Abstract
New Zealand’s 2020 General Election campaign was unusual, though not unprecedented, in featuring women as both Prime Minister (Jacinda Ardern, Labour) and Leader of the Opposition (Judith Collins, National). To explore the extent to which gender, party and style intersected in their social media positioning, we analysed all posts made on the two Party Leaders’ public Facebook pages. We found both quantitative (post frequency and composition, main topic and policy issues, audience reactions) and qualitative differences (tone, presentational style) but importantly, our research suggests that neither woman ‘performed’ gender in normatively stereotypical ways.
Keywords
Introduction and context
Social media platforms are now a key part of political campaigning and while politicians use a variety of options, Facebook maintains its lead (Bossetta, 2018; Magin et al., 2017; Yarchi and Samuel-Azran, 2018), not only because of its greater reach but also because of its more diverse audience (Auxier and Anderson, 2021). Although user numbers have probably peaked, the platform’s penetration is substantial with 83% of New Zealanders using it (Statista, 2021).
Historical, cultural and socio-political contexts are important in understanding campaign communication (Fountaine et al., 2019; Lalancette and Cormack, 2020), particularly online. With this in mind, our study of the New Zealand (NZ) 2020 General Election campaign contains much of interest. The campaign was dominated by the contrasting styles and strategic positions of the women leaders of the two main parties. Examining Facebook posts enabled us to investigate how aspects of gender, party difference, incumbency and generational allegiances were embodied in the messaging of the Leader of the Labour and the National Party respectively, and how this played out in the context of the Covid-19 pandemic crisis.
NZ’s 120-seat Parliament uses a mixed member proportional (MMP) electoral system where voters cast two votes: one for their electoral (geographic district) candidate and the other for their preferred party. The 2020 election date was delayed for four weeks to 17 October (with early voting from 3 October) due to Covid-19. The centre-left Labour Party entered the official campaign on September 13 with a strong incumbency advantage, largely reflecting its 40-year old leader Jacinda Ardern’s decisive and empathic leadership in the aftermath of the Christchurch Mosque shootings and during the early months of the Covid-19 pandemic, during which she fronted daily televised press briefings (Duncan, 2020).
In contrast, the National (centre-right) Party Leader, Judith Collins, faced several obstacles. When the election was called, it was only two months since she had assumed the party leadership after a period of unprecedented turmoil. Struggling to dent the popularity of Ardern, National had ousted its then leader for relative unknown Todd Muller, who subsequently had the shortest reign of any political leader in NZ when he resigned after just 53 turbulent days. Collins, who had twice before competed for the party leadership and was an often-divisive figure in her party, was subsequently elected. As a seasoned parliamentarian aged 61, she had been a cabinet minister in former National governments and relished her profile as a tough opponent, embracing her ‘Crusher Collins’ moniker which contrasted starkly with Ardern’s renowned reputation for kindness and compassion (Duncan, 2020).
The authentic politician: engagement, self-presentation and personalisation
A complex range of electoral and contextual factors intersect with the characteristics of social media to produce election outcomes. Politicians themselves insist that the particular affordances of social media platforms such as Facebook allow them to communicate and interact directly with citizens (Keller and Kleinen-von Königslöw, 2018; Magin et al., 2017) and mobilise voters (Ross et al., 2020), and there is some evidence that when political parties do interact with their followers, there is increased engagement, in the form of comments, likes and shares. Farkas and Bene (2021: 126) suggest that, ‘On Facebook, triggering user engagement is a strategic goal in itself’, because the structure of Facebook’s platform means that likes, shares and comments can boost the visibility of posts. However, in reality, politicians’ actual levels of interaction are very modest (Enli and Skogerbø, 2013; Magin et al., 2017; Sørensen, 2016). While more Facebook posts do not necessarily lead directly to more user reactions, they can attract more followers and thus potentially increase engagement. Keller and Kleinen-von Königslöw (2018) suggest that the number of followers that a politician engages will drive the number of reactions, and this is influenced by factors including the volume of legacy media coverage enjoyed by the politician, their visibility and activity in televised parliamentary debates and the urban nature of the electorate. Other predictors of audience reaction include posts which have an emotional pull, personal content and humour, all of which promote likes and shares (Larsson, 2015) whereas the more routine mobilisation post enjoys significantly less traction (Heiss et al., 2019). Thus, non-political content is generally more popular on Facebook than policy-oriented content (Farkas and Bene, 2021).
