Tweeting the Party Line: Did the 2016 Presidential Election Impact Legislative Communication?

Abstract

Given the ever-increasing importance of social-media platforms to the American political system, this thesis aims to make a novel contribution to the literature on legislative communication by analyzing whether senators modify their usage of Twitter in response to new information about their electorate. Specifically, senators’ Twitter accounts will be analyzed before and after the 2016 presidential election to see whether they changed their communication style in response to how their constituents voted. Automated tools are used to collect the hashtags used by senators in the 15 months preceding and following the 2016 election. A Pearson’s chi-squared calculation is used to create subsets of the most highly “politicized” phrases that represent each party (i.e. “Democratic buzzwords” and “Republican buzzwords”). Senators’ overall hashtag usage is then analyzed to see how frequently they use these politicized hashtags, which helps quantify the degree to which they are “tweeting the party line.” Senators will be treated as the unit of analysis, and are stratified by party to account for the fact that the outcome of the 2016 election placed different electoral pressures on Democrats and Republicans. The independent variable is how support for the senators’ party changed in their states from the 2012 presidential election to the 2016 presidential election. The dependent variable is how the senators’ usage of politicized hashtags changed between the pre- and post-election periods. Additionally, two interactions variables are included. The first is a senator’s margin of victory in their most recent race, which will help shed light on whether they modify their communication styles due to a re-election motive. The second is the proportion of a senator’s donations from that came from within their state, which will help shed light on whether they are driven by a desire to pander to their donors. The results of the analysis found a statistically significant and materially substantive relationship between how Democratic senators’ constituencies changed in their voting patterns from 2012 – 2016 and their usage of Democratic buzzwords in the aftermath of the election. However, no relationship was observed for their usage of Republican buzzwords, and the interactions variables did not have a significant impact. No relationship was observed for Republican senators’ usage of any of the politicized hashtags, even when including the interactions variables. These results show that Democratic legislators have responded to the results of the 2016 election in their state by trying to align with their constituents, which is notable as it rebuts the notion that the entire Democratic party has been uniformly liberalizing in the wake of the election. These results also show that Republican senators have not significantly changed their social media usage in response to the 2016 election, which may indicate that they do not view Trump’s victory as worrisome for their future electoral prospects. More generally, these results provide evidence that social media is a meaningful component of legislative communication, and provide some of the first evidence that senators will modify their communication style even in response to electoral information that does not directly pertain to them.