What makes the difference? Social media platforms and party characteristics as contextual factors for political parties’ use of populist political communication

Political populism is a multifaceted, ambivalent phenomenon being discussed in various academic disciplines, such as political science, sociology, and communication studies. According to Bossetta and Husted (2017), it has been described as an ideology, a style, a movement, and a logic (for a detailed discussion on different analytical approaches to the study of populism, see also Moffitt, 2016). Populism is a normatively loaded concept. Some authors consider populism an existential threat to liberalism (though not to democracy, as such), since it challenges the political elites and the existing representative institutions (Mouffe, 2019; Müller, 2016). Other authors, in contrast, highlight the emancipatory, progressive aims of populism and see it as a corrective for democracy, and therefore an essential part of the democratic process and democratisation (Gerbaudo, 2018; Rovira Kaltwasser, 2012). What differentiates populism from anti-system and extremist movements is that it does not aim to abolish democracy but rather “challenges democracies from within the democratic system” (Ernst et al., 2017: 1348), which makes it less of an existential threat, particularly in stable democracies (de Vreese et al., 2018), such as Norway.

While we acknowledge the importance of these research traditions, our study builds on another, more recent tradition rooted in communication studies that focuses on populist communication as a communicative phenomenon. It takes a minimal, communication-centred approach when defining populist communication. According to this understanding (and different from many definitions of populism in political science), populist communication is not related to any one political ideology but can rather be combined with several, for example, right-wing (e.g., nationalism) and left-wing (e.g., socialism, communism) ideologies (Mudde, 2004).

Our study is theoretically grounded in a well-established framework developed by Reinemann and colleagues (2017), based on Jagers and Walgrave (2007), which places the spotlight on the content of political messages rather than differentiating between populist and non-populist political actors or ideologies. Their minimal definition (Mudde, 2004) allows us to bypass normative views of the phenomenon and empirically determine to which degree political actors use populist communication, independent of their ideology (de Vreese et al., 2018). Populist communication can be utilised by all actors across the party spectrum (Blassnig, 2021; Engesser, Ernst et al., 2017). Reinemann and colleagues (2017) defined three core elements that distinguish populist from non-populist communication: people-centrism, anti-elitism, and exclusion of out-groups.

People-centrism refers to people and is considered the most central, “minimal defining element” (de Vreese et al., 2018: 427) of populist communication (Reinemann et al., 2017). Mudde (2004: 546) called the people an “imagined community [which is] neither real nor all-inclusive, but [is] in fact a mythical and constructed sub-set of the whole population”. A common populist claim is to give the people a voice and return power to them as true sovereigns (Canovan, 1999; Gerbaudo, 2018). Populist communication creates the imagination of the people as a completely homogeneous group (Jagers & Walgrave, 2007), but what exactly is meant by “the people” differs. The inherent vagueness of the term is part of the secret of success: Proponents of different views can find themselves in the same populist messages (Maurer, 2022), which makes populist communication an excellent strategy for reaching out to large audiences (Reinemann et al., 2017). Canovan (1999), for example, distinguished between three different kinds of people: “the united people”, who stand together against the power-holding bloc of the ruling, privileged elite; “the common people”, as opposed to the higher educated, aloof cultural elites who have lost contact with “the common people”; and “our people”, or the native people belonging to “our” nation. While the people, considered as the in-group, are opposed to another group, the nature or composition of this other group varies, too. Reinemann and colleagues’ (2017) conceptualisation allows for measuring people-centrism without any other group being explicitly mentioned, which Jagers and Walgrave (2007) referred to as “empty populism”. This is methodologically advantageous for measuring populist communication on social media, where its elements appear isolated from one another (Bene et al., 2023; Engesser, Ernst et al., 2017). In line with Canovan (1999), there are typically two kinds of other groups to which the people are opposed: either any elite (as opposed to “the united people” and “the common people”), or any out-group (as opposed to “our people”). These are addressed by the two other core elements of populist communication (Reinemann et al., 2017).