Some politicians acknowledge they use social media to subvert mainstream media agendas and there is also empirical evidence that political leaders use social media platforms to focus on the messages they wish to promote rather than mirroring the policy preoccupations of journalists (Ross et al., 2020). A crucial aspect of their social media behaviour relates to how politicians present themselves on social media and the increasing use of visuals, both photos and video, across all social media platforms has become an important element of politicians’ impression management strategies and accords with the personalisation turn of modern politics (Enli and Skogerbø, 2013; Kreiss and McGregor, 2018; McGregor et al., 2017). Obama’s 2008 presidential campaign is widely viewed as being the game-changer in politicians’ use of social media as an election strategy to grow their support base (Borah, 2016). Politicians and their campaign teams began to realise that to demonstrate their relevance and relatability, candidates needed to be messaging via the media platforms which were being used by a younger and potentially persuadable voting demographic (Enli and Skogerbø, 2013; Sweetser and Lariscy, 2008). Visual content in general is seen to increase user engagement, with video being particularly successful in boosting shares and thus extending a politician’s reach (Koc-Michalska et al., 2021). Most studies which look at visual images on social media, especially Instagram, suggest that politicians’ self-presentation strategies are more inclined to show themselves in professional rather than informal modes (Metz et al., 2020), often accompanied by campaign banners or other politicians or supporters, all of which contribute to what Grabe and Bucy (2009) have described as the ‘ideal candidate’ frame, as distinct from the ‘popular candidate’ frame (Uluçay and Melek, 2021).
The turn towards more professionalised modes of political communication, self-presentation strategies and impression management has led to a growing interest in the concept of authenticity and its relationship with truth. Enli (2016) used the term ‘authenticity illusions’ to demonstrate the constructed nature of politicians’ authenticity as a mediated performance, theorising seven types of authenticity: predictability, spontaneity, immediacy, confessions, ordinariness, ambivalence and imperfection. Another and smaller set of occasionally congruent dimensions emerges from Luebke’s (2021) review of authenticity theories comprising consistency, intimacy, ordinariness and immediacy. What seems clear, then, is the multi-layered, complex and contradictory nature of authenticity, where politicians and particularly leader candidates, strive to be both like us but also above us (Wood et al., 2016). We consider the issue of authenticity in our analysis.
The gendered politician
Although studies of politicians’ social media use still do not consistently report gender variables, the past decade has seen a growing interest in understanding the influence of gender, partly because of the increased number of women politicians and leaders but also because social media were, at least initially, thought to provide important opportunities for women politicians to gain the visibility often denied them by legacy media (Cardo, 2021; Fountaine et al., 2019). These gender-inflected studies have produced contradictory results, especially in relation to the behavioural traits traditionally associated with women and men. For example, some suggest that women are heavier users of social media (Jacobs and Spierings, 2016), are more interactive (Meeks, 2016) and attract higher levels of user engagement because of their more personalised content (Yarchi and Samuel-Azran, 2018). Others argue that men are more likely to engage in self-personalisation and send messages about their families (McGregor et al., 2017), not least because families are seen as an asset for political fathers but a liability for political mothers (Ross, 2017). This double bind, competence versus likeability, is compounded by gender-based stereotypes and assumptions about party issue ownership, for example that women/the left are better at ‘soft’ social issues like health and welfare and men/the right are better at ‘hard’ issues such as the economy and crime (Meeks and Domke, 2016).
Acknowledging these perceptions means that women (and indeed men) must carefully negotiate their messaging to attract maximum advantage and mitigate potential disadvantage from their imagined audience (Kelm, 2020). Women need to show sufficient masculinity traits to demonstrate their competence, tempered by sufficient femininity to demonstrate their likeability. McGregor et al. (2017: 276) suggest that women candidates combine personal and professional campaign-oriented posts on Facebook in order to show themselves as both, ‘engaged and competent leaders and as caregivers’. Similarly, Cardo’s (2021) study of Hilary Clinton, Theresa May and Jacinda Ardern’s use of social media suggests that they have all used a mix of masculine and feminine styles. While all three leaders focused on hard policy topics in their posts, they swerved potentially gender-incongruent (and thus possibly harmful) content by suggesting that economics was about people and discussed education and health in terms of their economic impact. On the other hand, Fountaine’s (2017) study of Ardern, in one of her earlier (non-leadership) electoral races when she was active on Twitter, identified a ‘likeability’ frame which was ‘rooted in the visual and rhetorical mix of personal and political’. Both Ardern and her then female opponent (Nikki Kaye) used humour, expressed sympathy and appreciation for others and tweeted about campaign activities alongside pictures of family pets. This style of political communication, also apparent with politicians such as Justin Trudeau, suggests that a younger generation of political leaders are embracing a political language which is more relatable to the historically disengaged younger citizen (Lalancette and Cormack, 2020).
