News, Persuasion, and Influence: American news media’s impact on public opinion of transgender people

Abstract

Transgender rights are under attack in the United States — nearly 300 anti-trans bills have been introduced in state legislatures nationwide in the first quarter of 2024, seeking to take away any number of rights from transgender people. As anti-trans rhetoric continues to proliferate, this thesis seeks to understand one method by which that rhetoric is built — the news. While previous studies have analyzed news coverage of transgender people and the impact of news on public opinion, no prior study has sought to examine the two in conjunction and draw conclusions from both together. This study uses two methods to do so: a content analysis in order to determine the variance in the emotional language used by different outlets, and a survey-based experiment in order to determine the impact of different language codings on the opinions held by readers on the trans community. For the content analysis portion of this thesis, I posed two hypotheses. Firstly, I hypothesized that news outlets would have an overall negative slant in their coverage of the transgender community. In conjunction with this, I hypothesized that, given that the study focused on six mainstream, legacy outlets, the emotions used in the articles would not vary significantly between outlets. The first hypothesis was demonstrated to be true, as five of the six outlets had an overall negative positivity score with the analysis methods used. The second, however, proved less true, as for seven of the eight emotional metrics measured, there was some statistically significant difference between the outlets. For the survey-based experiment, which used a 2x2 structure of stimuli — different versions of a fictitious news article — in the treatment groups, two further hypotheses were tested. The first was that the two treatment groups that used positively-coded language would result in a more positive public opinion than the groups that used negatively-coded language or the control group, and the second was that the two treatment groups that used personal language — that is, the inclusion of a quote — would have more positive public opinion than those with impersonal language (without a quote) or the control group. However, the experiment yielded largely statistically insignificant data, and as such, neither of these hypotheses could be proven to be true. However, given these results, I put forth one main potential conclusion — that while this thesis demonstrates a lack of persuasive ability of individual news articles, it does have potential implications for the broader news landscape and its treatment of the transgender community. The two main emotional pulls present in the articles analyzed in this thesis are surprise and anger — an important factor in the context of a broader sense of public outrage about changes in acceptable language around trans people. Even with statistically insignificant data in the experiment, news outlets and the people who work for them can work to make their writing more inclusive by following existing best practices in coverage of the transgender community.