A growing number of studies addressed the relationship between candidate use of social media (particularly Twitter) and preference voting, suggesting that those who embedded social media in their campaign strategy perform better than those who have not. This holds across a variety of countries and political systems.
Plausibly, this relationship could be a function of better, richer and more experienced candidates using Twitter more than the others. A second possibility is that the effects of social media travel through the activity of journalists. In this case, tweets merely have an indirect effect and provide journalists with one-liners to build their news stories. Here social media are considered, so to speak, nothing more than a new door into an old house. Alternatively, there could be a genuine direct effect of social media usage on the number of preference votes because candidates can use social media to interact with voters directly. We seek to shed some light on whether the effects of Social Media usage are direct, indirect or spurious.
We use data from the May 2014 Belgian that allow estimating the effects of social media campaign on preferential voting. This electoral appointment is of particular interest because on the same day Belgian voters chose their representatives to the regional, federal and European parliaments. We captured candidates twitter feeds during the campaign and we merge this with information of campaign spending, incumbency status and demographics. We also avail of data on the activity of journalists on Twitter, which allows addressing what best explains campaign effects on preferential voting. We aim at disentangling the effects of traditional media coverage, the interplay between old and new media to clarify the extent to which these overshade direct new media effects.