Method
We performed a content analysis of all Facebook posts made on the public pages of Jacinda Ardern and Judith Collins, from 13 September 2020 (Writ Day) to 16 October (the last day of campaigning before Election Day). Replicating aspects of our approach to analysis of the 2017 election campaign, our quantitative analysis coded for topic, policy, reference to family and young people, post composition, links and hashtags. We coded posts for the main topic and any policy posts were then coded for individual policy issues. The number of shares, comments and categories of emoticons was also recorded. The substantial amount of video content posted on Facebook during this campaign led to an additional set of content analysis questions, focused on video length, composition, style and tone. Our data collection comprised taking a screenshot of each post and the accompanying reaction data, every evening, at least 24 hours after the post was made. Subsequently, our reaction data represent at least a full day of follower responses.
The authors each coded approximately a third of the posts from each Party Leader, and a sub-set of posts was coded by all three authors and discussed as a pre-test of the code book which produced an intercoder reliability of 95%. Our qualitative content analysis extended these quantitative results by providing illustrative examples identified by the three authors during their independent coding and in their personal observations of key themes, all of which were subsequently discussed in multiple Zoom meetings over a 6-month period of analysis, discussion and writing.
Findings
Volume and frequency
A total of 279 Facebook posts were manually coded across the election campaign period. The first and most striking finding is that Judith Collins made nearly twice as many posts (182) as Jacinda Ardern (97). While this is perhaps not surprising as it mirrors findings which suggest opposition parties are more prolific users of social media (Larsson and Kalsnes, 2014), it is particularly interesting that the respective number of Party Leader posts is almost the same as in our study of the 2017 NZ election, albeit that the National Party had a different leader at that time (Ross et al., 2020). In that earlier case, Labour were in opposition and could therefore have been expected to post more frequently. This suggests that although incumbency may influence certain aspects of campaign activity and is likely to impact the number of social media followers a leader has, other factors are in play including the importance placed on social media as part of the broader campaign communication strategy. When we tracked frequency over the entire campaign, there were few differences other than volume, although both leaders posted more during the final week, with Collins making a significant final push with nine posts on the day before the vote, the majority of which were negative in tone, compared with Ardern’s more modest two posts.
Text versus image
Given previous research suggests that social media posts which include images gain more audience traction than text only (Bossetta, 2018; Koc-Michalska et al., 2021), we considered the broad composition of posts, how the two Party Leaders used photographs and moving images, and who featured in them. Figure 1 shows the broad composition of posts.
Post composition x party leader.
A relatively small number of posts comprised Party Election Broadcasts (PEBs) or static manifesto pledges (8% Ardern, 10% Collins) and only 1% of both Party Leaders’ posts were text-only. All the content which included photos and most of the posts which included video also had text captions, although a fifth of Ardern’s posts were video-only, compared with just 3% of Collins’ posts. Collins was much more likely to include photos (68%) than Ardern (47%), with Ardern more frequently including video (43%) compared with Collins (19%). There were also differences in the extent to which the two Party Leaders featured in the photos they posted, with Collins appearing in 55% of her own posts’ photos, compared with 28% for Ardern, although Ardern appeared as the primary focus in nearly all of her posted video content (see also Table 1). The almost complete absence of text-only posts suggests Party Leaders are paying attention to citizens’ preference for visual content, although each adopted a different strategy in response. Collins favoured photos of her visiting different constituencies with nearly half her posts featuring photographs of her with other people. Ardern, on the other hand, posted as much video as photo content, mostly of her campaigning at different venues, but also a number of Facebook Live events where she directly addressed the viewers and answered questions. While Ardern has a reputation for informality and regularly posts self-recorded video, including from her home (Kapitan, 2020), such posts comprised less than a fifth of her campaign content, though her informal style was evident across the range of video content she posted (see Table 1). Her selfie-style video posts were particularly appreciated by her followers in terms of reactions, as we show below.
Video composition, style, tone and length x party leader.
