Social Media in Australian Federal Elections:
Comparing the 2013 and 2016 Campaigns
Prof. Axel Bruns
Dr. Brenda Moon
Digital Media Research Centre, Queensland University of Technology, Brisbane, Australia
a.bruns / brenda.moon @ qut.edu.au
@snurb_dot_info / @brendam
Book Launch
Friday, 14.30, at the Routledge book stall:
Prof. Brian McNair launches
The Routledge Companion to Social Media and
Politics. Eds. Axel Bruns, Gunn Enli, Eli Skogerbø,
Anders Olof Larsson and Christian Christensen.
Basingstoke: Routledge, 2016.
Six continents. 37 chapters. 66
contributors. ~550 pages. #socmedpol
Come join us!
Australian Politics: A Crash Course
• Key features:
– Westminster(ish) system: first past the post in 150 federal electorates = 150 MPs
– But with voting preferences distribution from minor to major parties
– Plus Senate: 12 senators per state, 2 per territory – (12 x 6) + (2 x 2) = 76 senators
– Compulsory voting: 90%+ turnout, 3-5% informal (spoilt) votes
– ~3-year parliamentary terms, can be shortened to ~2.5 years on PM’s advice
• Parties:
– Australian Labor Party (centre left)
– Coalition = Liberal Party (conservative) + National Party (agrarian protectionists)
– Australian Greens (left, selected issues)
– Palmer United Party, Katter’s Australian Party, Nick Xenophon Team, One Nation, … (anti-
establishment parties formed around charismatic leaders)
Australian Politics: A Car Crash
• Considerable political instability since 2007:
– 1996-2007: PM John Howard (Liberal Party)
– 2007: PM Kevin Rudd (Labor Party), elected in landslide
– 2010: PM Julia Gillard (Labor Party) replaces Rudd in Labor coup
– 2010: PM Julia Gillard (Labor Party) wins election, forms minority government
– 2013: PM Kevin Rudd (Labor Party) replaces Gillard in Labor coup
– 2013: PM Tony Abbott (Liberal Party), elected in landslide
– 2015: PM Malcolm Turnbull (Liberal Party) replaces Abbott in Coalition coup
– 2016: PM Malcolm Turnbull (Liberal Party) wins election with one-seat majority
– (2016: Abbott building up for new coup attempt on PM Turnbull?)
 Five (six, with Rudd mk. II) Australian Prime Ministers in ten years
Social Media in Australian Elections
• Strong social media take-up in Australia:
– ~13m Facebook accounts; ~4m Twitter accounts, ~1m tweets/day (24m population)
– Political uses well-established: ~10k #auspol tweets/day; #ausvotes; #qanda; etc.
– Considerably larger volume of everyday political talk outside hashtags likely
– Most politicians have social media accounts – varying sophistication
• Australian political social media research:
– Studies of key hashtags and events (#spill / #libspill / #returnbull / …)
– Interviews with key political operatives
– Internal party research and post-election reviews
– Analyses of social media citations and sourcing practices in mainstream media
Our Approach
• Beyond hashtags:
– Identification of all candidates’ Twitter accounts before 2013 and 2016 elections
– Tracking of activities: tweets by candidates + tweets at candidates (@replies and retweets)
– Twitter Capture and Analysis Toolkit (TCAT) + Google BigQuery + Tableau
– Comparative analysis of activity around candidates and parties during both elections
• Timeframes:
– Last two calendar weeks before election day: Mon-Sun + Mon-Fri (election on Saturday)
– 26 Aug. to 6 Sep. 2013 + 20 June to 1 July 2016
• Analysis:
– Focus here on interpretation of quantitative patterns in electoral context
– Qualitative study of tweet texts and sentiments to come later
– Longer-term analysis of 2013 election patterns already published:
Axel Bruns. “Tweeting to Save the Furniture: The 2013 Australian Election Campaign on
Twitter.” Media International Australia (2016). DOI: 10.1177/1329878X16669001.
2013: Candidate Activity
2016: Candidate Activity
2013: Engagement with Candidates
2016: Engagement with Candidates
Observations
• Key patterns:
– Candidates in 2013:
• Labor campaign reliant on rank-and-file candidates; frontbench largely absent
– Aligned with strong doorknocking campaign: ‘local target’ strategy
• Coalition campaign driven by frontbenchers; other candidates quiet
– Strong central control, fear of missteps: ‘small target’ / ‘known target’ strategy
– Candidates 2016:
• Labor campaign more united; frontbenchers much more present
– Doorknocking retained, but also better central control: ‘local target’ + ‘known target’ strategies
• Coalition campaign more disorganised; limited control
– Indications of continuing internal disagreements over party direction: ‘small target’ vs. ‘friendly target’?
