Gendered debates? The use of gender-sensitive language in German televised debates, 1997–2022

Authors:
Jennifer Bast University of Kaiserslautern-Landau, Germany

Search for other papers by Jennifer Bast in
Current site
Google Scholar
Close
,
Jürgen Maier University of Kaiserslautern-Landau, Germany

Search for other papers by Jürgen Maier in
Current site
Google Scholar
Close
,
Georg Albert University of Kaiserslautern-Landau, Germany

Search for other papers by Georg Albert in
Current site
Google Scholar
Close
, and
Jan Georg Schneider University of Kaiserslautern-Landau, Germany

Search for other papers by Jan Georg Schneider in
Current site
Google Scholar
Close
Full Access
Get eTOC alerts
Rights and permissions Cite this article

Germany is a prime example of a country where the use of gender-sensitive language is the subject of public controversy and debate. This study provides a quantitative content analysis of all (N = 72) German televised debates aired on the federal and state level from 1997 until 2022 to explore the determinants of the use of gender-sensitive language by political candidates. We find that the use of gender-sensitive language has increased over time, but the effect is small and not robust. Party affiliation correlates with candidates’ linguistic behaviour, indicating that conservative candidates use less gender-sensitive language. Candidates’ gender and party affiliation, as well as parties’ socio-political positions, interact in predicting the use of gender-sensitive language. Candidates also change their use of gender-sensitive language from one debate to another. Compared to liberal/left-wing candidates, these changes are less likely to be associated with an increase in gender-sensitive language for conservative candidates.

Abstract

Germany is a prime example of a country where the use of gender-sensitive language is the subject of public controversy and debate. This study provides a quantitative content analysis of all (N = 72) German televised debates aired on the federal and state level from 1997 until 2022 to explore the determinants of the use of gender-sensitive language by political candidates. We find that the use of gender-sensitive language has increased over time, but the effect is small and not robust. Party affiliation correlates with candidates’ linguistic behaviour, indicating that conservative candidates use less gender-sensitive language. Candidates’ gender and party affiliation, as well as parties’ socio-political positions, interact in predicting the use of gender-sensitive language. Candidates also change their use of gender-sensitive language from one debate to another. Compared to liberal/left-wing candidates, these changes are less likely to be associated with an increase in gender-sensitive language for conservative candidates.

Key messages

  • Germany is a prime example of a country where the use of gender-sensitive language is highly debated.

  • Deciding whether to use gender-sensitive language in public appearances like televised debates is difficult for candidates.

  • There is no robust increase in gender-sensitive language in televised debates over time, but party affiliation is a significant predictor.

  • Candidate gender and party affiliation, as well as socio-political positions, interact in predicting the use of gender-sensitive language.

Introduction

Despite the massive changes in the ‘political communication ecosystem’ (Esser and Pfetsch, 2020), a dinosaur of election campaign communication has survived so far: televised debates. First introduced in the 1960s, they can still be considered ‘the most important (single) event in the course of an election campaign’ (Maier and Faas, 2011: 76), thus serving ‘as a point of reference on the candidates’ messages for the entire nation’ (Benoit, 2014: 4). Given the potentially large impact debates can have on voters (see, for example, McKinney and Carlin, 2010), candidates often prepare for them meticulously (see, for example, Schroeder, 2016).

This preparation traditionally includes decisions on ‘what to say’ and ‘how to say it’. Recently, however, the question of whether and how to verbally address different genders has become relevant. The discussion about the use of gender-sensitive language (GSL), that is, using terms that are inclusive of multiple genders rather than using words that name one gender (usually the male), has developed primarily in countries where languages are spoken in which gendered nouns and pronouns play a central role (for example, the Romanian and Scandinavian languages, or German) and leads to heated discussions on whether GSL should be used in the communication of public administrations or in journalism. While a male bias in language had already been a topic among feminists since the 1970s, GSL became a broadly discussed issue when the German Federal Constitutional Court ruled in 2017 that public authorities must offer a third category besides ‘male’ and ‘female’ in official documents (Schneider, 2021). In the last five to six years, we have also had a number of publicly received and politically controversial statements by linguists in German-speaking mass media (both pro and contra); for instance, a polemic signature campaign was carried out by the conservative language association Verein Deutsche Sprache, in which gender-sensitive language was labelled across the board as ‘Gender-Unfug’ (‘gender nonsense’) that should be abolished.1

In Germany, anecdotal evidence suggests that the political elite is divided on this issue (see, for example, Schneider et al, 2022). Although parties like the Left Party2 and the Green Party tend to support the use of GSL – in the case of the Green Party, even in the form of binding rules3 – conservative parties are more sceptical; recently, even some states governed by the Christian Democrats (CDU/CSU) have banned the use of GSL in administration and schools.4 Moreover, polls show that citizens are often highly polarized on this issue (Geiger and Graf, 2019; Infratest dimap, 2021). According to public opinion polls, in 2020, a slight majority supported GSL in written (54 per cent) or spoken language (52 per cent) in the media. In 2022, this proportion decreased to 41 per cent each (Infratest dimap, 2023). Furthermore, GSL is an important issue for a significant part of the population. In 2020 (2022), 38 (36) per cent perceived GSL as very or somewhat important (Infratest dimap, 2023). Overall, these figures suggest that GSL is of moderate importance for Germans and that Germans are becoming more critical of its use. Thus, the decision for candidates as to whether or not to use GSL in televised debates is not an easy one to make.

The difficulties in dealing with GSL are multidimensional. First, there are different approaches to implementing GSL, each with varying degrees of acceptance (Bonnin and Coronel, 2021; Michaux et al, 2021). Second, once implemented, time works in favour of a positive attitude towards GSL (Gustafsson Sendén et al, 2015). This means that candidates may need to review their behaviour from time to time. Third, (dis)approval of GSL is not evenly distributed among the electorate; research has shown that attitudes towards the use of GSL are correlated with gender, age and party identification (see, for example, Parks and Roberton, 2004; Gustafsson Sendén et al, 2015). Therefore, candidates who participate in televised debates face the problem of having to meet the expectations of their own supporters while always keeping an eye on the attitudes of the broader electorate. Finally, the use (or lack) of GSL will reveal clearly visible differences between candidates; however, the effects and dynamics of a candidate’s own strategy are difficult to assess, including with regard to a debate’s follow-up coverage.

Therefore, the question is how candidates deal with these challenges in a televised debate and what factors explain their behaviour. In addition, there is the question of whether candidates adapt their behaviour over time and which candidates do so to a particular extent. While recent studies have examined the determinants of politicians’ use of GSL, for example, in parliamentary speeches or online communication (Stecker et al, 2021; Fabry, 2023), to our knowledge, there are no studies yet on televised debates. In view of the great public attention given to televised debates – for example, the debates in the run-up to the 2021 federal election were watched by up to 11 million viewers (Geese and Hess, 2021) – and the heterogeneous debate audience (see, for example, Maier and Faas, 2011; Maier, 2024), the candidates’ behaviour here could be quite different than in other communication channels.

This interdisciplinary article at the intersection of political communication and linguistics addresses this shortcoming. We analyse the use of GSL (that is, simultaneous reference to both men and women) using a content analysis of all US-style televised debates aired on German TV in the run-up to federal and state-level elections over a period of a quarter-century (1997–2022). This extends the empirical evidence we have presented elsewhere (Schneider et al, 2022). In addition, we present for the first time data on the change in the use of GSL by candidates who participated more than once in televised debates. Our results indicate: (1) candidates’ use of GSL has increased over time, but the effect is small and not robust; (2) party affiliation, but not a party’s stance on social issues, correlates with candidates’ behaviour; (3) candidate gender and party affiliation, as well as parties’ socio-political positions, interact, suggesting differences between male and female candidates running for conservative parties or in conservative political contexts (with female candidates more likely to use GSL) but not between those running for progressive/left-wing parties; (4) a candidate’s use of GSL is correlated with the respective behaviour of other actors, particularly the moderators; and (5) candidates from the CDU/CSU have a higher probability to change their use of GSL from one debate to another, though these changes are less likely to be associated with an increase in GSL than is the case for liberal/left-wing candidates.

How to refer to unknown persons and mixed-gender groups in German: a brief introduction to the linguistic background and possibilities of using gendered expressions in spoken language

The German language can typologically be described as a gender language with three gender5 classes: feminine (‘die’), masculine (‘der’) and neuter (‘das’). Determiners, pronouns and adjectives have different forms to mark the gender class, while nouns have an inherent gender. Grammatical gender is therefore a prominent feature of German, whereas, for instance, English uses pronouns to distinguish gender but has no grammatical markers of gender in the proper noun phrase (Prewitt-Freilino et al, 2012: 269). In English, ‘most human nouns are not formally marked for lexical gender’ (Hellinger and Bußmann, 2001: 8), for example, ‘citizen’ or ‘patient’. For German nouns, however, grammatical gender is obligatory, even if the gender category is not always easily recognizable: ‘While gender-class membership is not consistently marked on the noun itself, it is overtly evident from (singular) dependent elements, primarily articles and pronominal forms’ (Bußmann and Hellinger, 2003: 143).

The inherent grammatical gender of nouns often poses the question of how to refer to a person of unknown or irrelevant gender or to a group of female and male persons. As in many other languages, masculine terms in German are used to generically refer to both male and female persons. The crucial point is the difference between grammatical gender and social gender and, consequently, the possible ambiguity of masculine/male, on the one hand, or feminine/female, on the other. The assumption underlying the feminist language critique is that ‘men and women interpret male generics in a gendered way’ and that those male generics thus ‘could have a lasting impact on real world gender stereotyping and role behavior’ (Prewitt-Freilino et al, 2012: 270). As in natural gender languages like English, however, it is also possible in German for some plural forms that are not specified in terms of genus to evoke one-sided masculine or feminine associations due to sociocultural stereotypes. For example, this is the case for ‘die Abgeordneten’ (‘the delegates’), which seems ‘to contain a male bias’ (Bußmann and Hellinger, 2003: 150), though this plural form equally allows for male and female referents. It is not a generic masculine form, but it is unspecified with respect to grammatical gender: the singular could be ‘der Abgeordnete’ (‘male delegate’) or ‘die Abgeordnete’ (‘female delegate’), while the (definite) plural article in German in the nominative is always ‘die’. Prewitt-Freilino et al (2012: 270) describe the whole issue as linguistic forms being ‘used to refer to males as well as females’. It follows that one should distinguish between the form and its usage in context.