Collins and Ardern rarely posted about their families and only three posts included photos of family members, two from Collins and one from Ardern, all of which showed them with their respective partners. Women politicians’ reticence to play the family card is something we also found in our earlier study (Ross et al., 2020) and is a privacy strategy which women are more likely to employ than their male counterparts (Meeks, 2016, 2019; Stalsburg and Kleinberg, 2016). Where other family members were mentioned (albeit infrequently) in posts, it was mostly indirectly for Collins (e.g. she mentioned ‘my ancestors’ and being a ‘farmer’s daughter’) but more directly for Ardern, who mentioned her niece and her daughter in two separate posts. Ardern also mentioned her daughter in some of her relatively lengthy, selfie-style Facebook Live videos, including in response to the follower questions she elected to answer. Table 1 summarises the video data.
Topics and policy posts
When we looked at the content of posts, the two most popular types of content for both Party Leaders were local campaigning (in either their own or colleagues’ constituencies) and posts with policy as the primary topic (excluding PEBs), with more of Collins’ posts (36%) including policy-oriented content compared with those posted by Ardern (25%). The third most frequent type of content for Ardern related to organised party events (15%) compared with Collins (3%), while Collins’ third most frequent topic of content was Leaders’ debates (10%), compared with Ardern (6%). Collins clearly regarded the live televised Leaders’ debates as key opportunities to ‘perform’ for voters, and as Table 1 shows, she also included various clips from these debates in her posts. Nearly a third of Collins’ video content came from these media events, including one clip edited in a way that was subsequently identified as a half-truth by #nzvotes (2020). Interestingly, and perhaps surprisingly, Ardern was notably more reticent in these media encounters and appeared to focus more of her performative energy on her live Facebook sessions with followers.
Outside of these four content categories, there were several other topics on which both Party Leaders posted but the orientation of their posts was often different. For example, both Party Leaders made eight posts primarily about voting, but whereas Ardern mostly encouraged her followers to simply get out and vote, most of Collins’ posts comprised an explicit exhortation to ‘vote National’. Similarly, while both posted about the Leaders’ debates, most of Collins’ content was either focused on her backstage preparation (including shots from the make-up room), self-congratulatory about her own performance and/or critical of Ardern’s, while Ardern’s posts tended to be fact-based in terms of the timing of the debates or mentioning that she was about to participate. Both Party Leaders made posts promoting gender equality generally (Ardern, five; Collins, two) but with mostly different foci. Ardern posted about visiting Kate Sheppard’s house and the work the suffragette did to promote NZ women’s rights in the 1880s, as well as posts encouraging young women to believe they can do anything and a post about the death of Ruth Bader Ginsburg which coincided with NZ’s Women’s Suffrage Day. Collins also posted about Kate Sheppard in the context of Women’s Suffrage Day (which fell during the campaign period), with another post encouraging more women to take up the law as a career, as she had done.
While several studies suggest that politicians do not use social media platforms to talk about policy issues because this is not what citizens want to read (Heiss et al., 2019), both Party Leaders posted relatively frequently about policy topics (see Figure 2). Overall, 33% of all posts had some policy-focused content: 25% for Ardern and 36% for Collins. These findings are similar to 2017, where Ardern mentioned policy topics in 26% of her posts compared with 40% of posts from then National Party Leader Bill English (Ross et al., 2020). While both Party Leaders posted about the economy, health, farming, Covid-19, housing and regeneration, albeit at different frequencies, there was also a range of topics which only one or the other posted about, and these are quite revealing.
Policy post topics x party leader.
We suggest that Ardern’s focus on the environment was highly tactical since her support base is perceived to be younger than Collins’ and she held rallies at two university campuses during the campaign. Given the increasing visibility of young people demonstrating their concern and support for planetary care and indeed the equally pressing problem of young people’s deteriorating mental health, the focus on these themes and the general orientation of Ardern’s policy posts is entirely logical in its appeal to her core constituency. Relatedly, around 14% of the photos in her posts featured young people and half of those were of Young Labour supporters involved in various campaign activities (10% of Collins’ posts included photos of young people with around a third of those featuring Young National supporters). Given that the election campaign was taking place when the world was in the devastating throes of the first wave of Covid-19, the volume of Ardern’s health and Covid-related posts is entirely to be expected.