– Engagement in 2013 and 2016:
• Much more engagement activity by ordinary users in 2016
• Almost no retweeting in 2013; strong retweeting especially of Labor in 2016
• Governing party always receives the greatest number of @mentions
• Audience focus strongly on PM candidates and leading frontbenchers
• Also some addressing of party dissidents: ex-PMs Gillard (2013) and Abbott (2016), internal critics Wayne Swan
(Labor, 2013) and Cory Bernardi (Coalition, 2016)
Thematic Patterns
• Theme construction:
– Themes drawn from 2013 ABC VoteCompass + iSentia media themes of 2013
– Iterative construction of distinct keyword baskets for each theme
– Three additional themes introduced for 2016 election
• Theme analysis:
– Classification of tweets based on keyword baskets
– Ordered by specificity: e.g. ‘carbon tax’  Environment; ‘tax’  Budget
– Iterative testing and refinement
– Tracking across timeframes and comparison across elections
• Success rates:
– 2013: 25% of tweets matched to themes / 2016: 34% of tweets matched to themes
Themes Comparison: 2013 / 2016
2013: Themes per Day
2016: Themes per Day
(*): new in 2016
Conclusion and Outlook
• Longitudinal perspective:
– Evolution in campaigning strategies:
• Labor rediscovery of local candidates and doorknocking + social media approach
– Counteracted by internal party turmoil:
• Campaign plans disrupted by lack of enthusiasm and discipline following leadership coups
– Ordinary users’ activities largely stable:
• Focus strongly on governing party, and party leaders
– Except for retweeting choices:
• Significant changes between 2013 and 2016 – especially Labor retweeting
• Further plans:
– More detailed content and thematic analysis (quantitative and qualitative)
– Network analysis of @mention / retweet interactions: evidence of polarisation?
http://mappingonlinepublics.net/
@snurb_dot_info
@brendam
@socialmediaQUT – http://socialmedia.qut.edu.au/
@qutdmrc – https://www.qut.edu.au/research/dmrc
This research is funded by the Australian Research Council through Future Fellowship and LIEF grants FT130100703
and LE140100148.

Social Media in Australian Federal Elections: Comparing the 2013 and 2016 Campaigns

  • 1.
    Social Media inAustralian Federal Elections: Comparing the 2013 and 2016 Campaigns Prof. Axel Bruns Dr. Brenda Moon Digital Media Research Centre, Queensland University of Technology, Brisbane, Australia a.bruns / brenda.moon @ qut.edu.au @snurb_dot_info / @brendam
  • 2.
    Book Launch Friday, 14.30,at the Routledge book stall: Prof. Brian McNair launches The Routledge Companion to Social Media and Politics. Eds. Axel Bruns, Gunn Enli, Eli Skogerbø, Anders Olof Larsson and Christian Christensen. Basingstoke: Routledge, 2016. Six continents. 37 chapters. 66 contributors. ~550 pages. #socmedpol Come join us!
  • 3.
    Australian Politics: ACrash Course • Key features: – Westminster(ish) system: first past the post in 150 federal electorates = 150 MPs – But with voting preferences distribution from minor to major parties – Plus Senate: 12 senators per state, 2 per territory – (12 x 6) + (2 x 2) = 76 senators – Compulsory voting: 90%+ turnout, 3-5% informal (spoilt) votes – ~3-year parliamentary terms, can be shortened to ~2.5 years on PM’s advice • Parties: – Australian Labor Party (centre left) – Coalition = Liberal Party (conservative) + National Party (agrarian protectionists) – Australian Greens (left, selected issues) – Palmer United Party, Katter’s Australian Party, Nick Xenophon Team, One Nation, … (anti- establishment parties formed around charismatic leaders)
  • 4.
    Australian Politics: ACar Crash • Considerable political instability since 2007: – 1996-2007: PM John Howard (Liberal Party) – 2007: PM Kevin Rudd (Labor Party), elected in landslide – 2010: PM Julia Gillard (Labor Party) replaces Rudd in Labor coup – 2010: PM Julia Gillard (Labor Party) wins election, forms minority government – 2013: PM Kevin Rudd (Labor Party) replaces Gillard in Labor coup – 2013: PM Tony Abbott (Liberal Party), elected in landslide – 2015: PM Malcolm Turnbull (Liberal Party) replaces Abbott in Coalition coup – 2016: PM Malcolm Turnbull (Liberal Party) wins election with one-seat majority – (2016: Abbott building up for new coup attempt on PM Turnbull?)  Five (six, with Rudd mk. II) Australian Prime Ministers in ten years
  • 5.