In anti-elitism, since “the elite” is assumed to oppose the will of the (however defined) people, populist communication attacks it. Hellström (2023) conceptualised the contrast between “the people” and “the elite” as the vertical dimension of what he calls “the populist divide”. As another vague term, the elite often gains its contours solely through juxtaposition with the people (Mudde, 2004; Rovira Kaltwasser, 2012). At the same time (and as a consequence thereof), there can be different kinds of elites, for example, the political elite (meaning mainly the mainstream parties in the respective country), the media elite (legacy media in contrast to alternative media and often social media), the economic elite, the moral elite, and the cultural elite (Hellström, 2023). Depending on the political ideology populist communication is related to, the conflict between the people and the elite will centre on different matters, and different elites will be attacked. Left-wing populism tends to attack economic elites, while right-wing populism tends to attack the political, media, or scientific elite (Maurer, 2022).

With exclusion of out-groups, while the elite is located “above” the people in the social hierarchy, the other assumed opponent is located “beside” or even “below” the people: The so-called dangerous others (one or several out-groups) must be excluded since they are considered a threat to the people (Engesser, Fawzi, & Larsson, 2017). Hellström (2023: 5) conceptualised the “articulated differences between ‘the people’, who ‘naturally’ belong here, and ‘the non-people’ (the Other), who do not” as the horizontal dimension of the populist divide (which he, however, called people-centrism). The “othering”, based on social identity and social comparisons (Reinemann et al., 2017), sharpens the contours of the people as the in-group. The “others” are considered a homogeneous group that is “unjustly favored by the elite or even their partner in a conspiracy against the people” (Engesser, Ernst et al., 2017: 1112). Depending on the political ideology with which populist communication is combined, different groups are perceived as “others”, such as immigrants and ethnic, sexual, or religious minorities (Reinemann et al., 2017; see also Hellström, 2023). According to Rovira Kaltwasser (2012), this exclusionary element of populism is more typical in Europe (the same seems to apply to the US; Hopster, 2021), while populism in Latin America aims more strongly at including underprivileged groups.

These three core elements can appear both separately and in various combinations in political communication (Jagers & Walgrave, 2007). This approach makes it possible not only to differentiate between populist and non-populist communication but also between the degree of populist communication and between different types of populist communication, thus allowing for a differentiated measurement thereof (de Vreese et al., 2018).

Reinemann and colleagues’ (2017) approach is not without alternatives. For example, Engesser, Ernst, Esser, and Büchel (2017) conceptualised populist communication as an emphasis on the sovereignty of the people, advocacy for the people, attacks against the elites, ostracising others, and invoking the heart-land. However, these dimensions are, upon closer inspection – even though they are further differentiated – very similar to the three core elements put forth by Reinemann and colleagues (2017). Ernst and colleagues (2017, 2019) conceptualised populist communication as consisting of three populist key messages (where they also use people-centrism and anti-elitism but replace exclusion of out-groups with restoring sovereignty) combined with three populist communication styles (negativity, emotionality, and sociability). However, it has been empirically shown that the exclusion of out-groups relates more strongly to right-wing populism, while anti-elitism is more closely related to left-wing populism (Engesser, Ernst et al., 2017; Schmuck & Hameleers, 2020). Therefore, we consider the exclusion of out-groups necessary to include, since it allows for a more balanced measurement of populist communication.

There is widespread agreement in the literature that populist communication and social media are closely related. Social media do not cause populism, but they foster the spread of populist communication (Maurer, 2022) due to the specific opportunity structures they provide (Engesser, Fawzi, & Larsson, 2017; similarly described as affordances by Hopster, 2021). Particularly three opportunity structures make using populist communication on social media strategically advantageous for political actors: direct communication, attention economy, and user metrics.

First, the technological infrastructure of social media allows direct communication between political actors and citizens which resonates with people-centrism (Engesser, Fawzi, & Larsson, 2017) and facilitates populist communication. Being able to spread their messages and bypass the media’s critical editorial filters can be favourable for all political actors. However, it is particularly favourable for populists, whose positions are often attacked by the news media (but are well suited to be distributed by means of populist communication) and whose supporters might be sceptical toward, and therefore difficult to reach via, the news media (Hopster, 2021).

Second, some authors relate populist communication to stylistic characteristics, such as negativity, emotionality, simplifying language (Ernst et al., 2019; Jacobs et al., 2020; Sandberg et al., 2022), and sensationalism (Hopster, 2021) – in other words, an attention economy. The popularity bias of social media algorithms favours content with such features that catch the attention of many users and generate user engagement (e.g., clicks, likes, emotional reactions, shares, comments). These patterns result in higher visibility and virality of populist compared with non-populist political content, which in turn creates even more attention for populist communication. This applies all the more, as content that is successful on social media is often reported in the news media, which increases its reach beyond the platforms.