Collins posted more frequently about the economy and taxation and in particular, Labour’s intention to impose a ‘wealth tax’ which she warned would significantly affect pensioners and anyone with savings. In one of her posts on International Older Persons’ Day (1 October), she included a manifesto pledge to further support older people and provide health service discounts, and another (on 8 October) included a clever poster mock-up which compared what National would do (give seniors $1000 tax relief and other benefits) compared with the ‘intentions’ of a Labour-Green coalition (impose a NZ$7000 tax on savings). Collins’ economy posts were focused on what National would do to support the business sector and many featured photos of her visiting various large and smaller companies, leveraging National’s reputation for being the party of business. This dual focus (these two topics combined comprised 51% of all Collins’s policy posts) made strategic sense for Collins because of both the wealthier/entrepreneurial and older age profile of her support base.
Unsurprisingly, even when the Party Leaders posted about the same topic, the tonal orientation was entirely opposite. A significant minority of all Collins’ posts (14%) were primarily coded as ‘attacks’, mostly generalised criticisms of the Government, and nearly half of all video content (43%) was coded similarly (see Table 1). However, 41% of Collins’ policy content was also coded as attacking in tone, especially in the second half of the campaign, nearly half of which was about tax. None of Ardern’s posts were attack-oriented and our earlier study of five Party Leaders noted that negative Facebook posts were almost entirely absent from NZ leaders’ social media campaigning in 2017. Again, the use of the attack frame by opposition politicians is in keeping with the extant literature (Borah, 2016) but the frequency and focus of Collins’ attack posts is interesting because it was not a feature of her party’s earlier Facebook campaigning and while it produced significant audience reaction, as we explore further below, it ultimately failed to push her party ahead of Ardern on polling day.
Style and substance
Both Party Leaders have been involved in politics for many years and are well-known for their particular style, with the younger Ardern being relatively informal and collegial and Collins having more of a ‘hard woman’ reputation linked to her previous roles in Government (Duncan, 2020). Ardern regularly used humour in her posts, often self-deprecating. For example, a playful post on 20 September featured her work bag with the caption: ‘A beautiful day for outdoor paperwork, even if you haven’t been home for a while and you’re trying to ignore that your garden/weeds are now taller than your briefcase’. Self-deprecation was also a feature of her responses in informal Facebook Live posts, where she downplayed the success of her efforts to answer all questions, referring to ‘my usual poor job of responding to comments’ and concluding her 16 September post (which prompted the most comments of all her election posts) with, ‘I’ve done a terrible job, forgive me’. While Collins also used humour, her posts often constituted a complicated mix of hubris and light-heartedness, especially in relation to her reputation as ‘Crusher Collins’. For example, one post on 15 September included a screenshot from the New Zealand Herald showing a tattooed bicep featuring Collins brandishing a gun in a typical ‘Bond-girl’ pose with her infamous moniker scrawled below the image, and captioned by Collins as, ‘I do get some lovely fan mail but this is a first’. This post self-consciously and unapologetically acknowledges her aggressive reputation while also conveying a sense of the admiration she enjoys from (at least some) citizens.
Another highly instructive difference in the two women’s style relates to claims of expertise and competence, with Collins rarely missing an opportunity to talk about her years in Government and her profession. On NZ’s Police Remembrance Day (29 September), Collins posted, ‘as a lawyer, a former Minister of Corrections, Police and Justice, stamping out corruption in our country is very important to me. . .’ None of Ardern’s posts focused on her own achievements before politics or as PM, such as the global praise she received for her handling of the Covid-19 pandemic, although her Facebook Live posts did allude to Government achievements, such as getting debt levels down, inflected with an inclusive ‘we’ language (e.g. ‘we like to make sure we’re prepared for a rainy day’).
Audience (at)traction
A significant aspect of social media use in political campaigning is its affordances in prompting audience reaction through comments and likes as well as extending the reach of messages through followers’ networks (Keller and Kleinen-von Königslöw, 2018; Yarchi and Samuel-Azran, 2018). Ardern’s posts attracted substantially more comments than Collins’, with 90% receiving at least 500 comments including 51% attracting more than 1000 comments, whereas the majority of Collins’ received at least 50 comments and 14% received more than 500 comments. In terms of shares, most of Ardern’s posts were shared at least 100 times, and more than 10% were shared more than 500 times. For Collins, most of her posts were shared between 20 and 50 times, with a smaller number being shared more than 100 times; none of her posts were shared more than 500 times and one of her posts had no shares. However, we need to put these data into context since Ardern had between 1.3 and 1.7 m followers during the campaign, compared with Collins’ more modest following of around 58 k. If we look at how engagement (Kapitan, 2020) is measured in relation to social media, then Collins actually prompted more engagement per follower than Ardern in terms of comments and shares.
