The linguistic features of bias in the news

Black, Amanda R (2024) The linguistic features of bias in the news. Doctoral thesis, Northern Arizona University.

Text Black_2023_linguistic_features_bias_news.pdf - Published Version Download (2MB)

Abstract

This dissertation aims to determine what linguistic characteristics can reveal about bias in news and whether variation in linguistic characteristics, namely grammatical and lexical features, happen systematically across clines of political leaning and extent (i.e., severity of bias). Through an examination of key features and keywords in news texts divided into groups of political leanings of left and right and divided into groups by extent (extreme bias and no bias), this aim is accomplished. This study demonstrates that media bias, namely gatekeeping, coverage, and presentation bias can be examined via corpus linguistic methods and suggests that lexical and grammatical information are contributing to perceptions of direction and extent of bias. Importantly, it also demonstrates ways in which a researcher can avoid making subjective decisions about bias by relying on exploratory methods of register variation. The research in this study is conducted through six major steps: (1) compilation of a corpus of newspaper publications for which topic and time are closely controlled, (2) the collection of reader perceptions on the extent and direction of bias of each collected newspaper text, (3) a nuanced examination of key grammatical features and a comparison of their functions across biased and unbiased texts, (4) an analysis and comparison of the keywords that occur in both biased and unbiased texts (5) a comparison of key features across texts perceived as right (a.k.a. conservative) to those perceived as left (a.k.a liberal) (6) and an analysis and comparison of keywords that occur in texts perceived as conservative to those perceived as liberal. Results suggest systematic linguistic differences in the following ways: key features of left leaning texts show that this group maintains a formal tone, involvement at the group level, and covertly reveals stance via non-finite clauses while the keywords reveal a concern for social issues (particularly those related to equality), a sentiment of aggression and political unrest, and a more frequent on Trump. Conversely, right leaning texts are colloquial in tone, involved at the individual level, and covertly reveal stance via reported speech. The keywords for the right leaning group demonstrate a concern for political issues (particularly illegal immigration and a loss of constitutional freedoms), the actions of former presidents, and a negative evaluation of the liberal left. Texts perceived as extreme in their bias include features marked for epistemic and attitudinal stance, clausal and phrasal elaboration, description, emphasis, and evaluation and the keywords reveal a concern for both political and social issues, political actors, and opposing negative evaluations of both democrats and republicans. Key features and keywords for texts rated as no bias include features related to reporting what happened (who, what, when, and where). The differences are, for the most part, attributable to a functional or evaluative difference related to direction or extent of bias.

Item Type:	Thesis (Doctoral)
Publisher’s Statement:	© Copyright is held by the author. Digital access to this material is made possible by the Cline Library, Northern Arizona University. Further transmission, reproduction or presentation of protected items is prohibited except with permission of the author.
Keywords:	Linguistic analysis; Political speech; News reporting;
Subjects:	P Language and Literature > PE English
NAU Depositing Author Academic Status:	Student
Department/Unit:	Graduate College > Theses and Dissertations College of Arts and Letters > English
Date Deposited:	23 Oct 2025 23:55
Last Modified:	23 Oct 2025 23:55
URI:	https://openknowledge.nau.edu/id/eprint/6276

Actions (login required)

IR Staff Record View

Downloads