Lastly, from the importance of user engagement on social media follows their ability to (allegedly) serve as the people’s voice, their rally (Gerbaudo, 2018), and an expression of the “general will” (Hopster, 2021), which relates to people-centrism. Even though user metrics on social media are by no means representative of the entire population (Magin, 2022), they are useful for campaign planning, since they are generated in real time and provide an incidence of which issues people currently care about (Hopster, 2021).

While these opportunity structures foster the spread of populist communication on social media in general, research shows that how well social media and populist communication interact depends on contextual factors. However, research into these contextual factors is still in its early stages (Schmuck & Hameleers, 2020). To shed light thereon, we investigated how the platforms used (Facebook, Instagram, and Twitter) and two party characteristics (populist vs. non-populist parties; party ideology) affected the use of populist communication. All three factors have been discussed in the literature but without any definite answers yet.

Concerning platforms, it is widely assumed (Hopster, 2021), and proven by empirical research (Blassnig, 2021; Schmuck & Hameleers, 2020), that the different opportunity structures of various social media platforms fuel the use of populist communication differently. Ernst and colleagues (2017) and Schmuck and Hameleers (2020) have shown that populist communication has been more widespread on Facebook than on Twitter. Furthermore, Jacobs and colleagues (2020: 628) found that politicians used populist communication differently on these two platforms, leading to the conclusion that “it is indeed important not to lump different platforms together under the header ‘social media’: the differences in user base and architecture matter empirically”. These findings show the importance of systematic cross-platform comparisons (Schwartz et al., 2022), also because not all parties use all platforms. However, so far, many studies have investigated only one platform – mainly Facebook (Bene et al., 2023; Kalsnes, 2019; Mazzoleni & Bracciale, 2018; Sandberg et al., 2022; Zulianello et al., 2018) and, more rarely, Twitter (Bracciale & Martella, 2017; Maurer & Diehl, 2020; Waisbord & Amado, 2017). Although Instagram has become a central platform in election campaigns and provides opportunity structures that distinguish it significantly from other platforms (Larsson, 2021), content analyses on populist communication there are so far rare and focus mainly exclusively on visual representations of populism (Bast, 2021; Moffitt, 2022). A few studies have compared populist communication on Facebook and Twitter (Bracciale et al., 2021; Engesser, Ernst et al., 2017; Ernst et al., 2017, 2019; Jacobs et al., 2020), but comparisons across more platforms are missing. An exception is Baldwin-Philippi (2019), who qualitatively investigated the use of people-centrism on Facebook, Twitter, and Instagram but neglected other dimensions of populist communication. This deficient state of cross-platform comparisons leads to our first research question:

RQ1. How widespread was populist communication in political parties’ posts on their official Facebook, Instagram, and Twitter accounts during the 2021 Norwegian election campaign?

Moreover, we investigated how two party characteristics affected the use of populist communication: populist versus non-populist parties; and party ideology. Research on the connection between populist communication and social media has so far strongly focused on so-called populist actors (Järviniemi, 2022; for examples, see Kalsnes, 2019; Sandberg et al., 2022; Waisbord & Amado, 2017) for whom social media seem to provide a particularly favourable communication space (Gerbaudo, 2018; Hopster, 2021). However, this research direction implicitly presupposes that populist actors make particular use of populist communication – which is not necessarily the case (Järviniemi, 2022; Lilleker & Balaban, 2021). Populist communication on social media can be an important strategic tool (Kalsnes, 2019) across the party spectrum (Bene et al., 2023), particularly in election campaigns when parties are forced to reach out to the electorate. Recently, the number of content analyses comparing the use of populist communication by populist and non-populist actors has been growing, but they have revealed ambiguous results. Some studies find that populist actors use a higher degree of populist communication (Ernst et al., 2019; Gründl, 2022; Järviniemi, 2022; Zulianello et al., 2018), while others show that being populist or not does not alone decide if and how much populist communication an actor uses (Bene et al., 2023; Engesser, Ernst et al., 2017; Lilleker & Balaban, 2021). From this unclear picture, we derive the following research question:

RQ2. How much did the degree of populist communication differ between populist and non-populist political parties?

