The Queensland (QLD) state general election on January 31 provides some interesting statistics and valuable lessons about the importance of social media. The Liberal National Party (LNP) failed to win a second term, while the Australian Labor Party won 44 of the 89 seats and is now governing Queensland with the support of independent MP Peter Wellington. An interesting feature of this election is that a large number of candidates campaigned via social media. In particular, 68.6% of candidates used a business/community Facebook page to run campaigns. The role that social media played in this election should not be understated.

Social media presence is highly correlated with election success

What is striking about the QLD state general election is that social media presence is highly correlated with election success. 92.1% of the 89 winners have a Facebook page, whereas only 62.5% of the other 344 candidates have a Facebook page. If we compare the probability of winning among those with a Facebook page and those without, we see that 27.6% of candidates with a Facebook page won an election, whereas only 5.1% of candidates without a Facebook page won an election. Candidates who have a Facebook page are 5 times more likely to win an election than those who don’t!

Figure 1

figure1

These statistics are obviously correlations. They do not imply that simply setting up a Facebook fan page will significantly change the election outcome. Candidates who are more resourceful, strategic, better organized, and interested in interacting with people are probably more into social media and at the same time popular to voters. These factors are likely more significant for their election success than having a Facebook fan page. Nonetheless, the power of social media should not be underestimated.

According to statistics compiled by SocialMediaNews.com.au in February 2015, Facebook has 13.8 million users in Australia. This represents approximately 60% of Australia’s population, giving political candidates a wide-reaching platform to effectively campaign and get their messages across to more people.

A 2014 report by Pew Research found that the vast majority of both Facebook users (86%) and Twitter users (77%) see some political content, such as posts from friends, news organisations and political leaders, and links to news stories. Moreover, 24% of Twitter users and 19% of Facebook users, a substantial percentage of each user group, report that political posts represent at least half of the posts they see. This demonstrates the successful shareability of political content on Facebook and Twitter, and potentially allows candidates who have the ability to master communication in these channels an upper hand in garnering support and votes.

Social media may also serve as a useful campaigning and monitoring device

When we correlate the relationship between Facebook page likes and primary votes or two-candidate preferred votes, a statistical and positive relationship emerges in each instance. The relationships are particularly strong when we exclude from the analysis leaders who command page likes outside of their electoral districts (e.g., Premier Campbell Newman, opposition leader Annastacia Palaszczuk, and controversial One Nation Party leader Pauline Hanson). On average, every additional Facebook page like translates into 3.1 primary votes and 1.1 two-candidate preferred votes if we compare candidates across all districts. If we compare the relative performance of candidates within electoral districts (rather than correlation across all candidates and all districts), then on average, every additional Facebook likes translates into 4.1 primary votes (Figure 2A) and 1.8 two-candidate preferred votes (Figure 2B).

Figure 2A

figure2a

Notes: 433 observations shown in Figure 2A. Solid dots are districts used to fit a linear relationship between deviations in primary votes from the average primary votes within district and deviations in Facebook page likes from the average Facebook likes within district. Hollow dots are observations of candidates who have no Facebook page; hollow dots with candidate names and district names labelled are districts where one of the leaders contested. The fitted line (using fixed-effects regression technique) has a slope of 4.1, meaning that for every extra facebook page like relative to the district average, two-candidate preferred votes increase by 4.1 relative to the district average.

Figure 2B

figure2b

Notes: 89 observations shown in Figure 2B. Solid dots are districts used to fit a linear relationship between differences in two-party preferred votes and differences in Facebook page likes between two candidates. Hollow dots are are districts with at least one candidate having no Facebook page; hollow dots with candidate names labelled are districts where one of the leaders contested. The fitted line (using first-difference regression technique) has a slope of 1.9, meaning that for every extra facebook page like relative to the rival’s, two-candidate preferred votes increase by 1.9.

Candidates’ social media performance gives strong indication of offline performance

Given these statistical relationships, it is not too surprising why ALP won the election with 44 seats against LNP’s 42 seats. ALP candidates did better than LNP candidates on social media. 94% of ALP candidates have a Facebook page, but only 85% of LNP candidates do. Among the non-party leader candidates who have a Facebook page, ALP candidates received 1125 likes on average, while LNP candidates received only 1052 likes on average. Candidates’ performance on social media says a lot about their ‘offline’ performance.

What do all these mean?

The results indicate that politicians (and also businesses and other types of organisations) can use Facebook page likes to infer whether their campaigns are going well on the ground. Social media provides instant feedback on their performance and allows them to target audience, modify strategies, and vary resources to achieve the desired outcomes. If a candidate has more genuine Facebook page likes than the rivals, it’s very likely the candidate will also win in an election. The same can be said for businesses. When a business is doing better than its competitors on social media, it’s likely that the business is also outperforming its competitors in sales.

By Choon Wang and Yuan Wang