and out-group exclusion (also appears as “anti-immigration” or xenophobia in the context of right-wing populism) (de Vreese, 2017). Populism itself is neither left nor right but adapts “host ideologies” (Bakker et al., 2020, p. 3). In liberal democracies, populism usually refers to right-wing populism (de Vreese, 2017; Mudde and Kaltwasser, 2017)1. Mudde (2019) alleges a “mainstreaming and normalization of the far right in general, and the populist radical right in particular, in the twenty-first century” (Introduction, para. 1). Regarding the claimed
1Mudde (2019) claims that populism is pro-democracy but anti-liberal democracy, the radical right can be
5
normalization of far right stances, this work intends to look at classic European right-wing populist parties. Within the European Union (EU), Mudde refers to a list of - commonly known right-wing populist - parties even as radical right populist parties2.
The focus on common right-wing populist parties rather than the extreme right has two reasons. First, a common argument of right-wing populists is that extreme right cases are often picked to depict the entire right spectrum. Variations among different parties are being neglected too much, as de Lange and Mügge (2015) also claim in their research about gender and right-wing populism. Second, literature often establishes a connection between anti-feminism and right-wing extremism or salient right-wing politicians (e.g. Blum, 2017; Keskinen, 2013), which is not necessarily applicable to the general stance of right-wing populist parties. Klammer and Goetz (2017) even claim that anti-feminism is a characteristic that needs to be included in the definition of right extremism:
Anti-feminism has a long tradition as a political strategy against gender equality. (…) Frequently, an anti-feminist ideology is mixed with other sexist, homophobic, racist and anti-Semitic ways of thinking, and therefore must be seen as a key element of extreme right ideology. (Klammer & Goetz, 2017, p. 87)
Another notable point is, that right-wing populism, as well as the radical right, often sell gender equality as something that has almost or already been achieved in Western
democracies, which enables belittling gender discrimination and concealing power structures that cause gender inequality (White, 2014; Ylä-Anttila and Luhtakallio, 2017). This study draws on these concepts and literature about ties between anti-feminism and the far right to observe and analyze to what extent these are present in everyday communication by common right-wing populist parties.
2He lists the following parties as examples for the radical right: Fidesz – Hungarian Civic Alliance, Law and
Justice (PiS) in Poland, the Patriotic Front, a former electoral alliance in Bulgaria, the Conservative People’s Party of Estonia (EKRE), the Northern League (LN) in Italy, the Slovak National Party (SNS), the Danish People’s Party (DF) in Denmark, and the Democratic Unionist Party (DUP) in the UK (Introduction, para. 1)
6
Far right groups are said to hold more traditional gender norms (Mudde, 2019), which is not per se anti-feminist. Women being more likely depicted in traditional roles is
presumably a general tendency that can be found. Yet, literature states that in right-wing populist parties usually the importance of women (meaning native women) is stressed because they view the traditional family as the core of society, but gender issues or women’s rights are often only addressed in the context of family issues or policies (de Lange & Mügge, 2015; Spierings, Zaslove, Mügge, & de Lange, 2015; Ylä-Anttila & Luhtakallio, 2017). Some work, however, suggests that the rhetoric of right-wing populist parties is anti-feminist. A study in Finland found an overlap in the rhetoric created by radical right-wing populists and
intellectuals in the anti-immigration movement, and the rhetoric in the antifeminist men’s rights movement (Keskinen, 2013). Similarly, in France, Germany, and the UK, right-wing populist parties were found to over-proportionally choose women to target in their social media posts, and the contents were often found to be of anti-female, violent nature (Spring & Webster, 2019). However, empirical evidence is fragmented in this area and more systematic research is needed.
Research on feminism barely provides a concept or a definition of what anti-feminism comprises and it is intricate as anti-anti-feminism is a precluding definition describing everything opposed to feminism (“anti-feminist,” n.d.). That makes it necessary to have a working definition of feminism for this work. Due to different waves of feminism, there are different beliefs and goals (e.g. Allen & LeVine, 2018; Grady 2018). Nevertheless, all of them build on the roots of feminism, referring to equal rights and the participation of women, aiming for an “end of ‘advancement’: ‘rights equal to those granted men.’” And criticism of “the standard of male adulthood as the norm” (Offen, 1988, p. 123). Later, feminism covered areas like pay, sex, reproduction, and abortion, then an increasing understanding of gender and sexuality not being binary but fluid categories (Allen & LeVine, 2018). Contemporary feminism is characterized by its diversity. It is based on the freedom to choose, which can
7
also mean that a woman can adapt traditional female roles, looks, or even oppose to values of feminism. It has also been extended by a characteristic of online activism, and a new calling-out nature of sexism on social media (Chamberlain, 2016; Munro, 2013).
Given the complexity of the concept, the working definition is rather broad and considers everything that opposes equal rights and limits the freedom of choice of women as anti-feminist.
Far right politics follow an ideology of hetero normativity, in which femininity and masculinity are defined categories (Köttig & Blum, 2017), which is contrary to the idea of gender and sexuality being fluid (Allen & LeVine, 2018). In an analysis of the magazine of the right-wing populist Finn’s Party, the researchers also found that the party stresses women as important but societal gender equality issues are mainly being ignored (Ylä-Anttila & Luhtakallio, 2017). Womanhood is usually equated with motherhood and females are often portrayed in their traditional roles as caregivers. Moreover, an example story about the founding of the Finn’s Party men’s organization shows a man washing a baby, depicting the man as a worker and claiming in the text that equality has “gone too far” and that it “is needed to ensure that men are not discriminated against and women favoured ‘in the name of
equality’.” (p. 43).
Based on this, traditional gender roles are expected to play a dominant role in the portrayal of women in this research. Even de Lange and Mügge (2015) who criticized in their study that feminist scholarship labels right-wing populist parties too generalizing, concluded that most parties touch on gender issues only sporadically in their manifestos and the
exception addresses family policies rather than labor market or social policies or ethical issues. However, as systematic data on the communication of right-wing populist parties is largely lacking when it comes to the portrayal of women, this study takes op this task and
8
explores how these parties portray women in their online communications. The first, exploratory research question thus reads:
RQ1 - Are there any general patterns that can be identified when studying the
communication by right-wing populist parties on Twitter addressing (all types/groups of) women?
There is some empirical evidence that points to the existence of structural differences in the way specific groups of women are addressed. For example, female politicians may be approached in a particularly violent way. In an investigation by the BBC, journalists found that right-wing populists disproportionally target female politicians compared to their male counterparts (Spring & Webster, 2019). Comparing Germany, France, and the UK, they found that women in politics face threatening, misogynist content and receive comments based on their gender, race, and physical appearance. As an example, AfD mentioned German Green politician Katharina Schulze ten times as often as any other German political figure within a year.