From a systematic matter-of-fact perspective, the following arguments regarding the use of generic masculine forms in German should be considered. The masculine is the unmarked basis in word formation processes. The word formation suffix ‘-er’ is added to a verb like ‘wähl-en’ (‘vote’) to form an agent noun like ‘Wähl-er’ (‘voter’), which basically denotes a person performing the activity in question (here, a person who votes) irrespective of their gender. As part of this word formation process, the agent suffix ‘-er’ assigns the masculine gender to the respective noun. The word formation suffix ‘-in’ is used to form a feminine noun and counts as marked because its sole function is to display the social or biological gender of the referent (for example, ‘Wähl-er-in’ [‘female voter’]). This suffix ‘is an indispensable means of achieving female visibility, a situation very different from current English, which has no productive word-formation pattern for the derivation of female terms’ (Bußmann and Hellinger, 2003: 153). The other side of the coin is that because the masculine class of personal nouns is used to form agent nouns but can also be used to exclusively address or refer to male persons, it is systematically ambiguous, and its use can – according to psycholinguistic experiments on written language (see, for example, Stahlberg and Sczesny, 2001; Blake and Klimmt, 2010) – result in a mentally less distinct representation of non-male persons. Hence, the use of forms that can be interpreted either generically or as referring to men exclusively can – depending on the co- and context, as well as the referential mode (see, for example, Zifonun, 2018) – prompt people to have male-biased associations. However, there is evidence that some GSL strategies have a reverse effect and make women more ‘visible’ than men (see, for example, Heise, 2000; Blake and Klimmt, 2010).

There are several proposals for gender-sensitive German (see, for example, Elsen, 2020), which are critically discussed (see, for example, Eisenberg, 2018; Zifonun, 2018). In spoken language, the means for explicit gender symmetry are basically double mentioning (‘Wählerinnen und Wähler’), substituting nouns of indefinite social gender – either epicene nouns or plural forms of nominalized adjectives and participles like ‘Wählende’ (‘voting people’, that is, people who are voting) – and various kinds of paraphrasing, for example, ‘Menschen, die uns wählen’ (‘people who vote for us’). In addition, another variant of pronouncing personal nouns with a glottal stop between the morphological stem and the feminine suffix ‘-in’ has emerged during recent years (for example, ‘/vɛːlɐʔɪnən/’, ‘Wähler-innen’). The glottal stop has further fuelled the already-heated debate on GSL; many people sense that their success in social situations is being influenced by their own communicative behaviour regarding gendered expressions (Schneider, 2021).

By using GSL (or not using it), one undoubtedly takes a socio-political and – as certain parties are for or against GSL – ideological position recognizable to the interlocutor. Using the plural morpheme ‘-er’ (that is, refraining from using GSL to refer to a potentially mixed-gender group, for example, ‘Politiker’, ‘Wähler’ or ‘Bürger’), as well as the use of GSL by double mentioning (for example, ‘Wähler und Wählerinnen’), substitute nouns (‘Wählende’) or a glottal stop (‘Wähler-innen’), marks the sender socially and politically, at least unconsciously. It is thus a relevant question to what extent the heated debate about GSL does influence the communicative behaviour of politicians during pre-election TV debates. Since televised debates address the entire electorate, the politicians’ wording might reveal not only their preferences about GSL but also their implicit assumptions about which forms of gender reference are socially accepted.

Expectations regarding the use of GSL in German televised debates

Televised debates – that is, a discussion between the most promising candidates for the head of government in front of running cameras – are considered the most important single events in election campaigns (Maier and Faas, 2011). Given the popularity of the format, as well as the careful preparation of candidates for this event, televised debates are a perfect source to analyse the (changing) use of GLS and its drivers.

Research has identified the factor of time (Gustafsson Sendén et al, 2015), as well as characteristics at the individual level – gender (see, for example, Koeser and Sczesny, 2014; Gustafsson Sendén et al, 2015) and political orientation (see, for example, Formanowicz et al, 2013; Gustafsson Sendén et al, 2015) – as affecting citizens’ attitude towards and the use of GSL (see Stecker et al, 2021; Fabry, 2023). Why are these factors influential? We argue that the use (or non-use) of GLS is a product of the perception of social norms and/or a form of identity signalling.

Social norms are unwritten rules that guide our behaviour because we align it with the perceived behaviour or morals of social groups (Rimal and Real, 2005). For our study, descriptive and injunctive norms are particularly relevant. Descriptive norms can be understood as ‘what is commonly done’ (Cialdini et al, 1991: 202), that is, the type of behaviour that is perceived to be that of most people. Descriptive norms can motivate behaviour and facilitate decision making by suggesting behaviour that promises success (Cialdini et al, 1991) – after all, what many do cannot be that futile. Descriptive norms are particularly effective in shaping behaviour when they are made salient (Cialdini et al, 1991), such as through presence in public debate. In the case of GLS, a visible outcome of the growing public attention to the topic is the increasing use of gendered language, especially in politics, public administration and the media (see, for example, Sczesny et al, 2016; Michaux et al, 2021). Thus, political candidates may increasingly observe the use of GSL by public figures and journalists and might perceive it as a norm that should be followed when speaking in public.

Since research on the use of GSL by politicians in the German parliament and on Twitter finds an increase over time (Stecker et al, 2021; Fabry, 2023), we thus expect a positive effect of this trend on politicians’ use of GSL, especially in public events like televised debates:

H1: The use of GSL by candidates in televised debates increases over time.

Injunctive norms can be understood as the perception of the type of behaviour that a specific social group expects or considers as appropriate, thus motivating behaviour by providing clues about which actions are likely to be rewarded or sanctioned by this group (Cialdini et al, 1991; Niemiec et al, 2020). In the case of political candidates, the reference group could be the party, as well as their own voters. Assumptions about how their constituents feel about GSL could thus influence the behaviour of political candidates at public events like televised debates in order to increase voter support. Research shows that people who consider themselves as liberal or politically left-wing are more sympathetic to GSL than conservatives or people who think about themselves as politically right-wing (Formanowicz et al, 2013; Gustafsson Sendén et al, 2015; Infratest dimap, 2020). This suggests that the use of GSL may represent an injunctive norm among voters on the left side of the political spectrum, whereas this is less the case among conservatives.

Additionally, the use of GSL can also be understood as a form of identity signalling. Identity signals are ‘any components of communication that inform receivers of the signaler’s membership (or non-membership) in a subset of individuals’ (van der Does et al, 2022: 1). Although the concept has been explored primarily within the electorate (see, for example, Smaldino et al, 2018; van der Does et al, 2022), there are indications that identity signalling can be beneficial to political candidates (Peskowitz, 2019), and there is evidence that politicians successfully communicate cues (see, for example, Maier et al, 2012). Thus, it may be advantageous for political candidates to signal their stance on the issue by using or not using GSL, depending on their political background and their assumptions about the views of their potential voters.

Recent research on the use of GSL by German politicians is supportive of these assumptions, showing that politicians from centre-right/right-wing parties use GSL less than those from centre-left/left-wing parties (Stecker et al, 2021). Therefore, we expect an impact of the candidate's party affiliation on the use of GSL6:

H2: Candidates from parties of the political left (the SPD, Green Party and Left Party) use more GSL than those from parties of the political right (the CDU/CSU).

While it seems reasonable to expect that parties on the political left take more progressive stances regarding socio-political issues like gender equality and should thus be more likely to endorse GSL in their communication, research suggests that these positions can vary. On the one hand, German political parties have shifted their stances on socio-political issues over time (Bräuninger et al, 2020: ch 4). This might also include their position on the use of GSL. On the other hand, German political parties are subdivided into regional entities that participate in state-level elections, whose socio-political positions can differ significantly (Müller, 2009; Bräuninger et al, 2020: ch 4); for example, the CDU in West Germany represents more progressive social positions than the CDU in East Germany (Bräuninger et al, 2020: 195). Following our previous argument, candidates from regional parties with more progressive stances on socio-political issues should be more likely to use GSL in public appearances, keeping in line with their parties’ injunctive norms and signalling their position to potential voters. Conversely, candidates representing parties or regions with more conservative or traditional socio-political views may be less inclined to adopt GSL, reflecting their party’s alignment with more traditional social norms and values. Therefore, we follow up our expectations regarding differences between parties with a more nuanced hypothesis that considers variations not only between but also within parties:

H3: Candidates from parties with progressive socio-political positions use more GSL than candidates from parties with conservative socio-political positions.

Women tend to be more sympathetic to the use of GSL than men (see, for example, Parks and Roberton, 2002; 2004; Koeser and Sczesny, 2014; Michaux et al, 2021) and use it more often (Gustafsson Sendén et al, 2015; Koeser et al, 2015; Stecker et al, 2021; Fabry, 2023; but see Koeser and Sczesny, 2014). This could be due to the fact that women benefit more than men from forms of expression that are intended to increase their visibility and linguistic identifiability (Klann-Delius, 2005) and are therefore more sensitized to the issue (Parks and Roberton, 2002: 205). In contrast, men – even if they are positive about the equal treatment of women in society – should recognize that a change in the status quo threatens their ancestral position.

With respect to candidates, these findings suggest that, first, female candidates may be more positive about GSL than males and, second, gender-specific preferences towards GSL can serve as an injunctive norm for candidates. Moreover, GSL could also be deliberately used as a form of identity signalling to appeal to voters of a particular gender by suggesting, ‘I am one of you’. Although political actors running for so-called ‘popular’ parties want to appeal to the entire electorate, female candidates may be particularly motivated to use GSL because they can use it in a credible way to appeal to female voters. Of course, this argument also applies to male candidates. They should rather refrain from using GSL because male voters see GSL more critically. Third, the decision to (not) use GSL may increase homophily effects, that is, the tendency of people to prefer others who have the same social profile or hold the same attitudes (McPherson et al, 2001). Homophily effects have also been shown for voting behaviour, particularly in relation to gender (see, for example, Dolan, 2008). Therefore, we expect the following difference between female and male candidates in the use of GSL:

H4: Female candidates use more GSL in televised debates than male candidates.

On top of the assumed direct effects, gender and candidates’ party affiliation, as well as their parties’ socio-political position, might interact with respect to the use of GSL: we find it less likely that men and women of left-wing parties and parties or regions with progressive views on society differ in their use of GSL, as both genders are equally aligned with the attitudes of their voters and are more likely to support gender equality measures themselves, and should thus both use GSL as a form of identity signal. Among conservative politicians, however, female candidates may have more reason to use GSL than male candidates because the considerations regarding possible benefits for women and the resulting strategic implications discussed earlier also apply to them. Since we are not aware of any research on the interaction between gender and party affiliation or parties’ socio-political positions, we take an exploratory approach to these aspects and formulate the following research questions:

RQ1: Is there an interaction between gender and party affiliation with respect to the use of GSL?

RQ2: Is there an interaction between gender and parties’ socio-political positions with respect to the use of GSL?

