Political Campaigns

Affective polarization on Reddit

The emergence of affective polarization and its origin is a much-discussed topic that plays a particular role in politics all over the world. Survey-based research confirmed the existence of affective polarization. Social media is often named as one of the drivers of polarization in general. For that, we want to measure online affective polarization in the context of political discussions on Reddit. We define a data set for this question, identify partisan users, categorize discussions as in- or out-group conversations and measure the sentiment for the different classes of conversations. Using this approach, we reject our research hypothesis for the present data set. Moreover, we find that the sentiment in conversations of users favoring different parties is more positive than conversions of users that almost all favor the same party. This opens several other questions about the existence of online affective polarization on Reddit and on other social media platforms.

Median attractor model

We introduce an attractor model of political competition in which the distribution of voters' ideal points is endogenous. Two candidates for office select policies they want to pursue in office. They also select a range of voters they attempt to persuade from the merits of their policies. Finally, the opinion formation process where each voter is influenced by neighboring opinions determines the distribution of ideal policies. Our main insight in this paper is that despite the endogeneity of the distribution of voters' ideal points, candidates adopt policy positions equal to the median voters' ideal point before the communication processes take place. This result is called the median attractor result. We also show that candidates choose smaller influence ranges to attract voter groups as such strategies trigger the most favorable shift of the voter's ideal distribution towards the candidates' policy positions.

The impact of technologies in political campaigns

Recent political campaigns have demonstrated how technologies are used to boost election outcomes by microtargeting voters. We propose and analyze a framework which analyzes how political activists use technologies to target voters. Voters are represented as nodes of a network. Political activists reach out locally to voters and try to convince them. Depending on their technological advantage and budget, political activists target certain regions in the network where their activities are able to generate the largest vote-share gains. Analytically and numerically, we quantify vote-share gains and savings in terms of budget and number of activists from employing superior targeting technologies compared to traditional campaigns. Moreover, we demonstrate that the technological precision must surpass a certain threshold in order to lead to a vote-share gain or budget advantage. Finally, by calibrating the technology parameters to the recent U.S. presidential election, we show that a pure targeting technology advantage is consistent with Trump winning against Clinton.