When it comes to party ideology, we are interested in how it affects 1) the degree and 2) the type of populist communication. Concerning the degree of populist communication, content analyses have consistently and across different countries shown that political actors with an ideology located more toward the fringes of the political-party spectrum use a higher degree of populist communication compared with more moderate actors (Engesser, Ernst et al., 2017; Ernst et al., 2017; Lilleker & Balaban, 2021; Mazzoleni & Bracciale, 2018; Schmuck & Hameleers, 2020). This leads to our first hypothesis:

H1. Parties closer to the fringes of the political-party spectrum used a higher degree of populist communication on their official social media accounts.

When turning to the type of populist communication, we looked at how much the three core elements of populist communication (Reinemann et al., 2017) were used. In a comparative content analysis of four countries, Engesser, Ernst, Esser, and Büchel (2017) showed that populist communication often appeared in a fragmented form on social media; that is, its elements appeared isolated from one another. Moreover, the frequency with which the elements in total are used differs (Lilleker & Balaban, 2021; Schmuck & Hameleers, 2020). However, parties with different political ideologies have been shown to use the elements of populist communication differently (Maurer, 2022; Reinemann et al., 2017): Anti-elitism has an affinity with left-wing populism, often attacking economic elites, while the exclusion of out-groups tends to be used by right-wing populism, often attacking minorities (Engesser, Ernst et al., 2017; Schmuck & Hameleers, 2020). Therefore, we hypothesise:

H2. Depending on their political ideology, political parties used different types of populism on their official social media accounts.

To test our hypotheses and answer our research questions, we conducted a standardised content analysis of posts published on the official Facebook, Instagram, and Twitter accounts of the nine established parties that were elected to the Norwegian national parliament in 2021, as well as the official accounts of their party leaders. In total, we collected 2,599 posts from the four weeks prior to election day (17 August–13 September 2021). The post content and metadata were collected using the CrowdTangle API.3 for Facebook and Instagram and the Twitter API for Academic Research for Twitter. At the time of coding, five posts had already been deleted: therefore, we could code the content of 2,594 posts. We decided to exclude the 76 Facebook posts from the Patient Focus [Pasientfokus] movement since it is not registered as a party, and they had no accounts on Instagram or Twitter. This left us with 2,518 posts (Facebook: 1,030; Instagram: 576; Twitter: 912). Table A1 in the Appendix provides an overview of the sample. Although Norwegian election campaigns may be personalised in the sense that the party leaders represent the parties in images and social media posts (Enli & Skogerbø, 2013), the election system is party-centred, and the campaigns of parties and party leaders are jointly planned and organised (Skogerbø & Karlsen, 2014). Therefore, we lump together the accounts of parties and party leaders and treat both as campaign materials for the parties.

This study builds on the international comparative project “Digital Election Campaigning Worldwide” (DigiWorld),

The project can be found at the following website: https://digidemo.ifkw.lmu.de/digiworld/

To measure populist communication, we coded the three core elements suggested by Reinemann and colleagues (2017) – people-centrism, anti-elitism, and the exclusion of out-groups – by means of seven categories, all of which were coded binary as present (1) or not present (0), independent from one another. Complete category descriptions can be found in the Appendix.

People-centrism comprised references to the “pure people” (c1) and was coded when a national community or the unity of people were addressed in a post. Indicators such as “(will of the) people”, “our nation”, and “the citizens” were coded as means to generate a sense of community in some way. For coding people-centrism, however, it was not enough that those terms were mentioned. Rather, the post had to create a contrast between “the entire people” (for which the party/actor stood) and “the elite/the others”.

Anti-elitism captured whether a post blamed any elite (e.g., political elite, bureaucrats, economic elites, the news media, supranational elites) (c2.1), questioned the elite’s legitimacy to make decisions (c2.2), called for resistance against the elite (c2.3), or accused the elite of betraying the people (c2.4).

Exclusion of out-groups comprised posts calling for the exclusion of ethnic or cultural “others” (c3.1) and political “others” (c3.2).

To measure the degree of populist communication contained in the posts, we computed sub-indices for anti-elitism [ae = (c2.1 + c2.2 + c2.3 + c2.4) / 4] and exclusion of out-groups [eo = (c3.1 + c3.2) / 2]. People-centrism [pc = c1] which consisted of only one variable, was kept. These three sub-indices were combined into an overarching additive populist communication index [pi = (pc + ae + eo) / 3]. All sub-indices and the overall index were standardised to a value ranging from 0 (no populist communication at all; no category coded as present) to 1 (highest degree of populist communication; all seven categories coded as present).