Furthermore, right-wing populist parties are known for their anti-immigration stances and policies. In the context of gender and equality issues, it is common rhetoric of the right spectrum to portray migrants, especially Muslims, as a threat to Western values of (allegedly already achieved) equality and liberality (Keskinen 2013; Vieten 2016). A comparative study of manifestos found that most right-wing populist parties only touch on gender issues
sporadically (de Lange & Mügge, 2015). However, right-wing populists and the far right do adapt gender and equality discourses when it comes to immigration, particularly from Islamic countries (e.g. Blum, 2017; de Lange & Mügge, 2015), portraying Muslims as “perpetrators” (Vieten, 2016, p. 622). The rhetoric of right-wing populists can even overlap with feminist rhetoric when it comes to Muslim women, claiming that women are not treated equally in
9
Islamic culture (de Lange & Mügge, 2015). A study interviewing Dutch-Moroccan women living in the Netherlands also identifies that gender and gendered culturalism play a
dominant role when it comes to Muslim immigrants (Vieten, 2016). While only immigrant men are seen as a threat and Muslim men are being stigmatized as violent (Blum 2017; Vieten, 2016), Muslim women are either being presented as less emancipated than native women (Blum 2017) or neglected and the right rhetoric primarily worries about (Muslim’s) “sexual violence against ‘native’ women in the public space” (Vieten, 2016, p. 622).
While RQ1 examines the general patterns, the second aim is to map possible differences between targeted groups. Reviewing the literature, theoretical ground leads to expect most differences between the portrayal of female politicians, native women, and women with a migrant background. However, as empirical research is scattered and scarce, no directional hypotheses but a second research question are formulated:
RQ2 - To what extent can structural differences in the portrayal of female politicians, native
women, and women with a migrant background be observed in the communication of right-wing populist parties on Twitter?
For this study, Tweets of the Freedom Party of Austria (FPÖ), the Danish People’s Party (DF), and the Alternative for Germany (AfD) were analyzed. These parties are relevant because they are influential. FPÖ has been part of the governing coalition in Austria until it collapsed due to the Ibiza affair in May 2019. DF in Denmark had been supporting the past two minority governments in Denmark and made headlines throughout Europe for wanting a wall towards Germany. AfD in Germany, in turn, is Europe-wide known since Germany is one of the biggest countries in the EU, and its economic power and central location influence many other countries in Europe. AfD has the third biggest representation in the German parliament, however, all other parties have systematically refused any kind of coalition with the right-wing populist party.
10
While Mudde (2019) claims that although “almost all far right groups hold more traditional gender norms within their own national cultural context”, many norms of traditional values of right-wing populists in Northern European countries “would be
considered progressive in many Southern countries” if not even conflicting (p. 157). In line with this, the Gender Equality Index has identified Denmark as the most gender-equal society in the EU, next to Sweden (“European Union: Index…”, 2019). Also, the share of women of DF politicians in the Danish parliament (Folketinget) is the highest among the three addressed right-wing populist parties in this paper (“Mandatfordelingen”, n.d.). While DF is with 37.5 percent only slightly underneath the general share of women in the Danish parliament (39.1 percent), AfD and FPÖ have with respectively 11.0 and 16.7 percent lower shares of women in their parliaments (“Tal og fakta…”, n.d; “Frauen und Männer”, 2019; “Frauenanteil im Nationalrat”, n.d.). While the Austrian parliament’s (Nationalrat) general share of women is almost equal with the one in Denmark, the German parliament has with 31.2 percent generally a lower representation of women. Albeit, it dropped by 5.9 percent with the last elections when AfD became the third biggest party in the parliament for the first time. The
development is similar in other countries, for example in Finland (Ylä-Anttila & Luhtakallio, 2017).
The three parties also differ in times and circumstances they have been founded. While FPÖ, being founded in the fifties is an established party in Austria, AfD has only been
founded in 2013 and gained a lot of international attention ever since. DF is particularly interesting to look at again, in the context of women, as it has been founded in 1995 by, among others, Pia Kjærsgaard, who has been able to promote her image as a modern but traditional woman. She has also been a chairperson of DF for almost 20 years (Spierings et al., 2015). That leads to the third and final research question that focuses on the existence of cross-country differences in the portrayal of women:
11
RQ3 - To what extent can any structural differences in the portrayal of women, as referred to in the previous RQs, across Austria, Denmark, and Germany be observed?
Guided by these research questions, this work seeks to answer the overarching question of how different categories of women are portrayed in the communications of FPÖ, DF, and AfD on social media (Twitter). The focus on social media is motivated by the following considerations. First, as an ever-growing segment of the media landscape, social media is increasingly shaping society. Numbers show that most people use social media for news consumption, which will further increase with younger, digital-native generations growing older (e.g. European Federation of Journalists [EFJ], 2017; Gottfried, 2014; Statista, 2019). Second, social media allow political actors to reach and communicate with audiences directly and hence, outgo the journalist as a gatekeeper. In that way, political actors can spread their contents cheap, easy and unfiltered throughout society (Jacobs & Spierings, 2019; Meier, 2013), which also makes the internal validity of this work high as it measures direct output of the political parties (David & Sutton, 2011). Third, Twitter is said to shape the new politics (e.g. Hinsliff, 2016). While Facebook has a larger base of users, Twitter users are found to be more politically attentive (Mellon & Prosser, 2017). Politicians and parties use Twitter to position themselves in real-time to current events and their communication is often picked up or cited by traditional news sources, which increases their reach and potential influence (Hinsliff, 2016; Jacobs & Spierings, 2019).
METHODS
To answer the research questions, a quantitative content analysis was conducted to identify and compare patterns in the portrayal of women in Tweets by the right-wing populist parties of Austria (FPÖ), Denmark (DF), and Germany (AfD) (Reinard, 2006). The Tweets for the analysis were collected in a computer-assisted way using the programming language Python. Through the authorization of a Twitter Developer account, the most recent 3,200
12
public, accessible posts of an account can be scraped. As the account of the national FPÖ did not reach this number by far, their most active regional accounts were included, which are the official FPÖ accounts of the Austrian state Steiermark3 and the district of Donaustadt (Danube city) in Vienna. For DF and AfD, the official national accounts were analyzed. These
differences should not compromise the validity of the analysis since the regional accounts also generate official FPÖ content and represent the same ideological view.
The collected Tweets were filtered by a selection of keywords and word stems revolving around women (Appendix 1), which became a challenge regarding the German word “sie”. “Sie” can mean “she” but also “you” in the polite form, or “it”, depending on the gender/article. Therefore, the number of filtered Tweets in German was higher than the Danish Tweets as in the Danish language this is less of a problem. The words used for the search terms took different use of words in German and Danish into account. As the author of this work has sufficient knowledge in Danish but is not a native speaker, the Danish keywords were also discussed and reviewed by a Danish native.
All Tweets were sorted out manually and if the Tweet and (if present) the first link did not contain any portrayal of women; the Tweets were dropped from the analysis. Links were usually either on Twitter embedded pictures, videos, or links to external pages like
newspapers, blogs, or the party’s own website. Since it is an active decision to share content, the portrayal of women in these external sources was also coded if the Tweet did not
explicitly disagree with that aspect.