    Social Media inAustralian Elections • Strong social media take-up in Australia: – ~13m Facebook accounts; ~4m Twitter accounts, ~1m tweets/day (24m population) – Political uses well-established: ~10k #auspol tweets/day; #ausvotes; #qanda; etc. – Considerably larger volume of everyday political talk outside hashtags likely – Most politicians have social media accounts – varying sophistication • Australian political social media research: – Studies of key hashtags and events (#spill / #libspill / #returnbull / …) – Interviews with key political operatives – Internal party research and post-election reviews – Analyses of social media citations and sourcing practices in mainstream media
  • 6.
    Our Approach • Beyondhashtags: – Identification of all candidates’ Twitter accounts before 2013 and 2016 elections – Tracking of activities: tweets by candidates + tweets at candidates (@replies and retweets) – Twitter Capture and Analysis Toolkit (TCAT) + Google BigQuery + Tableau – Comparative analysis of activity around candidates and parties during both elections • Timeframes: – Last two calendar weeks before election day: Mon-Sun + Mon-Fri (election on Saturday) – 26 Aug. to 6 Sep. 2013 + 20 June to 1 July 2016 • Analysis: – Focus here on interpretation of quantitative patterns in electoral context – Qualitative study of tweet texts and sentiments to come later – Longer-term analysis of 2013 election patterns already published: Axel Bruns. “Tweeting to Save the Furniture: The 2013 Australian Election Campaign on Twitter.” Media International Australia (2016). DOI: 10.1177/1329878X16669001.
  • 7.
  • 8.
  • 9.
  • 10.
  • 11.
    Observations • Key patterns: –Candidates in 2013: • Labor campaign reliant on rank-and-file candidates; frontbench largely absent – Aligned with strong doorknocking campaign: ‘local target’ strategy • Coalition campaign driven by frontbenchers; other candidates quiet – Strong central control, fear of missteps: ‘small target’ / ‘known target’ strategy – Candidates 2016: • Labor campaign more united; frontbenchers much more present – Doorknocking retained, but also better central control: ‘local target’ + ‘known target’ strategies • Coalition campaign more disorganised; limited control – Indications of continuing internal disagreements over party direction: ‘small target’ vs. ‘friendly target’? – Engagement in 2013 and 2016: • Much more engagement activity by ordinary users in 2016 • Almost no retweeting in 2013; strong retweeting especially of Labor in 2016 • Governing party always receives the greatest number of @mentions • Audience focus strongly on PM candidates and leading frontbenchers • Also some addressing of party dissidents: ex-PMs Gillard (2013) and Abbott (2016), internal critics Wayne Swan (Labor, 2013) and Cory Bernardi (Coalition, 2016)
  • 12.
    Thematic Patterns • Themeconstruction: – Themes drawn from 2013 ABC VoteCompass + iSentia media themes of 2013 – Iterative construction of distinct keyword baskets for each theme – Three additional themes introduced for 2016 election • Theme analysis: – Classification of tweets based on keyword baskets – Ordered by specificity: e.g. ‘carbon tax’  Environment; ‘tax’  Budget – Iterative testing and refinement – Tracking across timeframes and comparison across elections • Success rates: – 2013: 25% of tweets matched to themes / 2016: 34% of tweets matched to themes
  • 13.
  • 14.
  • 15.
    2016: Themes perDay (*): new in 2016
  • 16.
    Conclusion and Outlook •Longitudinal perspective: – Evolution in campaigning strategies: • Labor rediscovery of local candidates and doorknocking + social media approach – Counteracted by internal party turmoil: • Campaign plans disrupted by lack of enthusiasm and discipline following leadership coups – Ordinary users’ activities largely stable: • Focus strongly on governing party, and party leaders – Except for retweeting choices: • Significant changes between 2013 and 2016 – especially Labor retweeting • Further plans: – More detailed content and thematic analysis (quantitative and qualitative) – Network analysis of @mention / retweet interactions: evidence of polarisation?
  • 17.
    http://mappingonlinepublics.net/ @snurb_dot_info @brendam @socialmediaQUT – http://socialmedia.qut.edu.au/ @qutdmrc– https://www.qut.edu.au/research/dmrc This research is funded by the Australian Research Council through Future Fellowship and LIEF grants FT130100703 and LE140100148.