Post provocateur
Some posts made by the Party Leaders attracted significantly more attention than most in terms of comments and to better understand what kind of content prompted the most commentary, we undertook a close analysis of the content of the five best performing posts from each Party Leader. Two of Judith Collins’ top five most commented posts criticised a possible Labour/Green coalition but interestingly, her two most commented upon posts (1400 and 1300 comments respectively) were about the televised Leaders’ debates. Late on 22 September, following the first TVNZ debate, Collins shared a visual from the New Zealand Herald website showing that its readers had voted her the debate winner, which she captioned as, ‘What a fantastic night! I had so much fun in the debate that I could have kept going for another hour. . .PS: I’m rather chuffed with this on the NZ Herald website!’ Just over a week later, Collins received 1300 comments after she posted herself on the (mocked-up) front cover of Vogue, following a humorous interaction during that evening’s election debate where Collins was asked if her views on Ardern as a ‘celebrity’ politician meant that she herself would turn down the opportunity to feature in Vogue: ‘A staff member made this for me as a joke tonight after the TV3 debate. Actually, it’s rather a good idea’. Her final most commented upon post also related to the televised Leaders’ debates and featured a behind-the-scenes photo with the words ‘here we go!’
Although the televised Leaders’ debates were the focus of just one of Ardern’s most commented upon posts, the relationship between live material and follower commentary was also apparent in her top five, which for Ardern meant her use of Facebook Live. Her two most commented upon posts during the campaign featured live video. Further, in Ardern’s top posts, there was a clear connection between high levels of comment and the explicit centring of herself in the content. Four of Ardern’s most commented on posts were her informal live videos where she talked directly and intimately to the camera (often answering viewer questions) and a post which announced the change to her profile picture. Of the three posts that tied for fifth place, one was a Facebook Live fronted by her, one was a selfie of her wearing full television make-up just before the first election debate, and just one (the exception to our observation) featured a visual of a Labour campaign poster outlining the party’s ‘plan for working Kiwis’. Thus, for both women, there was a pattern whereby posts which directly or indirectly foregrounded aspects of their physical appearance or related to some form of visual ‘performance’ of this appearance, correlated with especially high levels of follower commentary.
Emotional tugs
We were also interested to see if posts performed differently in prompting emotional responses. While our study did not analyse the comments made on any of the posts, we captured the general tone of post appreciation through a quantitative analysis of emoji responses. The average number of combined emojis for an Ardern post was 14,688 and the average number of ‘likes’ was 10,308, with the second most popular being ‘love’ (3460), followed by ‘care’ (260), ‘sad’ (250) and ‘funny’ (194). For Collins, her average for combined emojis was 1419, with ‘like’ being the most frequent (1246), followed by ‘love’ (125) and then ‘care’ and ‘angry’ both averaging at six. All posts attracted at least 100 emoji responses with the majority attracting more than 500. None of Ardern’s posts attracted fewer than 1000 responses and the majority (57%) prompted more than 10,000 responses, including 9% which attracted more than 50,000 responses. However, once again, these scale differences need to be understood in the context of the very different follower numbers for each Party Leader. So, the average follower engagement for Collins using only emotional responses as the measure was 0.024% compared with 0.008% for Ardern, meaning Collins was considerably more successful in prompting follower engagement in relation to the average rate of response per follower per post.
We also looked at each Party Leader’s top five posts in terms of attracting ‘likes’ and other emojis. For Ardern, the post which attracted the highest volume of emojis (110 k – 6.5% engagement) was posted on 12 October and was simply a change to her profile picture: this post was also one of her top five in terms of comments (4.3 k). Her second most enjoyed post (72 k, on 29 September) featured a lunchbox with her daughter’s name on it, prepared by her partner, which Ardern described as an ‘appreciation post’. The number of ‘care’ emojis for this post was also in the top five of all Ardern’s posts. The third most popular post (58 k) featured her niece sitting at her desk in Parliament. Next in popularity (54 k) was a post about the death of Ruth Bader Ginsburg, which also prompted the highest number of ‘care’ and ‘sad’ emojis and the highest number of shares (6 k). Lastly, the fifth most popular (39 k – 2.3% engagement) featured a photo of a bride stopping off to vote before her wedding, with Ardern congratulating the bride and commenting that whatever else is happening, ‘. . .there’s always time to vote’. Posted on 11 October, it was also one of Ardern’s top five most shared posts.