Finally, we are interested in changes in the use of GSL by individual politicians. Although much research has examined the long-term dynamics of campaign communications – for example, whether campaigns have become more personalized over time (Kriesi, 2012) – campaigns can also change in the very short term. Campaigns often need to be adapted to changing circumstances, such as changes in polls or because the opponent has changed their strategy (see, for example, Hartman et al, 2017). Moreover, campaigns are usually evaluated after the election to find out what went well and what proved to be suboptimal or even detrimental.

Candidates who participate in multiple televised debates will therefore try to learn from previous appearances, and there is no reason why their handling of GSL should not be put to the test. If candidates who have previously refrained from using GSL conclude that using GSL will help them achieve electoral support, they will change their voter salutation accordingly. Conversely, candidates who have used GSL will refrain from doing so if they conclude that it is reducing their electoral chances. If candidates conclude that using (or not using) GSL was helpful, they will not change their behaviour.

Again, televised debates are an ideal event to analyse whether candidates really change their campaign communication over time. However, we are not aware of any research analysing individual-level changes in debate strategies. Due to the lack of previous findings, we formulate the following research question to exploratively investigate the determinants of change in GSL use:

RQ3: Which factors drive the change in candidates’ use of GSL between subsequent debates?

Method

Data

Content analysis

Our analyses are based on a quantitative content analysis of all (N = 72) German televised debates on the federal (N = 12) and state level (N = 60), covering a time period of 25 years (1997 to 2022). In these ‘US-style’ TV debates, the (two or three) most promising candidates for the head of government (federal level: chancellor; state level: prime minister) discuss the most pressing issues in the run-up to an election live on TV. Unlike in some other countries, such debates in Germany – often referred to as a ‘duel’ (two candidates) or ‘truel’ (three candidates) – are not compulsory or derived from customary law but based on successful negotiations between the broadcasters and the campaign teams (Maier, 2023), usually the top candidates of the parties leading in the polls. This format should not be confused with so-called ‘elephant rounds’, in which the leaders of all parties represented in a parliament participate (see, for example, Schrott and Lanoue, 1992).

We first set up a transcript for every debate. Second, the transcripts were searched for the words or word components that indicated that either citizens (‘bürger’) or voters (‘wähl’) were directly addressed.7

In order to determine and classify the relevant linguistic forms with regard to the use of GSL, we used the following categories:

  • generic masculinum (plural), for example, ‘(die) Wähler’ (‘voters’);

  • generic masculinum (singular), for example, ‘(der) Wähler’ (‘male voter’);

  • generic feminine (plural), for example, ‘(die) Wählerinnen’ (‘female voters’);

  • generic feminine (singular), for example, ‘(die) Wählerin’ (‘voter’);

  • class-referring8 feminine, for example, ‘Bei den Wählerinnen ist Friedrich Merz viel unbeliebter als bei den Wählern’ (‘Among female voters, Friedrich Merz is much more unpopular than among male voters’);

  • double mentioning/splitting, for example, ‘Bürgerinnen und Bürger’ (‘female citizens and male citizens’);

  • neutral substitute expression formed from the present participle, for example, ‘Wählende’ (‘voting persons’);

  • references to a particular female, for example, ‘Wählerin Hanna Schmitz’ (‘the voter Hanna Schmitz’);

  • references to a particular male, for example, ‘Wähler Wolfgang Müller’ (‘the voter Wolfgang Müller’);

  • non-gendered composite nouns, for example, ‘Bürgerversicherung’ (‘citizen insurance’);

  • gendered composite nouns, for example, ‘Wählerinnen-Verzeichnis’ (‘electoral register’; literally, ‘(female) voters register’);

  • non-gendered derivations, for example, ‘Wählerschaft’ (‘voting public’);

  • gendered derivations, for example, ‘wählerinnenmäßig’ (‘in a female voter’s manner’) or ‘Wählerinnenschaft’ (‘[female] voting public’);

  • non-specific references, for example, ‘Wenn ein Wähler keine Wahlbenachrichtigung dabei hat, muss er extra ein Formular ausfüllen’ (‘If a voter does not bring a voting card, he needs to fill in a special form’9); and

  • gendering with a glottal stop, for example, ‘Wähler-innen’ (‘voters’).

To assess intercoder reliability, 10 per cent of the references to the relevant groups were randomly selected and coded again. The reliability of our coding is high (Holsti’s formula: αmin = .99, αmax = 1.0; Krippendorff’s α: αmin = .98, αmax = 1).

Cross-sectional data

Data preparation

Content analysis data were aggregated for each debate and each candidate, so that each candidate’s appearance in a debate represented one observation. From the total of 158 observations, we selected the 138 candidate appearances that contain at least one reference to citizens or voters (for an overview of the steps involved in the data preparation, see Figure A1 in the Online Appendix, available at: https://osf.io/8ardj/).

A first descriptive analysis shows that double mentioning and generic masculinum are most important (see Table 1). Other options to use GSL, such as substitute participles like ‘Wählende’ (‘voting persons’), did not occur in our corpus, and other paraphrases (for example, ‘alle, die uns wählen’ [‘everyone voting for us’]) are too variable for the kind of corpus analysis we conducted (for a description of the data preparation, see the Online Appendix, available at: https://osf.io/8ardj/). The analysis is thus restricted to the categories ‘generic masculinum’ and ‘double mentioning’. In the following, we describe the variables included in our analysis.

Table 1:

Expressions used by political candidates in German televised debates to refer to voters and citizens

Expression Political candidates
References to voters (‘Wähler*’) References to citizens (‘Bürger*’) Total
n % n % n %
Generic masculinum (plural) 73 40 116 21 189 25
Generic masculinum (singular) 24 13 27 5 51 7
Generic feminine 1 1 2 0 3 0
Class-referring feminine 0 0 0 0 0 0
Double mentioning 72 39 216 38 288 38
Neutral substitute expressions formed from the present participle 0 0 0 0 0 0
Other neutral substitute expressions 0 0 0 0 0 0
Reference to a particular female 0 0 1 0 1 0
Reference to a particular male 0 0 0 0 0 0
Reference to a particular institution 0 0 0 0 0 0
Non-specific reference 0 0 0 0 0 0
Gendered composite nouns 0 0 1 0 1 0
Non-gendered composite nouns 14 8 166 29 180 24
Gendered derivations 0 0 0 0 0 0
Non-gendered derivations 0 0 36 6 36 5
Glottal stop 0 0 0 0 0 0
Total 184 100 565 100 749 100

Dependent variable

The dependent variable in our models is the proportion of double mentions in the sum of double mentions and generic masculines used to address voters or citizens. This reduces our data to N = 126 candidate appearances by 79 politicians in which at least one of these expressions was used to refer to these groups.10

Independent variables

The data contain information about the candidates’ gender (0 = male; 1 = female) and party affiliation (0 = SPD/Green Party/Left Party; 1 = CDU/CSU). Our data contain a total of 27 appearances by female and 99 appearances by male candidates. The CDU/CSU and SPD account for the majority of these, with 57 appearances each, while the Greens and the Left Party are less frequently represented, with seven and five appearances, respectively.11 Furthermore, we added the exact date on which the debate took place.

For the measurement of the socio-political positions of the parties, we used data from Bräuninger et al (2020; see also Benoit et al, 2009; Gross and Debus, 2018).12 The positions are derived from the election manifestos of the respective parties at the federal and state level using fully automated text analysis and the ‘Word scores’ method and then standardized (for a detailed description of the method, see Bräuninger et al, 2020: ch 3). Low values indicate progressive and high values conservative socio-political positions. As the positions are available for every party and election in our data set at the federal and state level under investigation, they could be matched with the corresponding candidates in our data set.

Control variables

We added information on the setting (0 = state level; 1 = federal level) and the format (0 = classic debate; 1 = town hall meeting) for each debate. Moreover, we controlled for potential linguistic adaptation effects with reference to other debate participants. Research shows that a female debate participant can affect how male candidates behave verbally (Maier and Renner, 2018). Thus, we incorporated information on whether the speaking candidate faced at least one female opponent (0 = no; 1 = yes). In addition, we calculated the proportion of double mentions used by the respective opponent(s) in a debate, as well as the moderators, to account for the fact that candidates’ use of language may be influenced by those of the other participants – possibly because it is perceived as a descriptive norm.13 Psychology recognizes such unconscious verbal adaptation mechanisms under the term ‘chameleon effect’ (see, for example, Chartrand and Bargh, 1999). Finally, we added candidates’ age as a control variable, as there are indications that attitudes towards the use of GSL vary by age (see, for example, Gustafsson Sendén et al, 2015; Infratest dimap, 2023).

Longitudinal data

Data preparation

To answer RQ3, we built an additional data set consisting only of candidates who appeared in two or more debates. We created dyads including information for each candidate’s appearance in a debate and the subsequent debate. For instance, former Chancellor Angela Merkel appeared in the four debates; therefore, we calculated three dyads for her capturing the change in GSL (2005–2009, 2009–2013, 2013–2017). In total, we identified N = 28 candidates with multiple debate appearances. The final data set consists of N = 47 dyads representing two consecutive debate appearances by the same candidate that included verbal reference to the relevant groups in both of the debates.

Dependent variable

Consistent with our cross-sectional data, we measure the proportion of double mentions and/or generic masculines used by candidates to refer to citizens and/or voters. To not overstretch the data, we computed one variable indicating whether a candidate changed (1) their use in GSL or not (0) and one indicating whether the change consisted of an increase (1) or decrease (0) in the relative amount of GSL.

Independent variables

To explain the change in the use of GSL over time, we computed the time difference between the debates (in days) and included information on candidates’ gender and party affiliation. Given the small sample, we chose not to include additional contextual variables.

Analysis strategy

Our main dependent variable to analyse H1, H2, H3, H4, RQ1 and RQ2 is a proportion (relative frequency) of GSL bounded by the interval [0,1]. The relationship of proportions with one or more continuous predictors often violates assumptions of ordinary least square regression, especially for data points close to 0 or 1 (Papke and Wooldridge, 1996; Chen et al, 2017). We thus followed recommendations by Papke and Wooldridge (1996) and specified generalized linear models with a logit link function – ‘the logit transformation of the link function’ (Baum, 2008: 301) – and the binomial distribution.

Our data also include multiple interdependences due to observations being clustered in debates and candidates. Ideally, such data would be analysed using a cross-classified multilevel approach. However, our data would consist of two levels (debates and candidates) with a large number of groups of very few observations, which can lead to bias (Moineddin et al, 2007). Thus, we decided to follow an approach used by Damore (2002; see also Maier and Jansen, 2017) and estimate robust standard errors clustered in debates for the models testing H1, H2, H4 and RQ1. Clustering standard errors in debates (instead of candidates) was chosen because we assumed the lack of independence among candidate appearances in the same debate to be the greatest. This consideration is based on the fact that the appearances of candidates who participate in more than two debates are usually at least four years (a legislative period at the federal level) apart, a relatively long period of time in which language can change. The language behaviour of two candidates in the same debate could therefore already show stronger interdependencies due to the identical time and context. However, we additionally calculated all models with robust standard errors clustered in candidates. The findings suggest identical interpretations with respect to our hypotheses and research questions (see Table A5 in the Online Appendix, available at: https://osf.io/8ardj/).