The final codebook (Appendix 2) was obtained by a mixed approach. The first
codebook, used for the test-coding, was developed deductively. After the test-coding process, which consisted of coding 10 percent of the sample, proportionally to each account, the structure of the codebook remained the same, however, some categories were added or further
13
differentiated. Generally, data like the individual Tweet ID given by Twitter, the account the Tweet was from, and the distinction between the Tweet being an “original Tweet”, a “re-Tweet”, or “a response” to someone, were collected (Q1-3). If there were several
responses/comments of one account on the same thread, and the comments all dealt with the same woman/topic, it was only coded once, and the info of the Tweets was used summarized to fill in the variables of the codebook. The Tweet ID was also used to look up Tweets in its original context if there was unclarity. However, FPÖ Steiermark deleted all its past Tweets a couple of days after the scraping, so looking at the context for these Tweets was not possible during the coding process and they were coded to the best of knowledge.
Based on the investigation about female politicians being disproportionally targeted by right-wing populists (Spring & Webster, 2019), the Tweets were coded whether a “private individual”, a “(group of) women in general”, or a “female politician” was/were covered (Q4). If the woman referred to, was a professional in another field, for example, journalist or activist, or a public figure like the Queen of Denmark, it was coded as “Other”. After the test-coding, the “female politician” category was split up further in “female politician of own party” and “female politician of another party”. Also, because some women seemed to be re-appearing frequently, a follow-up question captured name and party (Q4b). If the woman was a public figure, a follow-up question asked, what topic was mainly covered in the Tweet: “her work/function”, “individual characteristics/personality”, “private life”, “looks”, “other”, or “none” (Q4a).
In Q5, it was noted whether the woman in the Tweet was the “main focus” of the post or a “side note”. If the Tweet covered several women, for example, the party was posting a response Tweet, discussing with a female politician of another party about a female activist, the politician and the activist were coded as two different cases with the same Tweet ID. If women were just mentioned in a side note, maximum two women per Tweet were coded as
14
cases: Either the two most dominant side notes or in videos (e.g. election speeches) longer than 30 minutes and if the main topic was not a woman or women, the two first mentioned. If a female politician of the own party was simply quoted, or her contents were shared without an external link, she was not coded. However, if the Tweet contained a link to the party’s website and the politician of the own party was portrayed with a picture and her stance to a current situation, it was coded.
Furthermore, inspired by the codebook of a study on visual advertisement
(Holtzhausen, 2010), the Tweets were checked for the women being portrayed in any of the following roles: “mother”, “housewife”, “physically decorative/sex object”, or “in her
professional function” (Q6). Additionally, it was extended by a “cultural threat”, and after the test-coding, by a “victim(s) of cultural threat” category, based on literature stating that
Muslims tend to be portrayed as a cultural threat to Western values in right-wing populist’s contents (de Lange & Mügge, 2015; Vieten, 2016). If none of the roles applied, the possibility was provided to code the portrayal as “other” and explain how the woman/women were portrayed, or “none of the above”.
To find possible differences in the portrayal between native and migrant women, the next question asked whether the woman was “explicitly portrayed as [a] migrant
woman/women”, “explicitly portrayed as [a] native woman/women”, “not made explicit but subject(s) is/are known to be [a] migrant woman/women”, “not made explicit but subject(s) is/are known to be [a] native woman/women”, “not detectable”, or “other” (Q7). So either if the woman was portrayed in a picture and looked very German, but also if a politician (e.g. chancellor Angela Merkel in Germany) was known to be native, it was coded as known to be a native. If the post covered a foreign woman who is not a migrant (e.g. the Swedish activist Greta Thunberg), it was coded as “other”. Angela Merkel, who appeared repeatedly in Tweets of all parties, was only in AfD Tweets coded as known to be native while in DF and FPÖ
15
Tweets she was coded as “other” since “native” in this study referred to the respective own country. If the country or region migrants were from was mentioned, it was also noted (Q7a). Additionally, it was captured which posts contained references to immigration in general (Q8).
Lastly, all posts were rated on a scale from 0 to 4 (0 = not at all, 1 = rather not, 2 = slightly, 3 = moderately, 4 = very) whether it was insulting, attacking, or threatening the portrayed women in any way (Q9). If a woman was being criticized in her function, and it was not insulting her as a person, it was coded as 1. If it was also insulting to her as a person, it was coded as 2. If for example, Muslim women were called to drop their veils, depending on how harshly it was worded, it was coded as 3 or 4. When party members of AfD screamed in a video in a speech that they will hunt Angela Merkel, it was a 4. Additionally, general notes for saliences concerning country context or other aspects were made.
After the test-coding, the adjustments mentioned above were made in order to make the codebook compatible with reality and ensure its validity. Moreover, the first and last names of each woman that appeared in the Tweets during the test-coding, were added as search terms to the list (Appendix 1), as several women kept re-appearing. This change made the number of Tweets rise to a total of 1,545 (970 AfD, 303 FPÖ, and DF 272), while AfD and FPÖ had the highest “sorting out” rates when manually dropping Tweets that did not cover women, due to the “sie” issue.
Before starting with the content analysis of the actual dataset, the intercoder reliability needed to be established. To do so, the (German) author of the thesis and the Danish coder both coded a subset of the sample, using 10 percent of the Danish tweets. After an
independent read of the codebook, additional instructions and examples were given. The intercoder reliability test using Krippendorff’s Alpha generated results in between .92 and 1 (Appendix 3), which speaks for high reliability of the codebook. Finally, the Danish coder
16
moved on to coding the total of Danish Tweets while the author coded the total of the Austrian and German Tweets.
RESULTS
After removing the Tweets if the content was not about a woman or women, there was a total of 842 coded cases for analysis (some Tweets were coded more than once due to several women being portrayed). In the scraped Tweets of AfD, 484 coded cases were found, in the Tweets of DF 189, and the Tweets of FPÖ 169 (Appendix 4, Table 1), despite FPÖ having the most Tweets prior filtering, to begin with. For both, AfD and DF, the most recent 3,200 Tweets that were possible to scrape were filtered for the search terms. For FPÖ, the 74 Tweets that were communicated by their national account, 860 stemming from their state account, plus 3,200 that came from their regional account in Vienna (altogether 4,134) were used. For FPÖ and DF, the recent 3,200 Tweets reached back until 2016 while for AfD, they only reached back until January 2019.
General patterns
Results showed that almost half of the Tweets (49.2 percent) referred to “Female politician[s] of another party” (Q4), which made them the most covered female category (Table 2). The second most covered subjects were politicians of the own party, however with a clear drop, appearing in 14.1 percent of the Tweets. German chancellor Angela Merkel was tweeted about the most as “female politician of another party”, followed by Mette
Frederiksen, the prime minister of Denmark (Table 3). Angela Merkel was mentioned in 161 Tweets (147 of AfD) and Mette Frederiksen in 77 (exclusively by DF). Both were often discussed as representatives of the general government. Especially Angela Merkel’s name was often also used for word plays, for example, “vermerkelt” in AfD Tweets, meaning politics became too much the way Merkel wanted, or “merkelig politik” in a DF Tweet, referencing the word “mærkelig”, meaning weird. When referring to female politicians of the
17
own party, Beatrix von Storch and Alice Weidel (both AfD) were most mentioned (37 and 29 times).