For Collins, her most appreciated post (8.4 k – 14% engagement) was the aforementioned mock-up of her looking very glamorous on the front cover of Vogue. This post on 30 September was also one of her top five performing posts in terms of comments (1300), ‘funny’ and ‘care’ emojis. The post prompting the second highest volume of emojis and the highest number of comments (1400) was the one featuring a news site screenshot showing that Collins had ‘won’ the evening’s Leaders’ debate. In third place (4.4 k) was a post from 13 September featuring a profile picture change with two posts vying for fourth most popular (3.8 k – 6.5% engagement): a black and white photo of Collins with her husband on the day of National’s campaign launch (20 September), with Collins saying how excited she was and glad to have husband by her side and the other (16 October) being the only text-only post she made, on the eve of the election, where she thanked supporters and mentioned how privileged she was to have led National’s campaign. Interestingly, the post which had the most shares and comments did not feature in the top five ‘emotional’ posts but it did produce the highest number of ‘sad’ and second highest number of ‘angry’ emojis. The post contained a cartoon sequence featuring pensioner ‘Edith’, suggesting that after her husband died, she would have to make hard choices because under Labour she would be forced to pay a wealth tax.
What this brief analysis suggests is that Facebook followers are much more likely to make an emotional response to posts focused around the personal rather than the explicitly political (Farkas and Bene, 2021), perhaps because showing appreciation for a nice photo or a non-controversial post is easy and quick and does not connote support for a particular party position, thus avoiding potential censure from an individual’s network of friends and followers (Marder et al., 2016). This is not to say that explicitly political content such as the posts which focused on policy did not also prompt audience reaction since for both Party Leaders, policy posts were among those which attracted the most comments (tax for Collins, the economy for Ardern). However, for most followers, their preference was for the personal over the political.
Discussion and conclusion
We set out to explore what, if any, differences played out between Jacinda Ardern and Judith Collins in terms of the style and content of their Facebook posts during the 2020 election in NZ and found several distinguishing features. Importantly, our research complicates the ways in which gender is seen to act as an influencing factor in political communication by demonstrating the mediating effect of both party and style, especially in leadership campaigns. We suggest that Facebook was used strategically by Ardern to display a public presence, provide a focus for supporter appreciation and signal upcoming events, with relatively modest policy content. Collins used the platform for similar reasons, but policy issues were both more frequent and more targetted, and negativity was a distinctive feature, particularly in the last week of campaigning. Our findings are consistent with work reported by #nzvotes (2020) which also found that Collins made two ‘half-truth’ Facebook posts during the campaign (Ardern made none), one of which presented selectively edited comments made by Ardern in the first Leaders’ debate.
There were also obvious differences in terms of posting frequency, post composition, major policy topic and to a lesser extent, the relative scale of audience engagement, but also similarities in their tactical targetting of content, especially policy issues, towards their assumed audiences. While both politicians made frequent use of visuals, Ardern was much more likely to include video and use Facebook Live to engage followers through highly informal and personalised interactions. Successive studies of social media messaging by politicians reflect this personalised trend (Hermans and Vergeer, 2013) which happily seems to be appreciated by citizens who consider that it demonstrates a more authentic personality (McGregor et al., 2017). Parmelee et al. (2022) suggest that Gen Z like to see politicians engaging in two-way dialogue on social media and revealing the internal mechanics of the job, not just the external performance. However, the direction of cause-effect is hard to determine.
On the other hand, Collins was more likely to include photos of herself with both ‘ordinary’ citizens, especially older people, and local businesspeople, and to focus her (negative) policy content on the dangers of a ‘wealth tax’. While received wisdom has it that voters do not like attack campaigning (Borah, 2016), our analysis of followers’ online reactions to some of Collins’ negative posts provides tentative support for its increased tolerance amongst citizens, as noted by Enli (2017) and Nai (2022). However, while negative campaigning was appreciated by a proportion of Collins’ Facebook supporters, the outcome of the election suggests that this did not play well with voters at the ballot box since the National Party saw its second worst election performance of all time, with Labour taking a landslide victory.