We set up additional generalized linear models including the parties’ socio-political positions as an independent variable to test H3 and answer RQ2. Since these positions can be assumed to be dependent on the party more than the debate or individual candidate, we calculated robust standard errors clustered in parties for these models. To answer RQ3, we specified logistic regression models to detect the determinants of changes in the use of GSL and clustered standard errors in candidates.

Results

Cross-sectional analyses

Figure 1 shows the results of general linear regression models explaining the proportion of GSL used by candidates participating in a televised debate (for full details of the models, see Table A4 in the Online Appendix, available at: https://osf.io/8ardj/). Model 1 is the baseline model to test H1, H2 and H4. The results suggest that the proportion of GSL used by a candidate depends on the timing of the debate and the candidates’ party affiliation but not on the candidates’ gender; hence, H4 has to be rejected. The date that the debate took place is a significant predictor,14 but the effect seems to be quite small (b = 0.00). However, since the variable is measured in days, the finding of only a small effect for the length of the study period is not particularly surprising. If we extrapolate the effect to the period of one year, we see that the estimated difference between the logit of the mean of GLS is 0.06. However, the increase of GLS over time disappears in Model 2. Hence, H1 is only partly confirmed. In addition, H2 is supported. The average marginal effects indicate that, on average, being a candidate for the CDU/CSU (instead of a candidate for the SPD, Green Party and Left Party) is associated with a decrease of 0.16 in the probability of the dependent variable to take the value of 1, that is, the candidate (only) using GSL when referring to voters or citizens. This effect persists if we add control variables (see Model 2).

The coefficient plot for Model 1 shows that the proportion of GSL used by a candidate is predicted by the timing of the debate and the candidates’ party affiliation but not the candidates’ gender. The plot for Model 2 shows that only candidate gender, not the time of the debate, remains a significant predictor when control variables are added. Models 3 and 4 additionally include the interaction between candidate gender and party affiliation (which is visualized in Figure 2). The interaction is only significant in Model 3, not Model 4, which includes control variables.
Figure 1:

Coefficient plots from generalized linear models estimating factors influencing candidates’ use of GSL (Models 1–4)

Citation: European Journal of Politics and Gender 2024; 10.1332/25151088Y2024D000000023

Looking at the control variables, we find very interesting results. Model 2 includes candidates’ age, the level at which the election took place, the format of the debate, the presence of female opponents, the use of GLS by other candidates and the use of GLS by the moderators. While candidate age, opponent gender and the level of the election do not affect candidates’ use of GLS, the format of the debate is a significant predictor of candidates’ use of GLS. A town hall debate is associated with a decrease of 0.19 in the probability of candidates using GSL when referring to voters or citizens, compared with debates without an audience. Furthermore, there is strong evidence that a chameleon effect is at work. On the one hand, an increase in the use of GSL by a candidate’s opponent(s) is, on average, associated with an increase of 0.23 in the probability of the candidate using GSL when referring to voters or citizens. Furthermore, candidates’ use of GLS increases with the use of gendered language by the moderators. The effect is similar in strength to the effect measured for political opponents.

To answer RQ1, we include the interaction between candidates’ gender and their party background. The interaction term has a statistically significant impact when the candidates’ gender, the candidates’ party affiliation and the timing of the debate are held constant (see Model 3). Figure 2 shows that there is only a small difference in the use of GLS between female candidates of left-wing parties and female candidates running for the CDU/CSU. There is also no difference between male and female candidates of centre-left parties. In contrast, male candidates from the CDU use less GSL than men running for centre-left parties and women from their own party. However, the interaction effect disappears when including the full set of control variables (see Model 4).15

Figure 2 visualizes the interaction between candidate gender and party affiliation detected in Model 3. It shows that female and male candidates from the SPD, Green Party and Left Party are similar in their use of GSL. Male candidates from the CDU use less GSL than male candidates running for centre-left parties and women from their own party. Thus, gender differences only seem to be apparent among candidates from the CDU/CSU.
Figure 2:

Predictive margins for the interaction between candidates’ gender and party affiliation

Citation: European Journal of Politics and Gender 2024; 10.1332/25151088Y2024D000000023

Figure 3 shows the results of general linear regression models estimating candidates’ use of GSL considering parties’ socio-political positions (for regression tables, see Table A11 in the Online Appendix, available at: https://osf.io/8ardj/).16 Model 5 is the baseline model to test H3. The results suggest that candidates’ use of GSL is not affected by their parties’ socio-political position. This observation also holds for the model including all control variables (Model 6). H3 must therefore be rejected. As in the previous analysis (with party background instead of socio-political position as an independent variable), gender is not a significant predictor of the use of GSL, whereas the time of the debate is (albeit with a very small effect), reinforcing our conclusions regarding H1 and H4.

The coefficient plot for Model 5 shows that the socio-political position of their party does not significantly predict candidates’ use of GSL; this result holds for Model 6, which includes additional control variables. Models 7 and 8 show that the interaction between candidate gender and a party’s socio-political position is only evident in the model with all control variables (Model 8).
Figure 3:

Coefficient plots from generalized linear models estimating factors influencing candidates’ use of GSL (Models 5–8)

Citation: European Journal of Politics and Gender 2024; 10.1332/25151088Y2024D000000023

Looking at the control variables, we also find similar patterns as in the previous analysis: candidate age, opponent gender and the level of the debate do not significantly affect candidates’ use of GSL, while the format of the debate and opponents’ and moderators’ use of GSL are significant predictors.

To answer RQ2, we include the interaction between candidates’ gender and parties’ socio-political positions (see Models 7 and 8 in Figure 3). The interaction term has a statistically significant impact in the full model including all control variables (see Model 8). As can be observed in Figure 4, there are only significant differences between female and male candidates’ use of GSL among those from parties with conservative socio-political positions. For example, the predicted value for the use of GSL by a candidate from a party with very conservative positions on society (socio-political position = 15) is 0.69 for female and 0.41 for male candidates. Thus, we see similar patterns as in the findings for RQ1: in more conservative parties, female candidates seem to use GSL significantly more than male candidates, while there are fewer gender differences among candidates of more progressive or left-wing parties.

Figure 4 shows two different visualizations of the interaction between candidates’ gender and socio-political position for their use of GSL during televised debates. Both graphics show that there are only significant differences between female and male candidates’ use of GSL among those from parties with conservative socio-political positions. In these parties, female candidates use more GSL than male candidates.
Figure 4:

Interaction plots visualizing the interaction between candidates’ gender and parties’ social position

Citation: European Journal of Politics and Gender 2024; 10.1332/25151088Y2024D000000023

Overall, our results suggest that the use of GSL in German televised debates tends to increase over time, but the effect is small and not robust. Party background, the format of the debate and the use of gendered language by other participants in the debate have an important influence on candidates’ rhetorical behaviour. Differences between female and male candidates only appear to be relevant among candidates of conservative parties running in conservative socio-political contexts.

Longitudinal analyses

Since several candidates in our data set participated in more than one debate, we can also examine whether candidates change their use of GSL and which candidate characteristics explain whether or not they change their language use from debate to debate (see RQ3). Table A17 in the Online Appendix (available at: https://osf.io/8ardj/) shows that the proportion of GSL used by a candidate varies over time. However, no clear pattern emerges: in 22 of the total 47 cases, the use of GSL increased compared to the previous debate appearance. In contrast, in 16 appearances, candidates used less GSL than in the previous debate. In nine cases, the use of GSL did not change.

To explain the change in use of GSL from a candidate’s debate appearance to the next we specified logistic regression models (see Table 2). Model 1 analyses whether a candidate’s probability to change their use of GSL from one debate to another can be explained by their gender, party and the time difference between the debates. It shows that the odds of a candidate changing their use of GSL are 0.21 times as high for female candidates. Moreover, the odds of change are 7.61 times as high for candidates from the CDU/CSU as for those of the SPD/Green Party/Left Party. In contrast, the time passed between the debates does not significantly explain language change. Model 2 predicts the direction of the change in use of GSL (increase or decrease) from candidates’ gender and party and from time. While gender and time are not significant predictors, our results show that the odds of increasing their use of GSL are 0.35 times as high for members of the CDU/CSU as for those of the SPD/Green Party/Left Party.

Table 2:

Logistic regression estimating candidates’ change in use of GSL between debates and direction of change from candidates’ gender and party affiliation and the time difference between debates

Change in GSL (1 = change; 0 = no change) Direction of Change (1 = increase; 0 = decrease)
Model 1 Model 2
Exp(b) (SE) Exp(b) (SE)
Candidate profile
Gender (male) Female 0.21 (0.19)a 0.57 (0.45)
Party (SPD, Green Party, Left Party) CDU/CSU 7.61 (6.86)b 0.35 (0.20)a
Time (∆day) 1.00 (0.00) 1.00 (0.00)
constant 2.06 (1.11) 1.87 (0.85)
Wald Chi2 10.55 4.11
Prob > Chi2 0.014 .249
Pseudo R2 0.19 0.06
N 47 38

Note: Displayed are odds ratios (robust clustered standard errors in parentheses) of logistic regressions. Significance levels: a p < .1; b p < .05; c p < .01; d p < .001.

We must, however, emphasize that these results should be interpreted with caution. In general, the sample is small and consists of unequal groups. As a consequence, we were not able to examine interaction effects or further potential factors influencing change in linguistic behaviour.

Discussion

We have argued that deciding whether to use GSL is not an easy task for candidates, as they must not only meet the expectations of their supporters but also keep the electorate at large in mind. Especially when candidates are speaking in front of large and fragmented audiences, as in televised debates, they need to think carefully about how to communicate in this regard.

To analyse the use of GSL and its determinants, we conducted a quantitative content analysis of all German televised debates on the federal and state level from 1997 to 2022. Our results show that candidates who explicitly address ‘citizens’ or ‘voters’ do not use GSL in the majority of cases. In six out of ten references, they use generic masculinum (singular or plural), non-gendered composite nouns or non-gendered derivations. If candidates use GSL, they explicitly address both male and female citizens or voters (‘double mentions’). Interestingly, we did not find any case in which the highly controversial glottal stop was used.

