By Axel Bruns and Daniel Angus.
The 2020 Queensland election campaign is coming to a close. Held under the shadow of the COVID-19 pandemic, with its impact on the mode of campaigning, the volume of postal and early voting, and of course the themes of political debate, this was an election campaign like no other. Here is the QUT Digital Media Research Centre’s final-week update on the social media activities of Queensland’s political parties, following our posts on weeks one, two, and three.
In this final stage of the election campaign, it’s worth looking back over the parties’ and candidates’ activities overall. For the timeframe of 3-29 October, our data for Twitter and Facebook reveal some very different patterns, but also some obvious similarities – and in this post, we are also paying particular attention to the differential patterns of activity across the various regions of Queensland, with their distinct demographics and specific political issues.
On Twitter, it’s very clear that Labor is well ahead on all engagement metrics. With just over 2,500 tweets sent, Labor candidates have been substantially more active than their competitors from the LNP (286 tweets only), and well ahead of the Greens (632 tweets). This imbalance cannot be accidental – rather, it likely indicates a distinct prioritisation of other platforms by LNP candidates, possibly due to an unfounded but widespread belief that the Australian Twittersphere has a left-of-centre bias.
A very substantial proportion of the tweets sent by all major parties are @mentions or retweets (and often both); this reflects a typical pattern of mentioning the central party account, party leaders, and other candidates in campaigning tweets. Only 14% of Labor and Greens tweets, and only 4% of LNP tweets do not @mention or retweet other accounts.
Public engagement with the candidates’ tweets similarly favours Labor: to date, its candidates have received more than 92,000 @mentions and retweets, or more than double the 41,000 tweets directed at LNP candidate accounts. Of these, the vast majority are directed at the party leaders: Labor’s Annastacia Palaszczuk accounts for 58% of the tweets at Labor candidates (and her Deputy Steven Miles attracts another 34%), while fully 95% of all tweets directed at LNP candidates @mention or retweet the LNP’s Deb Frecklington.
This demonstrates the concentration of attention and debate on Twitter on these highly visible party leaders; the broader team of ministers and shadow ministers, let alone the wider range of local candidates, barely rate a mention.
Notably, too, Labor’s candidates (chiefly, Palaszczuk and Miles) attract a considerably larger percentage of retweets; we have observed this at several points during the campaign, and the overall pattern has held. Some 18% of all tweets at Labor’s candidates, and 11% of tweets at the Greens, are retweets, but only 3% of those directed at LNP candidates. This is due in part also to the low volume of tweets from LNP candidates, of course (Twitter users have nothing to retweet if candidates themselves don’t tweet), but also indicates that users are considerably less willing to share and thereby amplify the tweets posted by Frecklington and her team.
Patterns on Facebook for the full campaign period of 3-29 October are substantially different. Here, Labor and LNP candidates have been nearly equally active in posting to their official pages: with 9,200 posts, Labor candidates are only slightly ahead of LNP candidates’ 9,000 posts. Many of the minor parties have also used Facebook considerably more than Twitter: One Nation candidates published 2,900 posts, and Greens candidates nearly matched this with their 2,500 posts.
Different campaigning styles emerge here, too: it is notable that the major party candidates almost exclusively utilised photo and video posts: 94% of Labor’s and 97% of the LNP’s posts fell into these categories. This almost certainly reflects well-established marketing advice that visual content tends to receive more engagement on Facebook and other social media platforms. Many minor parties show similar patterns – but here, posts featuring links to external content or simple text-based status updates are more prominent (16% for One Nation, and 21% for the Greens).
These choices also result in different public engagement patterns for the various parties. Labor and the LNP are clearly well ahead in the total count of reactions, comments, and shares received for their candidates’ posts; despite their similar levels of activity, however, Labor candidates receive considerably more engagement than their LNP counterparts. As on Twitter, this appears to indicate a much greater willingness amongst Facebook users to endorse and share Labor content.
This reading is also supported by differences in reaction types. Ignoring the generic ‘like’ and grouping ‘sad’ and ‘angry’ into more strongly negative and ‘love’, ‘care’, ‘haha’, and ‘wow’ into more strongly positive reactions, further differences between the parties are obvious: while Labor and the LNP both receive some 3% negative reactions, Labor’s 17% positive reactions are nearly double the 9% received by LNP candidates.
Meanwhile, the 7% of negative reactions that One Nation candidates generate nearly match their 11% positive reactions; as One Nation posts also result in a disproportionately large number of shares, however, this points to a calculated negative campaigning strategy: candidates share content designed to elicit a negative reaction that is strong enough to encourage Facebook users to share such content with their own networks. By contrast, of the larger parties Greens followers are the most positive: only 2% of their reactions are negative, and a full 24% positive – but this also results in very limited commenting or on-sharing.
Compared to Twitter, the party leaders account for a somewhat smaller percentage of Facebook engagement with candidate pages: of the 903,000 reactions received by Labor candidates, 56% are for Palaszczuk’s and Miles’s pages; of the 436,000 reactions to their LNP counterparts, 44% are for Frecklington’s page. There is a concentration on the leaders here, too, but clearly it is less pronounced than on Twitter – this reflects the more localised network structure of Facebook as a platform, which favours local engagement.
