New study “Prevalence of Prejudice-Denoting Words in News Media Discourse: A Chronological Analysis” by David Rozado from ECL -Otago Polytechnic in New Zealand, Musa Al-Gharbi from Columbia University in NY, and Jamin Halberstadt from Otago University investigated the prevalence of prejudice-denoting words from a large corpus of news and opinion articles.
The articles were published from 1970 to 2019 and included in total 27 million news and opinion articles. They represented 47 different news outlets. For many outlets, the articles skewed toward newer ones due to availability issues. The news outlets were largely US-based.
The words that were sought after were not the prejudicial words themselves, but words denoting the existence of a prejudice, such as “racism”, “sexism”, or “anti-semitic”, and variations of them. 16 terms were analyzed with the focus on how the frequency of use has changed over time, and how the political leanings of the outlets affects the frequency.
Generally, it was shown that the frequency of prejudice-denoting words has grown significantly over time, with a particular uptick in the post 2010s period. Both written news outlets and cable news outlets were affected.
The “agenda setting” hypothesis was investigated. The results were in line with the previous literature on the subject: there appeared to be a correlation, even if the causal link cannot be claimed, between the frequency of prejudice-denoting words and public perceptions of prejudice in the US. Furthermore, some influential news outlets such as The Washington Post, The New York Times, and Bloomberg had a predictive influence on the usage in other news outlets: they were the agenda setters.
The authors also ponder as an alternative explanation whether Western societies have grown less tolerant of forms of discrimination, or whether the frequencies reflect other cultural shifts. Also of note is the U.S. presidential election in 2016, with the growing amount of prejudicial discourse and thus the need for prejudice-denoting words.
Limitations of the study, as the authors admit, include the lack of context in which the words are used. Frequency count alone cannot provide that. Also a limitation, as alluded, is the lack of causal directions that can be inferred from the data.
The study “Prevalence of Prejudice-Denoting Words in News Media Discourse: A Chronological Analysis” by David Rozado, Musa Al-Gharbi and Jamin Halberstadt is in Social Science Computer Review. (free abstract).
Picture: Untitled by Amador Loureiro