The Hashtag Ceiling: Gender Bias during the 2020 U.S. Presidential Primaries

• In Event: R-E-S-P-E-C-T: Gender Identiy, Bias, & Discrimination in Political Communication

Abstract

There is substantial evidence suggesting people use different words, phrases, and ideas when discussing female candidates and politicians versus their male counterparts. There is further evidence that such differences are more pronounced for female presidential candidates than they are for female congressional candidates. More needs to be known regarding how network effects modify, amplify, or mute the reach and impact of such language in networked communication environments. This project seeks to fill this gap by examining the network effects of gendered language within two distinctly different types of competitive political environments: the 2018 U.S. Senate Midterms and the 2020 U.S. Democratic Presidential Primaries.
First, I examine four competitive elections for U.S. Senate seats during the 2018 U.S. midterms featuring an equal distribution of male and female candidates from each party while drawing upon an original data set of roughly 5 million tweets gathered daily over a two-month period. Second, I analyze several large data sets of more than 1 million tweets each gathered live during the 2020 Democratic Primary Debates.

I employ several tests on these rich data sets, including content analysis, time-series analysis, network visualizations, and metrics analyses. Further steps are taken to estimate the party identification of users in order to assess the potential effects of political polarization in these networks. This research contributes to the political communication literature by providing important insight into how social networks shape gender bias where and when it exists.

Author

• Eric C. Vorst, University of Missouri – St. Louis