A new study reveals how the words we use to describe political events can impact how trustworthy we perceive the speaker to be, and even shape our own opinions about the event itself. The research, published in the journal Cognition, shows that people tend to trust speakers who use language that aligns with their own political views. However, this same partisan language can damage trust and increase disagreement when it’s heard by those with opposing political beliefs.
Political polarization is a growing problem, not just in the United States, but globally. Researchers wanted to understand how the language used to describe political events might contribute to this increasing divide. We’re seeing more and more partisan language used by politicians, news organizations, and even everyday people on social media.
For instance, when describing the events at the U.S. Capitol on January 6th, many Republicans use the word “protest,” while many Democrats prefer terms like “insurrection” or “riot.” The new study aimed to isolate the impact of these specific word choices, examining how they affect both our perceptions of the speaker and our own understanding of what happened. The researchers hypothesized that individuals may be using this type of language more often to appear in a favorable light to people who already agree with them.
“Data shows that people increasingly harbour negative feelings towards their political opponents. A primary goal of this research is to assess how aspects of the current political climate—such as the prevalence of politically-biased language—contribute to deepening political divides,” said study author Alexander Walker, a Banting Postdoctoral Fellow at Brown University.
The research involved two experiments with a total of 1,121 participants from the United States, all of whom identified as either Democrats or Republicans. These participants were recruited online through services like Amazon Mechanical Turk and Prolific.
In the first experiment, participants were presented with descriptions of various politically charged events. Each event had three different descriptions: one using language favored by liberals (e.g., “expand voting rights”), one using language favored by conservatives (e.g., “reduce election security”), and one using neutral language (e.g., “expand mail-in voting”). These descriptions were presented as public statements made by fictional people. Alongside each statement, participants were also given a detailed, factual account of the event, which they were told to consider completely accurate. This was done to ensure everyone had a common understanding of the underlying facts, regardless of the partisan language used.
Participants were then asked to rate the speaker on several measures. They judged how trustworthy and moral the speaker seemed, how open-minded they appeared, and how much criticism they deserved. Participants also indicated how interested they would be in having a political discussion with the speaker, and what political party they believed the speaker belonged to.
The second experiment focused on how partisan language influences people’s own opinions about the events. Participants were presented with the same events, again described using either liberal-leaning, conservative-leaning, or neutral language. However, in this experiment, some participants received additional, detailed information about the event, while others received only the brief, partisan or neutral description. The key measure in this experiment was how much participants agreed or disagreed with the actions described in the event.
The results of the first experiment showed that people respond very differently to partisan language depending on their own political views. When speakers used language that matched the participant’s political leaning (in-group language), they were seen as more trustworthy, moral, and open-minded. Participants were also more willing to engage in a political discussion with these speakers.
However, the opposite was true when speakers used language associated with the opposing political party (out-group language). These speakers were viewed as less trustworthy, less moral, more closed-minded, and more deserving of criticism. Participants were also much less interested in talking politics with them. The more strongly a participant perceived a speaker as belonging to their own political party, the more positively they rated them.
“People view speakers describing political events in a way that aligns with their ideological leanings as highly trustworthy,” Walker told PsyPost. “Thus, as audiences become more politically polarized, people may increasingly view politically-biased news as more trustworthy than that which aims to be more neutral.”
“While describing events using politically-biased language was the best way to garner trust from political in-group members, describing these events using more neutral language was the best way to garner trust across political lines (i.e., from both Democrats and Republicans alike). Thus, despite the political leanings of our participants, they still both agreed that the presented politically neutral statements accurately described a set of contentious political events.”
The second experiment revealed that partisan language not only affects our view of the speaker, but also polarizes our own attitudes about the event itself. When events were described using language aligned with a participant’s political views, their opinions about the event became more extreme in that direction.
This effect was particularly strong when participants didn’t have much additional information about the event. When more details were provided, the polarizing effect of the partisan language was reduced, though it didn’t disappear completely. This finding implies that when people know more about the details of an event, they are less likely to let a single word or phrase dictate their overall opinion. Even so, the tendency to interpret events in a way that reinforces preexisting political beliefs remained evident.
“We find evidence suggesting that people are incentivized to describe political events in a manner that aligns with the political biases of their audience,” Walker said. “We also show that exposure to politically-biased descriptions of political events a) promotes negative evaluations of political out-group members and b) shapes how people view political events.”
But as with all research, there are some caveats to consider. The scenarios used in both experiments, while carefully designed, represented only a subset of the many possible events and linguistic choices that occur in everyday political discourse. The researchers note that large-scale studies using real-world data from social media, political speeches, or news reports could help confirm and expand on these findings.
“One limitation of the current work is that, unlike in real-world contexts, participants in our study did not have information about the people describing political events,” Walker noted. “This allowed us to isolate the impact of politically-biased language on peoples’ political attitudes and judgments of trustworthiness (e.g., trust in speakers describing a political event). However, it is an open question whether knowledge of a person’s or organization’s political leanings limits the persuasive influence of partisan rhetoric.”
The study highlights an important societal challenge: even when information is objectively accurate, the way it is framed can have powerful effects on public opinion and interpersonal trust. While much attention is often focused on the dangers of misinformation, this research demonstrates that even truthful language can contribute to polarization if it is presented in a way that appeals to partisan identities.
“Much attention has been paid to the negative impact of misinformation,” Walker said. “However, we show that information need not be objectively false to facilitate distrust of political opponents or polarize opinion across party lines.”
“A long-term goal of this work is to better understand growing partisan divides by understanding the interplay between peoples’ political environment and political attitudes. For example, does politically-biased (yet not objectively false) news make people more extreme in their ideological attitudes? Do these more extreme attitudes result in people viewing politically-biased information as most trustworthy? If so, how can this divisive cycle be interrupted?”
The study, “Partisan language in a polarized world: In-group language provides reputational benefits to speakers while polarizing audiences,” was authored by Alexander C. Walker, Jonathan A. Fugelsang, and Derek J. Koehler.