Female politicians, from the own as well as from other parties, were consistently covered in the context of their work or professional function (Q4a), in 99.2 and 93.4 percent of the cases (Table 4). Also, 75.2 percent of the women coded as “other” were portrayed in their work/function; “other” referring to journalists, activists, experts, authors, and the Danish queen.
Going together with female politicians being most covered, the role (Q6) found most was professional function (75.4 percent). Then, followed by the “victim of cultural threat” role (10.8 percent) and “none of the above/other” (6.4 percent) (Table 5). “Housewife” as well as “physically decorative/sex object” was seldom applicable. The factor, rating how insulting the tone was, showed significance between the groups [F(6, 833) = 5.26, p < .001] (Table 6). The role that showed by far the highest result was the “cultural threat” category (M = 2.00, SD = 1.02), which women were not portrayed in frequently (3.2 percent of the Tweets).
Most women portrayed by the parties were known to be native, although this was not mentioned explicitly (69.5 percent) (Q7) (Table 7). In 12.2 percent of the (non-)migration background was not detectable and in 7.6 percent of the cases it was “Other”, which usually meant that the Tweet referred to someone foreign not living in the country (for example Greta Thunberg, or Angela Merkel in the Austrian or Danish Tweets). Further, 5.8 percent were portrayed explicitly as migrant women, and 3.6 percent were known to be migrants. Women were rarely explicitly referred to as native (1.1 percent).
With regard to the general tone of the content, the degree to which the content of the Tweets was characterized as insulting or attacking was examined, using a 5-point scale
18
running from 0 to 4.4 While the tone was not very insulting in general (M = 1.01, SD = .99), it was significantly higher [F(4, 835) = 56.90, p < .001] for “female politician[s] of another party” (M = 1.30, SD = .78) and “(group[s] of) women in general” (M = 1.30, SD = 1.38) (Table 8).
Based on these exploratory analyses, the following answer to the first research question can be formulated. General patterns that were found show that the women most referred to by right-wing populist parties were politicians from parties other than the own. They were usually portrayed by references to their professional function. (Groups of) women in general and private individuals were covered less frequently and if they were, both were often found in the “victim(s) of cultural threat” role while women in general also tended to be portrayed as “mother” and individuals as “cultural threat”. Most attention was given to women who are known to be native, as women with a migration background are much less prominent. Generally, and different from what was expected based on theory, references to traditional gender roles were less dominating than references to professional function.
Different portrayals of different groups of women: female politicians, native women, and women with a migrant background
To explore whether and to what degree female politicians, native women, and women with a migrant background were portrayed differently, the presence of different roles (Q6) across these categories was first examined (Table 9). The analysis showed that there was a significant association between type of category (Q4) and role portrayal (Q6) [X² (24, N =842) = 728.2, p < .001] being moderate in strength (Cramer’s V = .465). Female politicians (of the own as well as from other parties) have almost exclusively been portrayed in their professional function (98.8 and 95.8 percent). It must be noted that female politicians of the
4 The scale was labeled: 0 = not at all, 1 = rather not, 2 = slightly, 3 = moderately, 4 = very. The variable is
19
own party and another party were usually known to be native (96.6 percent and 88.6 percent) (Table 7). Comparing the factor again, whether the Tweets were insulting, the Tweets about “female politician[s] from another party” (M = 1.30, SD = .78) differed significantly from almost all categories (Table 8, see above). Therefore, Tweets about a “female politician of [the] own party” were, unsurprisingly, not insulting at all with the lowest score of all categories (M = .02, SD = .13).
Next, the portrayal of native women was examined. Out of all (known to be) native women, 92.6 percent were portrayed in the role of their work/function, which is connected to 82.4 percent of native women portrayed being female politicians, about three-quarters from another party, and one-quarter of the own party (Table 10). On the scale measuring how insulting the Tweet was, the native category (M = .91, SD = .84) showed significant
differences compared to the migrant group (p < .001). The explicitly native group (M = 1.00, SD = 1.25) did not show significance to any other group.
Out of all Tweets about women, 5.8 percent were including portrayals of explicitly referred to as migrant women, and 3.6 percent were known to be migrants (Table 7). The relationship to both, the type of women portrayed (Q4) and the role (Q6) turned out to be significant. Migration background (Q7) tested significant with the former (Q4) with a moderate association [X² (20, N = 842) = 568.2, p < .001, Cramer’s V = .411], and the latter (Q6) as well [X² (30, N = 842) = 681.6, p < .001, Cramer’s V = .402]. Out of the women explicitly portrayed as migrants, 65.2 percent were merely referred to as Muslim, and 16.2 percent were from African countries that were named individually. Adding the two migration background groups together, 41.8 percent of them were “(group[s] of) women in general”, 22.8 percent “private individual[s]” and 19.0 percent “female politician[s] of another party” (Table 12). When explicitly migrants, they were in 51.0 percent of the cases portrayed as “victim[s] of cultural threat” and in 40.8 percent of the cases as “cultural threat” (Table 10).
20
When they were known to be migrants, in 76.6 percent they were portrayed in their
professional function, which is related to two people being subjects in the Tweets. First, AfD had targeted politician Sawsan Chebli – a Social Democrat – several times, who is a Muslim but has also been subject of public debates by others than right-wing populist parties about some controversial Tweets and her (non-)professional behavior. Second, the regional FPÖ in Vienna criticized a local politician of the Green party several times, who has Greek heritage. The factor measuring whether the Tweets were insulting also showed high significance between the groups [F(4,835) = 13.98, p < .001] (Table 11). The explicitly migrant category showed the highest mean (M = 1.83, SD = 1.19), followed by the “other” category (M = 1.62, SD = 1.25). The known to be migrant category (M = 1.10, SD = 1.13) already showed a significantly lower mean.
In summary, female politicians are predominantly portrayed in their professional functions and were usually known to be native. Not surprisingly, Tweets about “female politician[s] of another party” were significantly more insulting (p < .001) than the Tweets about “female politician[s] of own party”, even though the general mean was rather low. As native women were the norm, not many references were made to their native backgrounds and they were consistently portrayed in terms of their professional functions. Women with a migrant background were often Muslim and portrayed by references to “victim(s) of cultural threat” or being a “cultural threat” themselves. Explicitly migrant women received the Tweets rated as most insulting in this study.
Country comparison
Finally, were the possible country-level variations of interest. In all countries, most Tweets referred to female politicians from other parties than the own. The representation was highest for the Danish party DF (54.5 percent), followed by the German AfD (50.0 percent), and the Austrian FPÖ (40.8 percent) (Table 2). For DF, the second most covered type of
21
woman was (groups of) women in general (17.5 percent), for AfD, female politicians of the own party (16.7 percent), and FPÖ the “other” category (21.3 percent). Most of the “other” category consisted of journalists, authors, and experts including political scientists,
virologists, and doctors (55.6 percent). Angela Merkel was frequently covered by all three parties. AfD and FPÖ tweeted about Angela Merkel the most while for DF, it was Mette Frederiksen, head of state in Denmark, however, Angela Merkel came second.