We also showed that the use of GSL has increased significantly over a quarter-century and from one debate to another, confirming our expectation that the increasing use of GSL by the public leads to the emergence of a descriptive norm and thus influences candidate behaviour. However, the effect does not hold up to a model including additional control variables, which is at odds with the results of studies on rhetoric in the German parliament, which find stronger increases over time (Stecker et al, 2021; Fabry, 2023). This could be because of methodological differences (we analysed a shorter time period and focused on specific forms of GSL referencing specific groups) or because of the characteristics of televised debates, in which the behaviour of candidates may reflect more strongly the fact that attitudes in Germany seem to have become more negative over time (Infratest dimap, 2023).

Furthermore, candidates’ use of GSL can be explained by party affiliation: as expected, candidates running for the CDU/CSU used gendered language significantly less often than candidates running for centre-left parties. Interestingly, the socio-political position of the party is not a significant predictor of the use of GSL. In other words, there is no direct relationship between socio-political differences within a party across time (for example, today versus 20 years ago) and space (for example, between the different regions) and candidates’ use of GSL.

There is also initial evidence that candidates’ gender and party affiliation, as well as parties’ socio-political positions, interact, suggesting differences between male and female candidates running for conservative parties or in conservative political contexts (with female candidates more likely to use GSL) but not between those running for progressive/left-wing parties. Based on our theoretical framework, our findings suggest that different types of candidates refer to different reference groups when deciding on the use of GSL. Candidates from the centre-left/progressive parties as well as male candidates running for the CDU/CSU/conservative parties seem to follow the norms of their supporters. In contrast, female candidates of the CDU/CSU or parties with more conservative socio-political positions seem to resolve the conflict arising from the different norms of their party and gender in favour of gender.

Furthermore, longitudinal data demonstrate that the vast majority of candidates change their use of GSL over time; however, an increasing use of GSL is only slightly more common than a decreasing use. Conservative candidates are more likely to adjust their use of GSL over time; however, with respect to the direction of this change, we show that they are less likely to increase their use of GSL than candidates running for centre-left parties. These results are in line with the study by Stecker et al (2021) and confirm our expectation that politicians align their behaviour with their core voters. However, as public support for GSL is decreasing in Germany, centre-left candidates face growing conflict: if they use GSL, they will presumably fulfil the expectations of left-leaning voters but simultaneously use a linguistic practice that a (growing) majority of Germans seem to be critical of. Moreover, whereas cross-sectional data suggest that candidates’ gender does not directly influence the decision to use GSL, contradicting our expectations and previous findings (see, for example, Stecker et al, 2021; Fabry, 2023), the panel data indicate that male candidates adjust their use of GSL more often than female candidates; however, the direction of change is undetermined.

Finally, our results demonstrate that the use of GSL depends on the context: town hall meetings decrease the use of GSL. Moreover, candidates’ behaviour is shaped by the other participants, indicating a spontaneous adaptation to a perceived context-specific descriptive norm. The causal direction of this effect is unclear, however. Nevertheless, this correlation shows how sensitive candidates are to the use of GSL and that gendered language is likely to be considered from a strategic point of view.

Our study comes with limitations. First, we have focused on two groups we considered central in the context of political debates: voters and citizens. Thus, our findings may not be transferable to candidates’ overall linguistic behaviour. Future studies should extend the analysis to references to other social groups; our results suggest that candidates often refer to groups other than voters and citizens. Second, our data are on televised debates. Since there is evidence that candidate communication depends on context (see, for example, Stecker et al, 2021; Fabry, 2023), future studies should therefore extend the scope and analyse the use of GSL beyond televised debates and parliamentary speeches. Third, although we measured the use of GSL for all candidates running in German televised debates, the number of observations for our collection of panel data is too small for more detailed analyses. Too few cases also prevent a more differentiated analysis of the differences between the left-wing parties, which we have summarized in a single category. Since bivariate analyses show that these parties tend to differ in their use of GSL, future analyses should shed more light on the differences. Fourth, our analysis focuses on verbal communication, leaving out many ways known in written language to address gender identities other than female and male. It must therefore be noted that the analysis of written texts (for example, election campaign flyers) might yield different results regarding inclusive language. It can, however, be assumed that GSL in TV debates is primarily understood by the candidates as the equal representation of women and men, as double mentions occur but no more inclusive forms, such as the glottal stop sound.

Despite these limitations, we believe that our study provides initial and important insights into the use of GSL among politicians. The debate over how to address different genders appropriately, effectively and as acceptably as possible is in full swing, and the outcome of this sometimes highly emotional debate remains to be seen. History shows, however, that language changes. Follow-up studies will have to shed light on further developments and how politicians deal with these changes.

Notes

1

See: https://vds-ev.de/aktionen/aufrufe/schluss-mit-gender-unfug/ (for a more detailed discussion of this public and linguistic discourse, see Schneider, 2021).

5

It would actually be clearer to speak here of ‘genus’ rather than (grammatical) ‘gender’. However, in English, the Latin term ‘genus’ does not seem to be common – not even in linguistics. In German linguistics, the distinction between genus and gender is very important because, in many cases, especially when words refer to things and not to living beings, the term ‘gender’ is misleading because the genus does not have much to do with sex or (social) gender.

6

The Left Party (‘Die Linke’) was formed in 2007 as a merger of the Social Democratic Party of Germany (SPD) split-off Labour and Social Justice – The Electoral Alternative (WASG) and the Party of Democratic Socialism (PDS) (the two parties already ran jointly in 2005 as ‘Die Linke. PDS’ [for the history of the Left Party, see, for example, Hough and Keith, 2019]). Due to considerable overlaps (PDS candidates ran for the Left Party after the merger) and the limited number of candidate appearances from both parties in the data set, the candidates from the former PDS and today’s Left Party are analysed together and hereinafter referred to simply as the ‘Left Party’.

7

We initially adopted a broader approach by analysing a random subsample of eight debates to identify the groups of people being addressed by the candidates. The number of groups addressed proved to be excessively large and diverse: we identified over 200 different nouns used to refer to people. Many of these appeared only once, tied to specific situational campaign issues. Due to the considerable number and significant variability among debates, we opted against including all nouns for addressing people. Instead, we made the strategic decision to focus on voters and citizens as groups of people that: (1) hold utmost relevance in the democratic system and election campaigns; (2) are consistently present across a wide range of debates, ensuring heightened comparability; and (3) lend themselves to semi-automated analysis.

8

In contrast to the generic femininum, expressions belonging to this category refer to all females of the specific ‘class’, for example, all female voters or all female citizens. Thus, the female gender is known, but there is no specific reference nonetheless.

9

Unlike the forms used to refer generically to any member of a group or category, this form prompts listeners or readers to envision a person or group of persons without actually referring to one. Individual cases may be difficult to relate to either the generic or the non-specific use.

10

Descriptive statistics for all variables can be found in the Online Appendix (available at: https://osf.io/8ardj/).

11

Due to the small number of cases for the Greens and the Left Party, it is not possible to provide a more detailed analysis of the differences in the use of GSL between left-wing parties. However, an analysis of variance (ANOVA) indicates that candidates from left-wing parties significantly (on a significance level of p < 0.1 [see also note 14]) differ in their linguistic behaviour (F[2,66] = 2.91, p = 0.062). A Tukey post hoc test further demonstrated that GSL usage was significantly higher among SPD candidates when compared to candidates from the Left Party (p = 0.051).

12

We would like to thank Marc Debus, University of Mannheim, for providing us with the very latest data.

13

In case of more than one opponent or moderator, the arithmetic mean was calculated.

14

Due to the relatively small sample size, we chose a significance level of p < 0.1 (Thiese et al, 2016).

15

Since our sample is relatively small, we performed additional analyses to inspect five influential observations that were identified by Cook’s distances and DFBETAs (difference in beta values). We repeated all models by excluding each outlier separately and compared them to the results with the full data set (the influential cases and additional models can be found in Tables A4–A8 in the Online Appendix). Removing these observations did change some of the results with respect to our hypotheses. However, as all cases are valid components of our complete survey, we only interpret the analyses with a complete sample.

16

It should be noted that we are unable to estimate the impact of parties’ socio-political positions on the use of GSL in the former models because they are clustered on debates. To assess the effect of parties’ socio-political positions, it seems more appropriate to cluster in parties. For more details, see the ‘Methods’ section.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Acknowledgements

The authors would like to thank the anonymous reviewers for their helpful comments and suggestions.

Author biographies

Jennifer Bast is a postdoctoral researcher at the Department of Political Science at the University of Kaiserslautern-Landau, Germany. Her research focuses on visual political communication, populism, and gender and political leadership.

Jürgen Maier is Professor of Political Communication at the Department of Political Science at the University of Kaiserslautern-Landau, Germany. His research focuses on the content and impact of campaign communication.

Georg Albert is a research assistant for German linguistics at the University of Kaiserslautern-Landau, Germany. His interests focus on spoken and written interaction, as well as on social and cultural aspects of language use.

Jan Georg Schneider is Full Professor of German Linguistics at the University of Kaiserslautern-Landau, Germany. His research focuses on the mediality and multimodality of linguistic interaction, as well as on aspects of gender in the German language.

Conflict of interest

The authors declare that there is no conflict of interest.

References

  • Baum, C.F. (2008) Stata tip 63: modeling proportions, The Stata Journal: Promoting Communications on Statistics and Stata, 8(2): 299303. doi: 10.1177/1536867x0800800212

    • Search Google Scholar
    • Export Citation
  • Benoit, K., Bräuninger, T. and Debus, M. (2009) Challenges for estimating policy preferences: announcing an open access archive of political documents, German Politics, 18(3): 44154. doi: 10.1080/09644000903055856

    • Search Google Scholar
    • Export Citation
  • Benoit, W.L. (2014) Political Election Debates: Informing Voters about Policy and Character, Lanham, MD: Lexington Books.

  • Blake, C. and Klimmt, C. (2010) Geschlechtergerechte Formulierungen in Nachrichtentexten, Publizistik, 55(3): 289304.doi: 10.1007/s11616-010-0093-2

    • Search Google Scholar
    • Export Citation
  • Bonnin, J.E. and Coronel, A.A. (2021) Attitudes toward gender-neutral Spanish: acceptability and adoptability, Frontiers in Sociology, 6. doi: 10.3389/fsoc.2021.629616

    • Search Google Scholar
    • Export Citation
  • Bräuninger, T., Debus, M., Müller, J. and Stecker, C. (2020) Parteienwettbewerb in den Deutschen Bundesländern, 2nd edn, Wiesbaden: Springer VS.

    • Search Google Scholar
    • Export Citation
  • Bußmann, H. and Hellinger, M. (2003) Engendering female visibility in German, in M. Hellinger and H. Bußmann (eds) Gender across Languages. The Linguistic Representation of Women and Men, Vol. 3, Amsterdam and Philadelphia, PA: Benjamins, pp 14174.

