Right Wing Media Radicalization and Gendered Harassment

• In Event: The Negative Effects of Traditional Media

Abstract

On his nightly broadcast on March 9, 2021 Tucker Carlson verbally attacked then New York Times journalist Taylor Lorenz. In the following days, Lorenz received a torrent of toxic speech directed at her on Twitter. This case, as well as others, prompts the three main questions of our paper: First, who do news commentators attack and how does this vary by media partisanship? Second, how much do these legacy media events drive toxic speech on social media and silence the targets of these attacks? And most centrally, to what extent are the incidence and effects of these attacks a gendered phenomenon?

In this paper we leverage a random sample of news commentator transcripts from three Fox and three MSNBC primetime news commentator shows (Tucker Carlson, Laura Ingraham, Sean Hannity, Rachel Maddow, Chris Hayes, and Lawrence O’Donnell). These transcripts were randomly sampled from a population of approximately 3,000 broadcasts by these six commentators aired between February 2020 and February 2022. Each transcript was hand coded by two research assistants, who identified within each transcript all individuals who were discussed by the commentators and or their guests. We also coded each mention by whether it was neutral or an attack and identified the gender of the person mentioned, amongst other attributes. After identifying 1,020 such instances where an individual was discussed across the 120 news transcripts, we linked the discussed individuals with Twitter data. We identified all posts written by the discussed individual around the time the commentator show aired. We also identified all posts which included the discussed individual’s twitter handle in a random sample of ten percent of all tweets on the Twitter platform around the time the broadcast aired.

Existing literature on media and gender suggests that we would observe a gendered pattern to both the incidence of commentator attacks and their effects. While women are typically underrepresented in news (Mitchelstein et. al. 2019; Shor et. al. 2015; Shor et. al. 2019), when they do appear in news sources they are more likely to be negatively portrayed (Shor et. al. 2022; Smith et. al. 2021), as their elevation to news worthiness violates gendered status expectations (Ridgeway, 2013). This theory would predict that women are more likely to be attacked by news commentators than men and, if they were attacked, would be more likely to face toxic speech on social media. We expect this gender difference to be particularly strong for high status women such as politicians and government officials.

We also develop a second set of hypotheses from the literature on right wing media radicalization (Benkler et. al. 2018). We expect the incidence of attacks to be much greater on Fox than MSNBC. We also expect that any gendered effects of toxic speech on Twitter may be exacerbated for right wing attacks, given the greater prevalence of misogynistic discourse on the right.

We find that Fox News commentators are, on average, much more likely to attack women than MSNBC commentators. Fox commentators attack the women they discussed on the show 73 percent of the time, versus 15 percent of the time for MSNBC commentators. This difference is particularly stark for high status female politicians and government officials (82.1% attacked for Fox versus 13.6% for MSNBC). This gap, however, is in part driven by the greater overall tendency for Fox commentators to attack any individuals. Fox commentators attacked 64.2% of the individuals they mentioned on their shows, as compared with 40.2% of individuals mentioned on MSNBC shows. This finding points to how right wing media radicalization has mapped onto an increase in toxic speech, even in legacy media television shows. This radicalization also, however, has a gendered nature.

We investigate the effects of these gendered attacks on social media participation and the experience of toxic speech on these platforms. We measure whether attacked individuals were targeted with bursts of toxic speech on the platform, as well as estimate changes in these individuals’ social media participation following the attack. We leverage a difference-in-difference design comparing attacked individuals and neutrally mentioned individuals. We find initial evidence for a chilling effect of news commentator attacks, especially for women who are not politicians. After being attacked, women who are not high status politicians and government officials were 14 percentage points less likely to tweet in the days following the attack. These findings point to the second-order implications of right wing media radicalization for social participation: it can drive a segment of the population away from participating in political discussion.

Authors

• Hannah Waight, New York University

• Megan Brown, University of Michigan

• Zeve Sanderson, New York University

• Joshua A. Tucker, New York University

• Jonathan Nagler, New York University

• Felicia Loecherbach, New York University