Concerning the professional function, DF most often portrayed females in their work/function (98.1 – 100.0 percent) (Table 13-15). AfD scored lowest on portraying women in their work/function (61.7 – 100.0 percent). Especially in the “other” type of woman
category, AfD had the highest percentage of “individual characteristics/personality” (14.9 percent), compared to the other parties.
Divided by country, the two most found roles remained the same (Table 5). The
“victim of cultural threat” role was most frequently found among the Danish Tweets with 12.7 percent, followed third by the role of a “cultural threat” with 6.3 percent. The two other
countries did not portray women as a “cultural threat” as often. Hence, DF portrayed almost every fifth woman in context with a cultural threat.
While native women dominated in the communications by all parties (Table 16-18), there were a few country-level differences. For AfD and DF, it was almost three out of four women (73.8 and 73.0 percent), it was about every second for FPÖ (53.3 percent). For the remaining Tweets, was DF by far the highest with 11.8 percent of portraying women explicitly as migrants. Contrary, for AfD and FPÖ it was both “not detectable”, that came second with 21.9 and 11.0 percent and both only had 4.1 percent in portraying women as explicitly migrant. DF also scored highest when it came to whether the Tweet contained general references to immigration (Table 19). For DF, it was more than half of their Tweets (52.5 percent) while for AfD 39.4 and FPÖ 29.6 percent. The relationship between country
22
and immigration reference showed significance with a weak association [X² (2, N = 840) = 19.7, p < .001, Cramer’s V = .153].
DF was also more likely to portray explicitly migrant women (11.6 percent) (Table 17). When women were explicitly migrant, the roles fit again with the “(group of) women in general” and the “private individual” categories being the most represented type of woman. But when they were known to be migrants, they were most likely to be politicians of another party in Austria and Germany while Denmark barely covered women who were known to be migrants (Table 16-18).
The relationship between the insulting-ness scale and the country variable did not turn out to be significant, and the means for Tweets of AfD (M = 1.02, SD = .99), DF (M = 1.03, SD = 1.04), and FPÖ (M = .94, SD = .97) were about the same.
Wrapping up, the most notable country differences were that Germany portrayed their own female politicians more frequently than the other countries. Denmark turned out to be blunter at portraying migrant women explicitly and mentioning them in the context of
“cultural threat”. Therefore, Denmark scored highest on covering a female politician or other professionals in their professional function. The expectation that Denmark could be
performing better, as it is a Northern European country, seems to only apply native women, not for migrants.
CONCLUSION & DISCUSSION
Based on the findings, the research questions can be answered as follows. General patterns showed that Tweets by right-wing populist parties mainly covered female politicians. Women in general and private individuals were not as much covered but if they were, they were often found in the victim role of a cultural threat. A frequent portrayal in traditional gender roles could not be confirmed. The women covered were usually native. Female
23
politicians were portrayed in their professional functions. Women with a migration background were often referred to in context with cultural threat and explicitly migrant women received the most insulting Tweets. In the country comparison, the Danish party showed the biggest variations with being sharper towards Muslim women and are even more likely to portray native women in their professional role than AfD and FPÖ.
Generally, it needs to be pointed out that Tweets by FPÖ and DF made references to Germany regularly, FPÖ even making supportive Tweets regarding AfD criticizing German politics while AfD barely referred to DF or FPÖ. It demonstrates Germany’s central, geo-economically strong position in the EU, with the smaller neighboring countries orientating themselves towards Germany and being affected by Angela Merkel and German politics on an EU level.
The German chancellor and the Danish prime minister alone make up more than a fourth of all coded cases. For the German and the Danish right-wing populist party, both being the third most-voted party in the last national elections, both in the opposition, it makes sense to criticize leaders of the current governments. Often, their names were also used representatively for the governing cabinet. The Tweets reached from wishing Angela Merkel to get better when she was suspected to have Covid-19, to referring to left violent
demonstrates as “Merkeljugend”, referencing “Hitlerjugend”, to accusing Mette Frederiksen of letting IS fighters that cut off heads into the country.
The result of women being portrayed as “victim(s) of a cultural threat” as the second most present portrayal is in line with the literature about anti-immigration rhetoric as
suggested by de Lange and Mügge (2015) or Vieten (2016). Especially when it comes to the portrayal of groups of women and private individuals, the findings resonate with existing literature. However, the idea that women are often portrayed in more traditional roles as suggested by Mudde (2019) is not confirmed by the findings, except for a few references in
24
which groups of women are portrayed as mothers. That would match with studies of
manifestos and party contents of right-wing populist parties defining families as the core of society (e.g. de Lange & Mügge, 2015; Ylä-Anttila & Luhtakallio, 2017), but the data in this study did not reflect that. This could be connected to data being collected on Twitter as the platform’s users are younger, more educated, and slightly more liberal than the average population and the contents might be influenced by the audience (Mellon & Prosser, 2017). However, to make more precise propositions, a further study only investigating the portrayal of private individuals or (groups of) women, could find better answers. The fact that women were barely portrayed as housewives or physically decorative online, could also be connected to the calling-out nature of sexism on the internet, which was pointed out as characteristic of contemporary feminism (Chamberlain, 2017).
Furthermore, the fact that the Danish party portrayed women more consistently in terms of their professional functioning, can be explained by work that suggested how Northern European countries are more progressed on traditional values and have a stronger societally entangled state feminism (Borchorst & Siim, 2008; Mudde, 2019). DF’s bluntness towards Muslims and the tendency to portray women more often in context with a cultural threat, or refer to them as explicitly Muslim might be connected to the past two governments being center-right (Crouch & Eriksen, 2015). Anti-immigration stances might, therefore, be more common and accepted in public debates and Denmark is also known to be more anti-immigration than other EU countries (“Statistics: More migrants…”, 2020).
One of the Tweets showed a video of a speech of AfD politician Mariana Harder-Kühnel, in which she refers to equality in a way that stood out. She referred to equality in the constitution saying, “[b]ut by this I mean real equality, in the sense of equal opportunity between men and women, not parity”. This phrase fits particularly well into a general
25
people to be artificial and negative, which should be overcome by an active state, whereas the right believes that inequalities between people are natural and positive, and should be either defended or left alone by the state.” (Terminology, para. 4). This is summing up and
explaining well, how right-wing populists approach gender issues and the equality of men and women.
Limitations of this study are the missing comparison to male politicians and Muslims. Also, the Austrian Tweets were more often coded as “other”, which could point to differences caused by regional accounts. Another aspect that could be further researched is how
individual characteristics affect stances on equality and how they overlap with political leanings. For example, a study by Bakker et al. (2020) shows that people who score low on the personality trait “agreeableness” are more susceptible to anti-establishment messages expressed by populists. Researching further in the field of feminism in connection with media and politics is important as the job is not done. Regarding the corona pandemic, concerned experts even address that feminism might be thrown back decades (Koch, 2020). When the author got to ask Megan Twohey, one of the journalists who broke the Weinstein sexual harassment story, about connections between (anti-)feminism and right-wing populists, she instantly started talking about allegations of US President Donald Trump. Research on feminism and gender in media and politics generate more knowledge and awareness, and as long as such people can be in powerful political positions, it shows how much awareness is still needed and how broadly accepted anti-feminist stances still are.