A number of studies have looked at the ways in which politicians present themselves ‘authentically’ on social media showing themselves to be relatable but also competent and professional (e.g. Enli, 2016; Fountaine, 2017). Women politicians are particularly vulnerable to gender stereotypes and therefore tread a careful line between showing sufficient emotion to be credible as women and sufficient professionalism to be seen as credible politicians (McGregor et al., 2017; Meeks, 2016, 2019). However, in this campaign the usual rules about gender performativity and the risks and benefits of conformity or challenge were arguably less important, both because of the Covid-19 dividend and positive reputation enjoyed by Ardern and the historic ‘hard woman’ persona enjoyed by Collins but also because they were both women. They presented themselves in ways which were both similar and different to each other, including in their styling of visual content, very occasionally showing themselves as partners and mothers but mostly reinforcing their role as politicians. The use of images is an important way in which to convey this complex notion of authenticity, not only serving to anchor meaning (Kress and van Leeuwen, 1996) but also prompting a more polyvalent and complex reading of a given post. As part of their repertoire of strategies aiming to deliver an authentic sensibility in their posts, we observed that Collins and Ardern used a mix of ‘backstage’ intimacy, drawing on Goffman’s (1959) dramaturgical perspective, and/or direct-to-camera address, fusing their personal with their political identity to convey what Forchtner and Kølvraa (2017: 262) describe as a ‘horizontal social relationality’. This is not to suggest that posts which encourage this reading through the use of such intimate content or humour are entirely tactical, but to suggest that posts made by any politician are always political. Nothing that politicians post about is entirely spontaneous or entirely fabricated, but posting this rather than that photo is a strategic act of image management with the intention of prompting a positive evaluation (Uluçay and Melek, 2021). In (social) mediated communication, politicians are always on display, always firmly in the public not the private sphere and the relatability they articulate is both real and consciously performed. As Yates (2019: 348) argues, women politicians are ‘refracted through the lens of an increasingly emotionalised political culture’ so further care needs to be taken to ensure that posts create the ‘right’ kind of affect, enabling a positive identification with the audience. We argue that although both Ardern and Collins mostly performed their political selves rather than their gendered selves, when they did post more normatively feminine and feminised content which emphasised their corporeal selves, from changes to their profile pictures to Vogue covers and children’s lunchboxes, those posts gained some of the highest levels of emotional engagement, demonstrating the endurance of feminine tropes as cultural currency.
Lastly, we also considered how online audiences reacted to political messaging and found that Collins posted twice as frequently as Ardern, and her levels of audience engagement, albeit with a significantly smaller number of followers, were proportionately higher. The top-performing posts for each Party Leader showed, as with other studies (Larsson, 2015; Stetka et al., 2019) that personalised content was far more popular than ‘straight’ policy posts. This echoes work by Metz et al. (2020) who suggest that citizens want to see more emotion and personal content from political figures, although we argue that it’s not clear in which direction the cause-effect is moving. Through their political communication tactics, as much on social media as through their televised debates, political leaders engage in a rhetorical performance, the content of which they hope will resonate with partisan groups of citizens. We argue that much like the echo chamber phenomenon, politicians’ Facebook followers (other than those following for hostile reasons) are likely to have already determined their political preferences which coalesce around an assemblage of identity-based and self-interested concerns, particularly in contexts like NZ where there are two main political parties. Thus, campaign communication functions to remind citizens who their ‘people’ are and the likes, shares and comments their posts prompt provide endlessly reinforcing and self-perpetuating evidence that politician and follower are indeed on the same page.
Limitations
No research project is ever perfect and ours is no exception. The most obvious limitation is that we cannot know for sure whether the posts were written by Ardern and Collins, or by members of their campaign team but in truth, that is an issue for all research which focuses on the written word, although the video content was less ambiguous. However, our analysis is based on what the politicians and/or their campaign team members wanted to communicate to their followers and we assume that at the very least, most posted content would have had Ardern and Collins’ eyes on it, however cursorily. A second limitation is our focus on the Prime Minister and the Leader of the Opposition, thus ignoring the Leaders of the smaller parties. But again, our purpose was to focus on women leaders to keep the gender variable constant and permit a manageable sample where in-depth analysis of a sub-sample of posts was feasible. Lastly, we focus on the specific context of an election campaign and it may be that Ardern and Collins’ posting behaviour in the interregnum would produce different results and it would be useful to focus on such non-election periods and perhaps over time, in future work.