    • Search Google Scholar
    • Export Citation
  • Chartrand, T.L. and Bargh, J.A. (1999) The chameleon effect: the perception–behavior link and social interaction, Journal of Personality and Social Psychology, 76(6): 893910. doi: 10.1037//0022-3514.76.6.893

    • Search Google Scholar
    • Export Citation
  • Chen, K., Cheng, Y., Berkout, O. and Lindhiem, O. (2017) Analyzing proportion scores as outcomes for prevention trials: a statistical primer, Prevention Science: The Official Journal of the Society for Prevention Research, 18(3): 31221. doi: 10.1007/s11121-016-0643-6

    • Search Google Scholar
    • Export Citation
  • Cialdini, R.B., Kallgren, C.A. and Reno, R.R. (1991) A focus theory of normative conduct: a theoretical refinement and reevaluation of the role of norms in human behavior, Advances in Experimental Social Psychology. Advances in Experimental Social Psychology, 24: 20134. doi: 10.1016/S0065-2601(08)60330-5

    • Search Google Scholar
    • Export Citation
  • Damore, D.F. (2002) Candidate strategy and the decision to go negative, Political Research Quarterly, 55(3): 669. doi: 10.2307/3088036

  • Dolan, K. (2008) Is there a ‘gender affinity effect’ in American politics?, Political Research Quarterly, 61(1): 7989. doi: 10.1177/1065912907307518

    • Search Google Scholar
    • Export Citation
  • Eisenberg, P. (2018) Misshandlung der Sprache. Gleichberechtigung der Geschlechter ist wünschenswert, doch das generische Maskulinum darf nicht angetastet werden. Eine Streitschrift, Tagesspiegel, 6 August.

    • Search Google Scholar
    • Export Citation
  • Elsen, H. (2020) Gender – Sprache – Stereotype: Geschlechtersensibilität in Alltag und Unterricht, Tübingen: Narr Francke Attempto Verlag.

    • Search Google Scholar
    • Export Citation
  • Esser, F. and Pfetsch, B. (2020) Political communication, in D. Caramani (ed) Comparative Politics, Oxford: Oxford University Press, pp 33658.

    • Search Google Scholar
    • Export Citation
  • Fabry, V. (2023) ‘To gender or not to gender?’ Examining the usage of gender-inclusive language by German politicians in parliament and on Twitter, paper presented at the Annual Conference of the International Communication Association (ICA), Toronto.

    • Search Google Scholar
    • Export Citation
  • Formanowicz, M., Bedynska, S., Cisłak, A., Braun, F. and Sczesny, S. (2013) Side effects of gender-fair language: how feminine job titles influence the evaluation of female applicants, European Journal of Social Psychology, 43(1): 6271. doi: 10.1002/ejsp.1924

    • Search Google Scholar
    • Export Citation
  • Geese, S. and Hess, C. (2021) Nutzung und Bewertung medialer Wahlinformationen. Ergebnisse einer Repräsentativbefragung zur Bundestagswahl 2021, Media Perspektiven, 12: 61024.

    • Search Google Scholar
    • Export Citation
  • Geiger, A.W. and Graf, N. (2019) About One-in-five U.S. Adults Know Someone who Goes by a Gender-Neutral Pronoun, Washington, DC: Pew Research Center, www.pewresearch.org/short-reads/2019/09/05/gender-neutral-pronouns/.

    • Search Google Scholar
    • Export Citation
  • Gross, M. and Debus, M. (2018) Does EU regional policy increase parties’ support for European integration?, West European Politics, 41(3): 594614. doi: 10.1080/01402382.2017.1395249

    • Search Google Scholar
    • Export Citation
  • Gustafsson Sendén, M., Bäck, E.A. and Lindqvist, A. (2015) Introducing a gender-neutral pronoun in a natural gender language: the influence of time on attitudes and behavior, Frontiers in Psychology, 6. doi: 10.3389/fpsyg.2015.00893

    • Search Google Scholar
    • Export Citation
  • Hartman, T.K., Pattie, C. and Johnston, R. (2017) Learning on the job? Adapting party campaign strategy to changing information on the local political context, Electoral Studies, 49: 12835. doi: 10.1016/j.electstud.2017.06.005

    • Search Google Scholar
    • Export Citation
  • Heise, E. (2000) Sind Frauen mitgemeint? Eine empirische Untersuchung zum Verständnis des generischen Maskulinums und seiner Alternativen, Sprache & Kognition, 19(1–2): 313. doi: 10.1024//0253-4533.19.12.3

    • Search Google Scholar
    • Export Citation
  • Hellinger, M. and Bußmann, H. (2001) Gender across languages. The linguistic representation of women and men, in M. Hellinger and H. Bußmann (eds) Gender across Languages. The Linguistic Representation of Women and Men, Vol. 1, Amsterdam and Philadelphia, PA: Benjamins, pp 125.

    • Search Google Scholar
    • Export Citation
  • Hough, D. and Keith, D. (2019) The German Left Party: a case of pragmatic populism, in G. Katsambekis and A. Kioupkiolis (eds) The Populist Radical Left in Europe, Abingdon and New York, NY: Routledge, pp 12944.

    • Search Google Scholar
    • Export Citation
  • Infratest dimap (2020) Vorbehalte gegenüber genderneutraler Sprache, www.infratest-dimap.de/umfragen-analysen/bundesweit/umfragen/aktuell/vorbehalte-gegenueber-genderneutraler-sprache/.

    • Search Google Scholar
    • Export Citation
  • Infratest dimap (2021) Weiter Vorbehalte gegen gendergerechte Sprache, www.infratest-dimap.de/umfragen-analysen/bundesweit/umfragen/aktuell/weiter-vorbehalte-gegen-gendergerechte-sprache/.

    • Search Google Scholar
    • Export Citation
  • Klann-Delius, G. (2005) Sprache und Geschlecht, Stuttgart: J.B. Metzler. doi: 10.1007/978-3-476-05072-4

  • Koeser, S. and Sczesny, S. (2014) Promoting gender-fair language, Journal of Language and Social Psychology, 33(5): 54860. doi: 10.1177/0261927x14541280

    • Search Google Scholar
    • Export Citation
  • Koeser, S., Kuhn, E.A. and Sczesny, S. (2015) Just reading? How gender-fair language triggers readers’ use of gender-fair forms, Journal of Language and Social Psychology, 34(3): 34357. doi: 10.1177/0261927x14561119

    • Search Google Scholar
    • Export Citation
  • Kriesi, H. (2012) Personalization of national election campaigns, Party Politics, 18(6): 82544. doi: 10.1177/1354068810389643

  • Maier, J. (2023) What factors explain the broadcasting of televised election debates? Empirical evidence from Germany, European Journal of Communication, 38(3): 27286. doi: 10.1177/02673231221123201

    • Search Google Scholar
    • Export Citation
  • Maier, J. (2024) Die Fernsehdebatten in der „heißen“ Phase der Bundestagswahl 2021: Nutzung, Wirkung auf die Wahlbeteiligung und Effekte auf das Wahlverhalten, in H. Schoen and B. Weßels (eds) Wahlen und Wähler. Analysen aus Anlass der Bundestagswahl 2021, Wiesbaden: Springer.

    • Search Google Scholar
    • Export Citation
  • Maier, J. and Faas, T. (2011) ‘Miniature campaigns’ in comparison: the German televised debates, 2002–09, German Politics, 20(1): 7591. doi: 10.1080/09644008.2011.554102

    • Search Google Scholar
    • Export Citation
  • Maier, J. and Jansen, C. (2017) When do candidates attack in election campaigns? Exploring the determinants of negative candidate messages in German televised debates, Party Politics, 23(5): 54959. doi: 10.1177/1354068815610966

    • Search Google Scholar
    • Export Citation
  • Maier, J. and Renner, A.-M. (2018) When a man meets a woman: comparing the use of negativity of male candidates in single- and mixed-gender televised debates, Political Communication, 35(3): 43349. doi: 10.1080/10584609.2017.1411998

    • Search Google Scholar
    • Export Citation
  • Maier, M., Adam, S. and Maier, J. (2012) The impact of identity and economic cues on citizens’ EU support: an experimental study on the effects of party communication in the run-up to the 2009 European Parliament elections, European Union Politics, 13(4): 580603. doi: 10.1177/1465116512453957

    • Search Google Scholar
    • Export Citation
  • McKinney, M.S. and Carlin, D.B. (2010) Political campaign debates, in L.L. Kaid (ed) Handbook of Political Communication Research, New York, NY: Routledge Taylor & Francis Group, pp 20334.

    • Search Google Scholar
    • Export Citation
  • McPherson, M., Smith-Lovin, L. and Cook, J.M. (2001) Birds of a feather: homophily in social networks, Annual Review of Sociology, 27: 41544. doi: 10.1146/annurev.soc.27.1.415

    • Search Google Scholar
    • Export Citation
  • Michaux, V., Méndez, J. and Apel, H. (2021) Mündlich Gendern? Gerne. Aber wie genau? Ergebnisse einer Akzeptanzuntersuchung zu Formen des Genderns in der Mündlichkeit, Sprachreport, 37(2): 3441. doi: 10.14618/sr-2-2021-mich

    • Search Google Scholar
    • Export Citation
  • Moineddin, R., Matheson, F.I. and Glazier, R.H. (2007) A simulation study of sample size for multilevel logistic regression models, BMC Medical Research Methodology, 7: 110. doi: 10.1186/1471-2288-7-34

    • Search Google Scholar
    • Export Citation
  • Müller, J. (2009) The impact of the socio-economic context on the Länder Parties’ policy positions, German Politics, 18(3): 36584. doi: 10.1080/09644000903055815

    • Search Google Scholar
    • Export Citation
  • Niemiec, R.M., Champine, V., Vaske, J.J. and Mertens, A. (2020) Does the impact of norms vary by type of norm and type of conservation behavior? A meta-analysis, Society & Natural Resources, 33(8): 102440. doi: 10.1080/08941920.2020.1729912

    • Search Google Scholar
    • Export Citation
  • Papke, L.E. and Wooldridge, J.M. (1996) Econometric methods for fractional response variables with an application to 401(k) plan participation rates, Journal of Applied Econometrics, 11(6): 61932. doi: 10.1002/(sici)1099-1255(199611)11:6<619::aid-jae418>3.0.co;2-1

    • Search Google Scholar
    • Export Citation
  • Parks, J.B. and Roberton, M.A. (2002) The gender gap in student attitudes toward sexist/nonsexist language: implications for sport management education, Journal of Sport Management, 16(3): 190208. doi: 10.1123/jsm.16.3.190

    • Search Google Scholar
    • Export Citation
  • Parks, J.B. and Roberton, M.A. (2004) Attitudes toward women mediate the gender effect on attitudes toward sexist language, Psychology of Women Quarterly, 28(3): 2339. doi: 10.1111/j.1471-6402.2004.00140.x

    • Search Google Scholar
    • Export Citation
  • Peskowitz, Z. (2019) Ideological signaling and incumbency advantage, British Journal of Political Science, 49(2): 46790. doi: 10.1017/s0007123416000557

    • Search Google Scholar
    • Export Citation
  • Prewitt-Freilino, J.L., Caswell, T.A. and Laakso, E.K. (2012) The gendering of language: a comparison of gender equality in countries with gendered, natural gender, and genderless languages, Sex Roles, 66(3): 26881. doi: 10.1007/s11199-011-0083-5

    • Search Google Scholar
    • Export Citation
  • Rimal, R.N. and Real, K. (2005) How behaviors are influenced by perceived norms, Communication Research, 32(3): 389414. doi: 10.1177/0093650205275385

    • Search Google Scholar
    • Export Citation
  • Schneider, J.G. (2021) Zum prekären Status sprachlicher Verbindlichkeit: Gendern im Deutschen, in J. Raab and J. Heck (eds) Prekäre Verbindlichkeiten: Studien an den Problemschwellen normativer Ordnungen, Wiesbaden: Springer VS, pp 1743.