26
REFERENCE LIST
Allen, M., & LeVine, S. (2018, May 29). A global women's liberation movement. Retrieved from https://www.axios.com/women-equality-movement-politics-tech-industry-metoo-feminism-46451fa1-80b4-4511-b4ea-94e67e64e90d.html
anti-feminist. (n.d.). Retrieved from https://www.merriam-webster.com/dictionary/anti-feminist
Bakker, B.N., Schumacher, G. & Rooduijn, M. (2020). The Populist Appeal: Personality and Anti-establishment Communication. Paper forthcoming in the Journal of Politics.
Blum, A. (2017). Men in Battle for the Brains: Constructions of Masculinity Within the “Identitary Generation”. In M. Köttig, R. Bitzan, & A. Petö (Eds.), Gender and far right
politics in Europe (pp. 321-334). Cham, Switzerland: Palgrave Macmillan.
Bongen, R., & Schiele, K. (2019, November 1). Rechtsextremismus: Feminismus als Feindbild. Retrieved from https://www.tagesschau.de/investigativ/panorama/frauenhass-rechtsextremismus-101.html
Borchorst, A., & Siim, B. (2008). Woman-friendly policies and state feminism: Theorizing Scandinavian gender equality. Feminist Theory, 9(2), 207–224.
https://doi.org/10.1177/1464700108090411
Chamberlain, P. (2016). Affective temporality: towards a fourth wave. Gender and
Education: “If Not Now, When?”: Feminism, Activism and Social Movements in the European South and Beyond, 28(3), 458–464.
https://doi.org/10.1080/09540253.2016.1169249
Crouch, D., & Eriksen, L. (2015, June 18). Denmark swings to the right as centre-left coalition accepts defeat. Retrieved from
27
https://www.theguardian.com/world/2015/jun/19/denmark-swings-right-centre-left-coalition-faces-defeat
David, M., & Sutton, C. D. (2011). Social research an introduction (2nd ed.). London: Sage.
de Lange, S., & Mügge, L. (2015). Gender and right-wing populism in the Low Countries: ideological variations across parties and time. Patterns of Prejudice, 49(1-2), 61–80. https://doi.org/10.1080/0031322X.2015.1014199
de Vreese, C. H. (2017). Political Journalism in a Populist Age. Shorenstein Center on Media, Politics and Public Policy.
European Federation of Journalists. (2017, April 3). The use of social media for news is growing, says report on digital news. In europeanjournalists.org. Retrieved from https://europeanjournalists.org/blog/2017/04/03/the-use-of-social-media-for-news-is-growing-says-report-on-digital-news/
European Union: Index: 2019: Gender Equality Index. (n.d.). Retrieved from https://eige.europa.eu/gender-equality-index/2019
Flade, F., & Mascolo, G. (2020, February 21). Nach Anschlag in Hanau: Die große Ratlosigkeit der Behörden. Retrieved from https://www.tagesschau.de/inland/hanau-investigativ-101.html
Frauenanteil im Nationalrat. (n.d.). Retrieved May 21, 2020, from
https://www.parlament.gv.at/SERV/STAT/PERSSTAT/FRAUENANTEIL/frauenanteil_N R.shtml
Frauen und Männer. (2019, July). Retrieved May 21, 2020, from
https://www.bundestag.de/abgeordnete/biografien/mdb_zahlen_19/frauen_maenner-529508
28
Gilloz, O., Hairy, N., & Flemming, M. (2017, May 8). Getting to know you: mapping the anti-feminist face of right-wing populism in Europe. Retrieved October 4, 2019, from https://www.opendemocracy.net/en/can-europe-make-it/mapping-anti-feminist-face-of-right-wing-populism-in-europe/
Gottfried, J. (2014, November 12). Facebook and Twitter as political forums: Two different dynamics. Retrieved from https://www.pewresearch.org/fact-tank/2014/11/12/facebook-and-twitter-as-political-forums-two-different-dynamics/
Grady, C. (2018, March 20). The waves of feminism, and why people keep fighting over them, explained. Retrieved from https://www.vox.com/2018/3/20/16955588/feminism-waves-explained-first-second-third-fourth
Hinsliff, G. (2016, July 31). Trash talk: How Twitter is shaping the new politics. Retrieved from https://www.theguardian.com/technology/2016/jul/31/trash-talk-how-twitter-is-shaping-the-new-politics
Holtzhausen, T. (2010, August). Content Analysis of Roles Portrayed by Women in Advertisements in selected South African Media [PDF]. University of Pretoria. Jacobs, K., & Spierings, N. (2019). A populist paradise? Examining populists’ Twitter
adoption and use. Information, Communication & Society, 22(12), 1681–1696. https://doi.org/10.1080/1369118X.2018.1449883
Kantor, J., & Twohey, M. (2019). She Said. London, Great Britain: Bloomsbury Publishing Plc.
Keskinen, S. (2013). ANTIFEMINISM AND WHITE IDENTITY POLITICS: Nordic
Journal of Migration Research, 3(4), 225–232. https://doi.org/10.2478/njmr-2013-0015
Klammer C., & Goetz, J. (2017). Between German Nationalism and Anti-Muslim Racism: Representations of Gender in the Freedom Party of Austria (FPÖ). In M. Köttig, R. Bitzan,
29
& A. Petö (Eds.), Gender and far right politics in Europe (pp. 79-93). Cham, Switzerland: Palgrave Macmillan.
Köttig, M., & Blum, A. (2017). Introduction. In M. Köttig, R. Bitzan, & A. Petö
(Eds.), Gender and far right politics in Europe (pp. 1-10). Cham, Switzerland: Palgrave Macmillan.
Mandatfordelingen. (n.d.). Retrieved from
https://www.ft.dk/da/medlemmer/mandatfordelingen
Meier, K. (2013). Journalistik (3rd ed.). Konstanz and Munich, Germany: UVK Verlagsgesellschaft mbH.
Mellon, J., & Prosser, C. (2017). Twitter and Facebook are not representative of the general population: Political attitudes and demographics of British social media users. Research &
Politics, 4(3). https://doi.org/10.1177/2053168017720008
Mudde, C. (2019). The far right today. Cambridge, UK: Polity Press.
Mudde, C., & Kaltwasser Cristóbal Rovira. (2017). Populism: a very short introduction. New York, NY: Oxford University Press.
Munro, E. (2013). Feminism: A Fourth Wave? Political Insight, 4(2), 22–25. https://doi.org/10.1111/2041-9066.12021
Offen, K. (1988). Defining Feminism: A Comparative Historical Approach. Signs, 14(1), 119–157. https://doi.org/10.1086/494494
Rechtspopulismus. (2017, January 24). Retrieved from
https://www.bpb.de/politik/extremismus/rechtspopulismus/
Reinard, J. C. (2006). Communication research statistics. Thousand Oaks, CA: Sage Publications.