All posts were coded by three student coders who underwent intensive training in the joint English codebook. We tested intercoder reliability based on 140 randomly selected posts (5.6% of the full sample) using Brennan and Prediger’s kappa, which is chance-corrected and more robust than Krippendorff’s alpha regarding variables with a skewed distribution (Quarfoot & Levine, 2016). In addition, we report the “raw” agreement levels (% agreement). Reliability was sufficient for all categories (see Table A2 in the Appendix).

We base our analyses on the sub-indices for the three elements of populist communication to investigate the type of populist communication, as well as the overall index that allows us to determine the degree of populist communication used in the posts. Analyses were performed using a series of statistical techniques. Given the skewed nature of the indices used, we opted for non-parametric analysis methods.

To answer the first research question on the degree of populist communication used on the three platforms, we investigated how strongly the three elements of populist communication were used across the three platforms (see Table 2). Overall, the Norwegian political parties were rather restrained from using populist communication: Only 8.5 per cent of all posts contained any form of populist communication (not shown in Table 2). Anti-elitism emerged as the comparatively most common element. Comparing the three platforms across the three dimensions, Facebook emerged as the most populist platform. Reading each row in Table 2 from left to right, the percentages decrease as we move from Facebook to Twitter via Instagram. Thus, while the presence of populist communication is limited across all three platforms, we nevertheless see some discrepancies between them, which might be related to the platforms’ opportunity structures. Longer posts provide more space, possibly also to use a higher degree of populist communication. Typically, posts on Facebook are more text-heavy than both those on the visual-dominated platform Instagram and on the micro-blogging service Twitter, which limits tweets to a maximum of 280 characters. An additional analysis of the relationship between post length and the overarching additive index using Spearman’s rho suggests a significant correlation of .13 (p < .001) – not very strong, but still indicative that longer posts tend to include a higher degree of populist communication.

In order to answer our second research question, we investigated how party characteristics affected the use of populist communication, starting with the comparison between populist (i.e., the Progress Party, the only populist party in our sample) versus non-populist (i.e., the remaining) parties. For Facebook, the mean overall index for the Progress Party (M: 0.048, Sd: 0.118) was higher than the combined mean of all other parties (M: 0.022, Sd: 0.078). A Wilcoxon rank-sum test indicated that this difference was statistically significant (p = 0.024). For Instagram, the mean reported for the Progress Party emerged as lower (M: 0.006, Sd: 0.038) than the mean for all other actors combined (M: 0.016, Sd: 0.066), but according to a Wilcoxon rank-sum test for these means (p = 0.137), this difference is not statistically significant. On Twitter, the Progress Party was not active during the study period.

The first hypothesis postulates that a party ideology toward the fringes of the political-party spectrum leads to a higher degree of populist communication. Figure 1 details the mean of our additive index for each studied party on each platform, which allows us to compare the degrees of populist communication across platforms and parties. The error bars indicate the standard deviations of said index for each party on each platform.

FIGURE 1

Starting with Facebook (represented by purple bars in Figure 1), a Kruskal–Wallis H test suggested significant differences among the reported means. Post hoc testing using Dunn’s test revealed the highest mean reported – for the Progress Party (M: 0.048, Sd: 0.119) – to be significantly different (p < .005) from the lowest means reported – for the Christian People’s Party [Kristelig Folkpartie] (M: 0.003, Sd: 0.016) and the Liberal Party [Venstre] (M: 0.004, Sd: 0.025). The Labour Party (M: 0.035, Sd: 0.107) emerged as similarly significantly different (p < .005) from the Christian People’s and Liberal Parties. Thus, while the differences between them were not statistically significant, the top populist parties, according to their communication on Facebook, were the Progress Party and the Labour Party, followed by the Red Party (M: 0.036, Sd: 0.092) and the Socialist Left Party (M: 0.032, Sd: 0.093).