    • Search Google Scholar
    • Export Citation
  • Schneider, J.G., Albert, G., Bast, J. and Maier, J. (2022) Geschlechtersensibler Sprachgebrauch im Wahlkampf? Eine diachrone Analyse von TV-Duellen in Deutschland, Zeitschrift für Angewandte Linguistik, 2022(76): 93123. doi: 10.1515/zfal-2022-2081

    • Search Google Scholar
    • Export Citation
  • Schroeder, A. (2016) Presidential Debates: Risky Business on the Campaign Trail, 3rd edn, New York, NY: Columbia University Press.

  • Schrott, P.R. and Lanoue, D.J. (1992) How to win a televised debate: candidate strategies and voter response in Germany, 1972–87, British Journal of Political Science, 22(4): 44567. doi: 10.1017/s0007123400006487

    • Search Google Scholar
    • Export Citation
  • Sczesny, S., Formanowicz, M. and Moser, F. (2016) Can gender-fair language reduce gender stereotyping and discrimination?, Frontiers in Psychology, 7: 25. doi: 10.3389/fpsyg.2016.00025

    • Search Google Scholar
    • Export Citation
  • Smaldino, P.E., Flamson, T.J. and McElreath, R. (2018) The evolution of covert signaling, Scientific Reports, 8(1): 4905. doi: 10.1038/s41598-018-22926-1

    • Search Google Scholar
    • Export Citation
  • Stahlberg, D. and Sczesny, S. (2001) Effekte des generischen Maskulinums und alternativer Sprachformen auf den gedanklichen Einbezug von Frauen, Psychologische Rundschau, 52(3): 13140. doi: 10.1026//0033-3042.52.3.131

    • Search Google Scholar
    • Export Citation
  • Stecker, C., Müller, J., Blätte, A. and Leonhardt, C. (2021) The evolution of gender-inclusive language. Evidence from the German Bundestag, 1949–2021. doi: 10.31219/osf.io/fcsmz

    • Search Google Scholar
    • Export Citation
  • Thiese, M.S., Ronna, B. and Ott, U. (2016) P value interpretations and considerations, Journal of Thoracic Disease, 8(9): 92831. doi: 10.21037/jtd.2016.08.16

    • Search Google Scholar
    • Export Citation
  • Van der Does, T., Galesic, M., Dunivin, Z.O. and Smaldino, P.E. (2022) Strategic identity signaling in heterogeneous networks, Proceedings of the National Academy of Sciences of the United States of America, 119(10): 2117898119. doi: 10.1073/pnas.2117898119

    • Search Google Scholar
    • Export Citation
  • Zifonun, G. (2018) Die demokratische Pflicht und das Sprachsystem: erneute Diskussion um einen geschlechter-gerechten Sprachgebrauch, Sprachreport, 4: 4456.

    • Search Google Scholar
    • Export Citation
  • Figure 1:

    Coefficient plots from generalized linear models estimating factors influencing candidates’ use of GSL (Models 1–4)

  • Figure 2:

    Predictive margins for the interaction between candidates’ gender and party affiliation

  • Figure 3:

    Coefficient plots from generalized linear models estimating factors influencing candidates’ use of GSL (Models 5–8)

  • Figure 4:

    Interaction plots visualizing the interaction between candidates’ gender and parties’ social position

  • Baum, C.F. (2008) Stata tip 63: modeling proportions, The Stata Journal: Promoting Communications on Statistics and Stata, 8(2): 299303. doi: 10.1177/1536867x0800800212

    • Search Google Scholar
    • Export Citation
  • Benoit, K., Bräuninger, T. and Debus, M. (2009) Challenges for estimating policy preferences: announcing an open access archive of political documents, German Politics, 18(3): 44154. doi: 10.1080/09644000903055856

    • Search Google Scholar
    • Export Citation
  • Benoit, W.L. (2014) Political Election Debates: Informing Voters about Policy and Character, Lanham, MD: Lexington Books.

  • Blake, C. and Klimmt, C. (2010) Geschlechtergerechte Formulierungen in Nachrichtentexten, Publizistik, 55(3): 289304.doi: 10.1007/s11616-010-0093-2

    • Search Google Scholar
    • Export Citation
  • Bonnin, J.E. and Coronel, A.A. (2021) Attitudes toward gender-neutral Spanish: acceptability and adoptability, Frontiers in Sociology, 6. doi: 10.3389/fsoc.2021.629616

    • Search Google Scholar
    • Export Citation
  • Bräuninger, T., Debus, M., Müller, J. and Stecker, C. (2020) Parteienwettbewerb in den Deutschen Bundesländern, 2nd edn, Wiesbaden: Springer VS.

    • Search Google Scholar
    • Export Citation
  • Bußmann, H. and Hellinger, M. (2003) Engendering female visibility in German, in M. Hellinger and H. Bußmann (eds) Gender across Languages. The Linguistic Representation of Women and Men, Vol. 3, Amsterdam and Philadelphia, PA: Benjamins, pp 14174.

    • Search Google Scholar
    • Export Citation
  • Chartrand, T.L. and Bargh, J.A. (1999) The chameleon effect: the perception–behavior link and social interaction, Journal of Personality and Social Psychology, 76(6): 893910. doi: 10.1037//0022-3514.76.6.893

    • Search Google Scholar
    • Export Citation
  • Chen, K., Cheng, Y., Berkout, O. and Lindhiem, O. (2017) Analyzing proportion scores as outcomes for prevention trials: a statistical primer, Prevention Science: The Official Journal of the Society for Prevention Research, 18(3): 31221. doi: 10.1007/s11121-016-0643-6

    • Search Google Scholar
    • Export Citation
  • Cialdini, R.B., Kallgren, C.A. and Reno, R.R. (1991) A focus theory of normative conduct: a theoretical refinement and reevaluation of the role of norms in human behavior, Advances in Experimental Social Psychology. Advances in Experimental Social Psychology, 24: 20134. doi: 10.1016/S0065-2601(08)60330-5

    • Search Google Scholar
    • Export Citation
  • Damore, D.F. (2002) Candidate strategy and the decision to go negative, Political Research Quarterly, 55(3): 669. doi: 10.2307/3088036

  • Dolan, K. (2008) Is there a ‘gender affinity effect’ in American politics?, Political Research Quarterly, 61(1): 7989. doi: 10.1177/1065912907307518

    • Search Google Scholar
    • Export Citation
  • Eisenberg, P. (2018) Misshandlung der Sprache. Gleichberechtigung der Geschlechter ist wünschenswert, doch das generische Maskulinum darf nicht angetastet werden. Eine Streitschrift, Tagesspiegel, 6 August.

    • Search Google Scholar
    • Export Citation
  • Elsen, H. (2020) Gender – Sprache – Stereotype: Geschlechtersensibilität in Alltag und Unterricht, Tübingen: Narr Francke Attempto Verlag.

    • Search Google Scholar
    • Export Citation
  • Esser, F. and Pfetsch, B. (2020) Political communication, in D. Caramani (ed) Comparative Politics, Oxford: Oxford University Press, pp 33658.

    • Search Google Scholar
    • Export Citation
  • Fabry, V. (2023) ‘To gender or not to gender?’ Examining the usage of gender-inclusive language by German politicians in parliament and on Twitter, paper presented at the Annual Conference of the International Communication Association (ICA), Toronto.

    • Search Google Scholar
    • Export Citation
  • Formanowicz, M., Bedynska, S., Cisłak, A., Braun, F. and Sczesny, S. (2013) Side effects of gender-fair language: how feminine job titles influence the evaluation of female applicants, European Journal of Social Psychology, 43(1): 6271. doi: 10.1002/ejsp.1924

    • Search Google Scholar
    • Export Citation
  • Geese, S. and Hess, C. (2021) Nutzung und Bewertung medialer Wahlinformationen. Ergebnisse einer Repräsentativbefragung zur Bundestagswahl 2021, Media Perspektiven, 12: 61024.

    • Search Google Scholar
    • Export Citation
  • Geiger, A.W. and Graf, N. (2019) About One-in-five U.S. Adults Know Someone who Goes by a Gender-Neutral Pronoun, Washington, DC: Pew Research Center, www.pewresearch.org/short-reads/2019/09/05/gender-neutral-pronouns/.

    • Search Google Scholar
    • Export Citation
  • Gross, M. and Debus, M. (2018) Does EU regional policy increase parties’ support for European integration?, West European Politics, 41(3): 594614. doi: 10.1080/01402382.2017.1395249

    • Search Google Scholar
    • Export Citation
  • Gustafsson Sendén, M., Bäck, E.A. and Lindqvist, A. (2015) Introducing a gender-neutral pronoun in a natural gender language: the influence of time on attitudes and behavior, Frontiers in Psychology, 6. doi: 10.3389/fpsyg.2015.00893

    • Search Google Scholar
    • Export Citation
  • Hartman, T.K., Pattie, C. and Johnston, R. (2017) Learning on the job? Adapting party campaign strategy to changing information on the local political context, Electoral Studies, 49: 12835. doi: 10.1016/j.electstud.2017.06.005

    • Search Google Scholar
    • Export Citation
  • Heise, E. (2000) Sind Frauen mitgemeint? Eine empirische Untersuchung zum Verständnis des generischen Maskulinums und seiner Alternativen, Sprache & Kognition, 19(1–2): 313. doi: 10.1024//0253-4533.19.12.3

    • Search Google Scholar
    • Export Citation
  • Hellinger, M. and Bußmann, H. (2001) Gender across languages. The linguistic representation of women and men, in M. Hellinger and H. Bußmann (eds) Gender across Languages. The Linguistic Representation of Women and Men, Vol. 1, Amsterdam and Philadelphia, PA: Benjamins, pp 125.