30
Spierings, N., Zaslove, A., Mügge, L., & de Lange, S. (2015). Gender and populist radical-right politics: an introduction. Patterns of Prejudice, 49(1-2), 3–15.
https://doi.org/10.1080/0031322X.2015.1023642
Spring, M., & Webster, L. (2019, July 15). A web of abuse: How the far right disproportionately targets female politicians. Retrieved November 3, 2019, from https://www.bbc.com/news/blogs-trending-48871400
Statista. (2019, September 30). News sources used in European countries in 2019. In Statista
– The Statistics Portal. Retrieved from
https://www.statista.com/statistics/422687/news-sources-in-european-countries/
Statistics: More migrants left Denmark than entered in 2019. (2020, May 04). Retrieved from https://apnews.com/b4f4c00c59aae9da3495a1d109c72ae6
Tal og fakta om Folketinget. (n.d.). Retrieved May 21, 2020, from
https://www.ft.dk/da/folkestyret/folketinget/tal-og-fakta-om-folketinget
Vieten, U. (2016). Far Right Populism and Women: The Normalisation of Gendered Anti-Muslim Racism and Gendered Culturalism in the Netherlands. Journal of Intercultural
Studies, 37(6), 621–636. https://doi.org/10.1080/07256868.2016.1235024
Walton, S. (2012). Anti-feminism and Misogyny in Breivik’s “Manifesto.” NORA - Nordic
Journal of Feminist and Gender Research, 20(1), 4–11.
https://doi.org/10.1080/08038740.2011.650707
White, J. (2014, March 31). Anti-euro party turns anti-feminist. Retrieved from https://www.thelocal.de/20140331/german-anti-euro-party-afd-turns-anti-feminist-alternative-for-germany-facebook
31
Ylä-Anttila, T., & Luhtakallio, E. (2017). Contesting Gender Equality Politics in Finland: The Finns Party Effect. In M. Köttig, R. Bitzan, & A. Petö (Eds.), Gender and far right politics
32
APPENDIX
Appendix 1: Search terms for filtering Tweets
German search term used English translation Danish search term used
Frau woman / Ms. / Mrs. / wife kvinde / dame / kone / fru
/ frøken
weiblich female Femini / kvinde
gender gender
femini feminis femini
Geschlecht gender køn
Mutter mother Mor / moder
Mütter mothers mødre
emanz emanc Emancip / frigørelse /
frigjort
Sie / ihre She / her Hun / hende
Added names after test-coding:
Olga Ilse Dronning
Alice Groß Margrethe
Weidel Angela Angela
Beatrix Merkel Merkel
Storch Margarete Mette
Renate Schramböck Frederiksen
Künast Natascha Jette
Ferda Strobl Skive
Ataman Franziska
Angela Giffey
Merkel Claudia
Nicole Plakolm
Diekmann Eva Maria
Annette Holzleitner
Behnken Barbara
Mariana Neßler
33 Greta Thunberg Sawsan Chebli
34
Appendix 2: Codebook Codebook
Introduction. Go through the Tweets and read them carefully. If the Tweet includes a link, copy and open the link in order to skim it if there are women portrayed in any way. If the Tweet does not mention an individual woman or women at all, drop the Tweet. If yes, code the Tweet according to the following questions in this codebook. Additional links to outlets or external contents do not all have to be coded and analyzed in detail, but they should be
skimmed to decide if they are of relevance and if they are, the information should be used to complete or specify the coding. If the context of the Tweet is unclear (that is often the case when the Tweet is a response, which typically starts with “@username…”), open the original Tweet by copying this link in the URL and insert the respective Twitter ID:
https://twitter.com/anyuser/status/[insert Tweet_ID]
Several Tweets that are responses/comments on the same thread, address the same topic are coded as one Tweet. If a Tweet is about two women (e.g. discussing the content of an activist with a politician), code them as two Tweets.
Q1. What’s the Tweet ID of the Tweet? _________________________
Q2. Which country/party is the Tweet from? 1 = Austria/FPÖ
2 = Denmark/DF 3 = German/AfD
Q2a. If the Tweet is by the FPÖ, which account is it? ____________
Q3. Is the Tweet original content by the party, a Re-Tweet or a response to someone? 1 = Original Tweet
2 = Re-Tweet 3 = A response
Q4. Does the Tweet refer to a private individual, (a group of) women in general or to a female politician - from another party or from the own party? If the woman/women is/are portrayed
35
in a different way or in another professional function, use the “Other” code and describe briefly in which function.
1 = Private individual
2 = (Group of) women in general 3 = Female politician of another party 4 = Female politician of own party 6 = Other: ____________
If Q4 = 3, 4 or 6 is another professional function, continue with Q4a, if not, continue with Q5.
Q4a. If the woman is a public figure, what topic does the Tweet mainly cover? 1 = Her work/function 2 = Individual characteristics/personality 3 = Private life 4 = Looks 6 = Other: ____________ 9 = None
Q4b. If the woman is a politician, who is she and which party is she from? ____________
Q5. Is the woman/Are the women the main focus of the post or just mentioned as a side note? 1 = Main focus
2 = Side note
Q6. In which (if any) of the following roles is the woman mainly being portrayed? (If none is applicable, go for “Other” and explain how.)
1 = Mother 2 = Housewife
36
4 = In her professional function 5 = Cultural threat5
6 = Victim(s) of cultural threat
9 = None of the above/Other: ____________
Q7. Is the woman in the Tweet portrayed as a migrant or native woman and if so, is it explicitly or implicitly? Implicitly also entails if the nationality is not mentioned but an individual is known to be a native/a migrant (e.g. Angela Merkel in Germany). If the post covers a foreign woman who is also not a migrant (e.g. Greta Thunberg), code as “Other”. 1 = Explicitly portrayed as (a) migrant woman/women
2 = Explicitly portrayed as a native woman/women
3 = Not made explicit but subject(s) is/are known to be (a) migrant woman/women 4 = Not made explicit but subject(s) is/are known to be (a) native woman/women 5 = Not detectable
6 = Other
Q7a. If Q7=1 or 3 and the country or region of origin is explicitly mentioned, where is she from? ____________
Q8. Does the content contain references to immigration? 1 = Yes
2 = No
Q9. On a scale from “not at all” to “very” is the content insulting, attacking or threatening a woman or women in any way? (If e.g. a politician’s content or opinion is highly criticized without being insulting, it is 1. If women e.g. are told to drop their veils, depending on how harshly it is worded, go for 3 or 4.)
0 = Not at all 1 = Rather not
5 Clarification to what is considered a “cultural threat”: Is the Tweet questioning the loyalty of the
woman to the host country, to liberal Western values, like feminist values? Does it contain references to (facilitation of) terrorist activity?
