The 2016 US presidential election was close – and while traditional polls favored Hillary Clinton, the probabilistic polls (shown above) had Donald Trump ahead for most of the year, including the last 2 weeks before the election (but note the uncertainty expressed in the grey confidence band).
But which topics and channels drove the probabilistic poll? Our analysis of the 2016 US presidential election showed that news coverage was ten times less influential than fake news (disinformation). The most impactful news topic was the continued coverage of rivalry between Hillary and Bernie fans. In contrast, the key topic of social media disinformation involved Clinton’s emails, while disinformation about her connection to Muslims didn’t hurt, and actually helped her chances in some demographics (voter reaction by demographics).
How can campaigns drive the election outcome? Through a wise use of topics and platforms: stick to important points of difference with your opponent, but tailor the tone to the medium. Trump effectively used TV to highlight terror threats, while consistently trolling his rival on Twitter, bypassing the typical gatekeepers of media and party apparatus. As shown in below figure, Clinton got a boost from discussing women’s issues on Instagram but not on Twitter. A key damaging factor was negative moral language about lack of authority and virtue for Clinton on her own Facebook page. Thus, our findings reveal an important caveat to the 2016 advice to use any and all social media platforms.
What is our advice for 2020? Candidates should choose their topics and platforms wisely: TV is great for generally important topics to voters, such as the Economy and Health – while Instagram is great to rally support of fans and Twitter for stirring sentiment against opponents. However, candidates must avoid demeaning statements and work towards uniting across fractions – note how continued news coverage of Clinton vs Sanders damaged the Democratic candidate. At the same time, campaigns should monitor negative comments on their owned social media and take appropriate action. In contrast to commercial brands, where negative comments help rally the troops to the brand’s benefit (Ilhan et al. 2018), disinformation on a political brand can confuse fans and discourage them from voting.
What do you think? Let us know in reaction to the full paper with all specifics about the data and analysis: http://marketingandmetrics.com/how-social-media-drove-the-2016-us-presidential-election-a-longitudinal-topic-and-platform-analysis/