    • Search Google Scholar
    • Export Citation
  • Hough, D. and Keith, D. (2019) The German Left Party: a case of pragmatic populism, in G. Katsambekis and A. Kioupkiolis (eds) The Populist Radical Left in Europe, Abingdon and New York, NY: Routledge, pp 12944.

    • Search Google Scholar
    • Export Citation
  • Infratest dimap (2020) Vorbehalte gegenüber genderneutraler Sprache, www.infratest-dimap.de/umfragen-analysen/bundesweit/umfragen/aktuell/vorbehalte-gegenueber-genderneutraler-sprache/.

    • Search Google Scholar
    • Export Citation
  • Infratest dimap (2021) Weiter Vorbehalte gegen gendergerechte Sprache, www.infratest-dimap.de/umfragen-analysen/bundesweit/umfragen/aktuell/weiter-vorbehalte-gegen-gendergerechte-sprache/.

    • Search Google Scholar
    • Export Citation
  • Klann-Delius, G. (2005) Sprache und Geschlecht, Stuttgart: J.B. Metzler. doi: 10.1007/978-3-476-05072-4

  • Koeser, S. and Sczesny, S. (2014) Promoting gender-fair language, Journal of Language and Social Psychology, 33(5): 54860. doi: 10.1177/0261927x14541280

    • Search Google Scholar
    • Export Citation
  • Koeser, S., Kuhn, E.A. and Sczesny, S. (2015) Just reading? How gender-fair language triggers readers’ use of gender-fair forms, Journal of Language and Social Psychology, 34(3): 34357. doi: 10.1177/0261927x14561119

    • Search Google Scholar
    • Export Citation
  • Kriesi, H. (2012) Personalization of national election campaigns, Party Politics, 18(6): 82544. doi: 10.1177/1354068810389643

  • Maier, J. (2023) What factors explain the broadcasting of televised election debates? Empirical evidence from Germany, European Journal of Communication, 38(3): 27286. doi: 10.1177/02673231221123201

    • Search Google Scholar
    • Export Citation
  • Maier, J. (2024) Die Fernsehdebatten in der „heißen“ Phase der Bundestagswahl 2021: Nutzung, Wirkung auf die Wahlbeteiligung und Effekte auf das Wahlverhalten, in H. Schoen and B. Weßels (eds) Wahlen und Wähler. Analysen aus Anlass der Bundestagswahl 2021, Wiesbaden: Springer.

    • Search Google Scholar
    • Export Citation
  • Maier, J. and Faas, T. (2011) ‘Miniature campaigns’ in comparison: the German televised debates, 2002–09, German Politics, 20(1): 7591. doi: 10.1080/09644008.2011.554102

    • Search Google Scholar
    • Export Citation
  • Maier, J. and Jansen, C. (2017) When do candidates attack in election campaigns? Exploring the determinants of negative candidate messages in German televised debates, Party Politics, 23(5): 54959. doi: 10.1177/1354068815610966

    • Search Google Scholar
    • Export Citation
  • Maier, J. and Renner, A.-M. (2018) When a man meets a woman: comparing the use of negativity of male candidates in single- and mixed-gender televised debates, Political Communication, 35(3): 43349. doi: 10.1080/10584609.2017.1411998

    • Search Google Scholar
    • Export Citation
  • Maier, M., Adam, S. and Maier, J. (2012) The impact of identity and economic cues on citizens’ EU support: an experimental study on the effects of party communication in the run-up to the 2009 European Parliament elections, European Union Politics, 13(4): 580603. doi: 10.1177/1465116512453957

    • Search Google Scholar
    • Export Citation
  • McKinney, M.S. and Carlin, D.B. (2010) Political campaign debates, in L.L. Kaid (ed) Handbook of Political Communication Research, New York, NY: Routledge Taylor & Francis Group, pp 20334.

    • Search Google Scholar
    • Export Citation
  • McPherson, M., Smith-Lovin, L. and Cook, J.M. (2001) Birds of a feather: homophily in social networks, Annual Review of Sociology, 27: 41544. doi: 10.1146/annurev.soc.27.1.415

    • Search Google Scholar
    • Export Citation
  • Michaux, V., Méndez, J. and Apel, H. (2021) Mündlich Gendern? Gerne. Aber wie genau? Ergebnisse einer Akzeptanzuntersuchung zu Formen des Genderns in der Mündlichkeit, Sprachreport, 37(2): 3441. doi: 10.14618/sr-2-2021-mich

    • Search Google Scholar
    • Export Citation
  • Moineddin, R., Matheson, F.I. and Glazier, R.H. (2007) A simulation study of sample size for multilevel logistic regression models, BMC Medical Research Methodology, 7: 110. doi: 10.1186/1471-2288-7-34

    • Search Google Scholar
    • Export Citation
  • Müller, J. (2009) The impact of the socio-economic context on the Länder Parties’ policy positions, German Politics, 18(3): 36584. doi: 10.1080/09644000903055815

    • Search Google Scholar
    • Export Citation
  • Niemiec, R.M., Champine, V., Vaske, J.J. and Mertens, A. (2020) Does the impact of norms vary by type of norm and type of conservation behavior? A meta-analysis, Society & Natural Resources, 33(8): 102440. doi: 10.1080/08941920.2020.1729912

    • Search Google Scholar
    • Export Citation
  • Papke, L.E. and Wooldridge, J.M. (1996) Econometric methods for fractional response variables with an application to 401(k) plan participation rates, Journal of Applied Econometrics, 11(6): 61932. doi: 10.1002/(sici)1099-1255(199611)11:6<619::aid-jae418>3.0.co;2-1

    • Search Google Scholar
    • Export Citation
  • Parks, J.B. and Roberton, M.A. (2002) The gender gap in student attitudes toward sexist/nonsexist language: implications for sport management education, Journal of Sport Management, 16(3): 190208. doi: 10.1123/jsm.16.3.190

    • Search Google Scholar
    • Export Citation
  • Parks, J.B. and Roberton, M.A. (2004) Attitudes toward women mediate the gender effect on attitudes toward sexist language, Psychology of Women Quarterly, 28(3): 2339. doi: 10.1111/j.1471-6402.2004.00140.x

    • Search Google Scholar
    • Export Citation
  • Peskowitz, Z. (2019) Ideological signaling and incumbency advantage, British Journal of Political Science, 49(2): 46790. doi: 10.1017/s0007123416000557

    • Search Google Scholar
    • Export Citation
  • Prewitt-Freilino, J.L., Caswell, T.A. and Laakso, E.K. (2012) The gendering of language: a comparison of gender equality in countries with gendered, natural gender, and genderless languages, Sex Roles, 66(3): 26881. doi: 10.1007/s11199-011-0083-5

    • Search Google Scholar
    • Export Citation
  • Rimal, R.N. and Real, K. (2005) How behaviors are influenced by perceived norms, Communication Research, 32(3): 389414. doi: 10.1177/0093650205275385

    • Search Google Scholar
    • Export Citation
  • Schneider, J.G. (2021) Zum prekären Status sprachlicher Verbindlichkeit: Gendern im Deutschen, in J. Raab and J. Heck (eds) Prekäre Verbindlichkeiten: Studien an den Problemschwellen normativer Ordnungen, Wiesbaden: Springer VS, pp 1743.

    • Search Google Scholar
    • Export Citation
  • Schneider, J.G., Albert, G., Bast, J. and Maier, J. (2022) Geschlechtersensibler Sprachgebrauch im Wahlkampf? Eine diachrone Analyse von TV-Duellen in Deutschland, Zeitschrift für Angewandte Linguistik, 2022(76): 93123. doi: 10.1515/zfal-2022-2081

    • Search Google Scholar
    • Export Citation
  • Schroeder, A. (2016) Presidential Debates: Risky Business on the Campaign Trail, 3rd edn, New York, NY: Columbia University Press.

  • Schrott, P.R. and Lanoue, D.J. (1992) How to win a televised debate: candidate strategies and voter response in Germany, 1972–87, British Journal of Political Science, 22(4): 44567. doi: 10.1017/s0007123400006487

    • Search Google Scholar
    • Export Citation
  • Sczesny, S., Formanowicz, M. and Moser, F. (2016) Can gender-fair language reduce gender stereotyping and discrimination?, Frontiers in Psychology, 7: 25. doi: 10.3389/fpsyg.2016.00025

    • Search Google Scholar
    • Export Citation
  • Smaldino, P.E., Flamson, T.J. and McElreath, R. (2018) The evolution of covert signaling, Scientific Reports, 8(1): 4905. doi: 10.1038/s41598-018-22926-1

    • Search Google Scholar
    • Export Citation
  • Stahlberg, D. and Sczesny, S. (2001) Effekte des generischen Maskulinums und alternativer Sprachformen auf den gedanklichen Einbezug von Frauen, Psychologische Rundschau, 52(3): 13140. doi: 10.1026//0033-3042.52.3.131

    • Search Google Scholar
    • Export Citation
  • Stecker, C., Müller, J., Blätte, A. and Leonhardt, C. (2021) The evolution of gender-inclusive language. Evidence from the German Bundestag, 1949–2021. doi: 10.31219/osf.io/fcsmz

    • Search Google Scholar
    • Export Citation
  • Thiese, M.S., Ronna, B. and Ott, U. (2016) P value interpretations and considerations, Journal of Thoracic Disease, 8(9): 92831. doi: 10.21037/jtd.2016.08.16

    • Search Google Scholar
    • Export Citation
  • Van der Does, T., Galesic, M., Dunivin, Z.O. and Smaldino, P.E. (2022) Strategic identity signaling in heterogeneous networks, Proceedings of the National Academy of Sciences of the United States of America, 119(10): 2117898119. doi: 10.1073/pnas.2117898119

    • Search Google Scholar
    • Export Citation
  • Zifonun, G. (2018) Die demokratische Pflicht und das Sprachsystem: erneute Diskussion um einen geschlechter-gerechten Sprachgebrauch, Sprachreport, 4: 4456.

    • Search Google Scholar
    • Export Citation
Jennifer Bast University of Kaiserslautern-Landau, Germany

Search for other papers by Jennifer Bast in
Current site
Google Scholar
Close
,
Jürgen Maier University of Kaiserslautern-Landau, Germany

Search for other papers by Jürgen Maier in
Current site
Google Scholar
Close
,
Georg Albert University of Kaiserslautern-Landau, Germany

Search for other papers by Georg Albert in
Current site
Google Scholar
Close
, and
Jan Georg Schneider University of Kaiserslautern-Landau, Germany

Search for other papers by Jan Georg Schneider in
Current site
Google Scholar
Close

Content Metrics

May 2022 onwards Past Year Past 30 Days
Abstract Views 525 525 0
Full Text Views 212 212 73
PDF Downloads 183 183 67

Altmetrics

Dimensions