37
2 = Slightly 3 = Moderately 4 = Very
Q10. Is there anything else of particular notice in the light of country context? Enter text: ___________________________
Q11. Is there anything else in the Tweet that might be of particular interest? (Pay special attention if a Tweet is about a migrant woman who is a professional or politician.) Enter text: ___________________________
38
Appendix 3: Intercoder Reliability Test Krippendorff's Alpha Reliability Estimate
Variable Measure Alpha LL95%CI UL95%CI Units Observrs Pairs Q3 Nominal ,9244 ,7732 1,0000 27,0000 2,0000 27,0000 Q4 Nominal 1,0000 1,0000 1,0000 22,0000 2,0000 22,0000 Q5 Nominal 1,0000 1,0000 1,0000 22,0000 2,0000 22,0000 Q6 Nominal 1,0000 1,0000 1,0000 22,0000 2,0000 22,0000 Q7 Nominal ,7978 ,4944 1,0000 23,0000 2,0000 23,0000 Q8 Nominal 1,0000 1,0000 1,0000 22,0000 2,0000 22,0000 Q9 Ordinal ,9241 ,8200 ,9930 22,0000 2,0000 22,0000
Note: The Alpha at Q7 was the lowest and found to be rooted in the non-native Danish coder not understanding the complexity of the Danish language in all its facets. All cases that were coded differently in Q7 were discussed and the coders turned out to agree on all of them.
39
Appendix 4: Results
Table 1: Number of Tweets per country (political party)
Frequency Percent
Austria (FPÖ) 169 20.1
Denmark (DF) 189 22.4
Germany (AfD) 484 57.5
Total 842 100.0
Table 2: Type of woman/women portrayed in Tweet (Q4), specified by political party Country
Total
Austria Denmark Germany
Private Individual Count 28 11 38 77
% within Country
16.6% 5.8% 7.9% 9.1%
(Group of) women in general
Count 17 33 65 115
% within Country
10.1% 17.5% 13.4% 13.7%
Female politician other party
Count 69 103 242 414
% within Country
40.8% 54.5% 50.0% 49.2%
Female politician own party Count 19 19 81 119 % within Country 11.2% 10.1% 16.7% 14.1% Other Count 36 23 58 117 % within Country 21.3% 12.2% 12.0% 13.9% Total Count 169 189 484 842 % within Country 100.0% 100.0% 100.0% 100.0%
Table 3: Top mentioned politicians of other parties (Q4b), specified by political party
Germany Denmark Austria
1 Angela Merkel, CDU
147 Mette Frederiksen, S 147 Angela Merkel, CDU
9
60.7 % 60.7 % 13.0 %
2 Ursula von der Leyen, CDU 15 Angela Merkel, CDU 6 Sandra Frauenberger, SPÖ 7 6.2 % 5.8 % 10.1 % 3 Annegret Kramp-Karrenbauer, CDU 14 Mette Abildgaard, DKF 3 Pamela Rendi-Wagner, SPÖ 6 5.8 % 2.9 % 8.7 % 4 Annalena Baerbock, B‘90/Die Grünen
6 Followed by three politicians, all mentioned only twice (= 1.9 %) Maria Vassilakou, Die Grünen 5 2.5 % 7.2 % 5 Sawsan Chebli, SPD 6 Klaudia Tanner, ÖVP 4 2.5 % 5.8 %
40
Table 4: Topics mainly covered (Q4a) of Tweet’s about public figures
Type of Woman Total
Female politician other party
Female politician
own party Other
Her
work/function
Count 384 117 79 582
% within Kind of Woman 93,4% 99,2% 75,2% 91,4%
Individual characteristics / personality
Count 10 0 10 21
% within Kind of Woman 2,4% 0,0% 9,5% 3,3%
Private Life Count 8 1 2 11
% within Kind of Woman 1,9% 0,8% 1,9% 1,7%
Looks Count 1 0 0 1
% within Kind of Woman 0,2% 0,0% 0,0% 0,2%
Other Count 7 0 10 17
% within Kind of Woman 1,7% 0,0% 9,5% 2,7%
None Count 1 0 4 5
% within Kind of Woman 0,2% 0,0% 3,8% 0,8%
Total Count 411 118 105 637
41
Table 5: Main role of the woman/women portrayed in Tweet (Q6), specified by political party Country
Total
Austria Denmark Germany
Role Mother Count 6 2 19 27
% within Country 3,6% 1,1% 3,9% 3,2% Housewife Count 2 0 5 7 % within Country 1,2% 0,0% 1,0% 0,8%
Physically decorative Count 1 0 0 1
% within Country
0,6% 0,0% 0,0% 0,1%
Professional function Count 122 146 367 635
% within Country
72,2% 77,2% 75,8% 75,4%
Cultural threat Count 3 12 12 27
% within Country 1,8% 6,3% 2,5% 3,2% Victim of cultural threat Count 19 24 48 91 % within Country 11,2% 12,7% 9,9% 10,8% Other Count 16 5 33 54 % within Country 9,5% 2,6% 6,8% 6,4% Total Count 169 189 484 842 % within Country 100,0% 100,0% 100,0% 100,0%
Table 6: Anti-feminism factor (Q9), specified by role
N Mean Std. Deviation Std. Error 95% Confidence Interval for Mean
Minimum Maximum Lower Bound Upper Bound Mother 27 ,93 1,412 ,272 ,37 1,48 0 4 Housewife 7 1,00 ,816 ,309 ,24 1,76 0 2 Physically decorative 1 ,00 . . . . 0 0 Professional function 634 1,00 ,888 ,035 ,93 1,07 0 4 Cultural threat 26 2,00 1,020 ,200 1,59 2,41 0 4 Victim of cultural threat 91 ,82 1,235 ,129 ,57 1,08 0 4 Other 54 ,98 1,173 ,160 ,66 1,30 0 4 Total 840 1,01 ,989 ,034 ,94 1,07 0 4
42
Table 7: Migration background (Q7), specified by type of woman Type of Woman Total Private Individual (Group of) women in general Female politician other party Female politician own party Other Migration background Explicitly migrant Count 18 28 1 0 2 49 % within Kind of Woman 23,4% 24,3% 0,2% 0,0% 1,7% 5,8% Explicitly native Count 3 6 0 0 1 10 % within Kind of Woman 3,9% 5,2% 0,0% 0,0% 0,9% 1,2% Known to be migrant Count 0 5 14 3 8 30 % within Kind of Woman 0,0% 4,3% 3,4% 2,5% 6,8% 3,6% Known to be native Count 13 15 367 115 75 585 % within Kind of Woman 16,9% 13,0% 88,6% 96,6% 64,1% 69,5% Not detectable Count 38 52 2 1 10 103 % within Kind of Woman 49,4% 45,2% 0,5% 0,8% 8,5% 12,2% Other Count 5 9 30 0 21 65 % within Kind of Woman 6,5% 7,8% 7,2% 0,0% 17,9% 7,7% Total Count 77 115 414 119 117 842 % within Kind of Woman 100,0% 100,0% 100,0% 100,0% 100,0% 100,0%
43
Table 8: Anti-feminism factor (Q9), specified by type of woman
N Mean Std. Deviation Std. Error 95% Confidence Interval for Mean
Minimum Maximum Lower Bound Upper Bound Private Individual 76 ,59 ,912 ,105 ,38 ,80 0 4 (Group of) women in general 115 1,30 1,376 ,128 1,04 1,55 0 4 Female politician other party 413 1,30 ,781 ,038 1,23 1,38 0 4 Female politician own party 119 ,02 ,129 ,012 -,01 ,04 0 1 Other 117 ,95 ,990 ,092 ,77 1,13 0 4 Total 840 1,01 ,989 ,034 ,94 1,07 0 4