This local engagement becomes more obvious if we examine the engagement patterns in electorates across the various regions of Queensland. We focus here only on the four most significant parties, and have excluded the electorates of Inala (Palaszczuk), Nanango (Frecklington), and Murrumba (Miles) from our analysis – as these candidates receive extraordinary levels of state-wide and national engagement because of their special roles in government and opposition, they would skew the overall analysis.
Several significant patterns emerge from this analysis, and these provide clues to the parties’ relative standing in the regions. In Far North Queensland (FNQ) and North Queensland (NQ), the two major parties are relatively evenly matched, with a slight lead for Labor across all forms of engagement; between the smaller parties, the Greens generate more engagement in the Far North than One Nation, but are largely dormant in North Queensland electorates. And while it is not pictured here, the party with the greatest engagement in NQ electorates – ahead even of the major parties – is Katter’s Australian Party, reflecting its status as a regional force in the North.
Next, Central Queensland (CQ) stands out for its strong engagement with One Nation candidates, and the exceptional volume of comments and shares for both the LNP and One Nation. This is almost certainly linked to the LNP’s controversial proposal of a nighttime curfew for young people in Townsville, which clearly generated a substantial amount of debate. It remains to be seen whether this translates into electoral success, however: the strong engagement for One Nation here indicates that Central Queensland remains a difficult region for both major parties.
One Nation also receives comparatively strong engagement on the Sunshine Coast (SC), which is also one of the few regions where LNP candidates have generated more engagement overall than their Labor competitors (even after excluding Deb Frecklington’s electorate of Nanango). These patterns are similar, but even more pronounced, on the Gold Coast (GC): here, engagement with LNP candidates is well ahead of their Labor challengers. And at much lower levels of overall activity, the same is also true for South Western Queensland (SWQ) electorates.
If these three regions can be considered to be LNP heartland, then the same has traditionally been the case for Labor and the greater Brisbane region. In both the northern and southern parts of Greater Brisbane (GB-N/GB-S, excluding Palaszczuk’s and Miles’s electorates), as well as in immediately adjacent South East Queensland (SEQ) electorates, engagement with Labor candidates easily outpaces that with their LNP competitors.
There are notable regional differences here, however: in Brisbane, and especially in its southern electorates, the Greens are especially well represented; they already hold the seat of Maiwar (GB-N), and there is a chance that they might also win the seat of South Brisbane (GB-S), currently held by Labor’s former Deputy Premier Jackie Trad. In broader SEQ electorates, meanwhile, One Nation also performs strongly, and even generates engagement levels similar to those for the LNP.
These Facebook engagement patterns show the parties’ difficulties in developing a single set of campaign messages that are attractive to voters across all Queensland regions. Both in their organic messages and in their targeted advertising, different candidates may highlight very different aspects of the parties’ policies and plans – and such a diversification of messages can also create dissonance and confusion, of course.
Such dissonance was exploited by the Greens with a Facebook ad in metropolitan Brisbane that was largely a cut-and-paste of a Facebook ad that Labor had run in the seat of Mackay. The ad features the incumbent Labor member Julieanne Gilbert in high-vis clothing and a hardhat at a mine site, with the text “18 new coal mines approved in Central QLD. The Palaszczuk Government is powering our economic recovery”. Metropolitan Brisbane Labor candidates have largely remained quiet on the issue of coal mining during the campaign.
More generally, while a large proportion of the material posted to Facebook by candidates contained visual content (images/videos), many of these posts also included brief sections of text. We profiled this text and extracted bigrams to understand the key messages that candidates were promoting throughout the campaign. A bigram is a pair of words that appear together, and here we focussed on a subset of the most frequent bigrams, sorted by party.
There is a small subset of highly frequent bigrams that include ‘One Nation’, ‘Deb Frecklington’, Economic Recovery’, and ‘State Election’.
The most frequent of these, One Nation, was mentioned in 936 posts, mostly mentions by One Nation candidates (883 posts). More interesting, however, is the second ranked bigram, ‘Deb Frecklington’, who was mentioned more frequently by Labor candidates than by those of her own party. The same is true for the third-ranked bigram, ‘economic recovery’, which was also used predominantly by Labor. This suggests a two-pronged strategy by Labor of directly challenging rival candidates while also selling their own policy vision, framed around the post-COVID economic recovery.
The remaining bigrams are a mix of slogans, policy foci, and self- and other-party references. The LNP’s major slogans include the trigram ‘Get Queensland Working’, ‘Change Government’, and ‘Create Jobs’. Labor candidates have tended to highlight post-COVID pandemic plans, with references to ‘Recovery Plan’, ‘Queensland Border’, ‘Across State’, ‘Reelected Palaszczuk’, and ‘Palaszczuk Government’. The minor parties tended to try to differentiate themselves from the major parties’ politics, with the Greens focussing attention on ‘Mining Corporations’, ‘Public Health’, and a ‘Plan (to) Create’. One Nation are clearly seeking a ‘Balance (of) Power’ position in the new government, and UAP candidates have focussed most strongly on their ‘Death Tax’ scare campaign.
On the topic of the UAP’s ‘Death Tax’ disinformation campaign, we can see a clear push early in the campaign (on 7 October) by UAP candidates, and a subsequent push on the 16 October. Labor may have resisted responding to these unfounded claims early, perhaps as a strategy to deprive the disinformation campaign of oxygen, but in the last few days of the campaign have decided to address it more directly. We can also see this disinformation being picked up and uncritically repeated again by One Nation candidates at this late stage.