Second, for Instagram (indicated by orange bars in Figure 1), the relative scarcity compared with the Facebook posts suggests that Instagram was not prioritised when it comes to the use of populist communication, as defined here. Indeed, while all parties made use of populist communication on Facebook, the results suggest that four parties – Conservative, Christian People’s, Centre, and Liberal – did not engage on Instagram in this way at all. Among the parties that did engage, the Labour (M: 0.027, Sd: 0.094), Red (M: 0.026, Sd: 0.078), and Socialist Left (M: 0.022, Sd: 0.077) Parties were more populist with regard to our additive populist communication index. Interestingly, while the Progress Party emerged with the highest mean index value for Facebook, the mean reported for their Instagram account (M: 0.007, Sd: 0.038) suggests different content strategies for different platforms. However, a Kruskal-Wallis H test found no significant differences among the means reported for Instagram (p > 0.05), suggesting that while we can discern some tendencies regarding usage, we should be careful not to overstate them.

Finally, even fewer parties – namely only four – spread populist communication on Twitter: Red (M: 0.014, Sd: 0.044), Socialist Left (M: 0.007, Sd: 0.036), Liberal (M: 0.004, Sd: 0.025), and Green (M: 0.002, Sd: 0.016). A Kruskal-Wallis H test indicated significant differences among the means (p < .001). Post-hoc testing using Dunn’s test showed as only significant differences a higher mean for the Red Party than for the Liberal (p < .005) and Green (p < .005) Parties.

To sum up, we can classify the Progress, Red, Labour, and Socialist Left Parties as those using the most populist communication, with the described differences across platforms: the Progress Party is most populist on Facebook (albeit not significantly so, as discussed above); the Labour, Red, and Socialist Left Parties are on top for Instagram (statistically indistinguishable from each other); and the Red and Socialist Left Parties rank highest on Twitter. Compared to the right-wing Progress Party, the left-wing Labour, Socialist Left, and Red Parties employed populist communication more consistently across all three platforms. Thus, in line with the first hypothesis, we see a clear differentiation of the parties that we have categorised as being more towards the fringes of the Norwegian political-party spectrum (Progress, Socialist Left, Red) across platforms – with the Labour Party as the odd one out.

Finally, we tested the second hypothesis, which postulates that parties will use different types of populist communication depending on their ideologies. We investigated the means and standard deviations for our sub-indices of people-centrism (see Figure 2), anti-elitism (see Figure 3), and exclusion of out-groups (see Figure 4) for all parties. To avoid too few cases, we lump the platforms together here.

FIGURE 2

FIGURE 3

FIGURE 4

Figure 2 suggests that the Labour Party (M: 0.058, Sd: 0.234) made the most frequent use of people-centrism, with their future coalition partner the Centre Party (M: 0.044, Sd: 0.207) in second place. A Kruskal–Wallis H test indicated significant differences between the means. A post hoc test using Dunn’s test revealed that the means reported for the Labour and Centre Parties were not significantly different from each other (p > .005). However, the mean for the Labour Party emerged as significantly higher when compared with all other means (p < .005 for all such comparisons). The mean reported for the Centre Party, however, was not significantly higher than any other mean (p > .005 for all such comparisons). Thus, during our study period, the Labour and Centre Parties used references to the people most often in the posts on their official social media accounts.

A different picture emerges when we assess the means of our anti-elitism index (see Figure 3). This shows that the three socialist parties emerged at the top in this regard. The far-left Red Party (M: 0.047, Sd: 0.116) used anti-elitism to a higher degree than the left-wing Socialist Left Party (M: 0.028, Sd: 0.095) and the social-democratic Labour Party (M: 0.023, Sd: 0.091). A Kruskal-Wallis H test for differences between the means reported in Figure 4 proved significant differences (p < .001). A post hoc test using Dunn’s test revealed that the mean for the Red Party was significantly different from all other means (p < .001). The mean reported for the Socialist Left Party emerged as higher than those reported for the Liberal, Christian People’s, Green, or Conservative Parties (p < .005 for all means) but failed to reach statistical significance when compared with the means found for the Labour and Progress Parties. In sum, while parties belonging to the red bloc prevailed with regard to anti-elitism, not all the means reported for these parties were found to be significantly different from each other.

While more parties employed people-centrism and anti-elitism, Figure 4 reveals a very different picture of the exclusion of out-groups. This element was used almost only by the right-wing populist Progress Party (M: 0.068, Sd: 0.18). A Kruskal-Wallis H test (p < .001) in combination with a subsequent Dunn’s test showed that the mean for the Progress Party was significantly different from all other means (p < .001 for all means). No other significant differences emerged from this analysis (p > .005 for all comparisons).