🧪 This platform is in early beta. Features may change and you might encounter bugs. We appreciate your patience!
word_choice
Word choice bias occurs when journalists or editors select terminology that encodes political, moral, or evaluative judgments — without stating those judgments explicitly and without attributing the language to a source who holds those views. The choice between 'freedom fighter' and 'terrorist,' 'pro-life' and 'anti-abortion,' 'spending' and 'investment,' 'illegal alien' and 'undocumented migrant' is not neutral: each set activates different associations and frames the reader's understanding before the facts are presented.
Two stories about the same protest use different language: one describes 'demonstrators expressing their concerns,' another describes 'agitators disrupting the city.' Both may be factually defensible; the chosen language encodes opposite moral evaluations without making any explicit evaluative claim.
Immigration stories in different outlets describe the same group as 'refugees fleeing persecution,' 'migrants seeking economic opportunity,' or 'illegal aliens.' Each label encodes a different legal status, moral claim, and policy implication — yet all are presented as neutral description.
Stories about economic policy use loaded synonyms: 'investing in public services' vs. 'government spending,' 'tax relief' vs. 'tax cuts for the wealthy,' 'deregulation' vs. 'removing worker protections.' The chosen phrase arrives in the neutral register of news language while carrying the full weight of one side's framing.
Binary (yes/no) questions an LLM must answer to identify this aspect:
Does the coverage use terminology that carries political, moral, or evaluative connotations beyond its neutral descriptive meaning?
Type: binaryWould a different but equally accurate word choice produce a different emotional or political impression of the same facts?
Type: binaryIs the chosen terminology consistently aligned with one political or ideological perspective across stories?
Type: binaryIs the loaded terminology used in news contexts (not labelled opinion) without attribution to a source who uses that language?
Type: binaryWord choice bias occurs when journalists or editors select terminology that encodes political, moral, or evaluative judgments — without stating those judgments explicitly and without attributing the language to a source who holds those views. The choice between 'freedom fighter' and 'terrorist,' 'pro-life' and 'anti-abortion,' 'spending' and 'investment,' 'illegal alien' and 'undocumented migrant' is not neutral: each set activates different associations and frames the reader's understanding before the facts are presented.
Framing effects are well-documented: semantically equivalent descriptions of the same facts produce different judgments when different language is used. Word choice operates pre-analytically — the framing is absorbed before the audience consciously evaluates the content. Loaded words prime interpretations that persist even when audiences try to reason through them.
Notice when terminology carries political associations and ask whose language is being adopted. Is the loaded term attributed to a source — or is the outlet using it as its own? What would a neutral equivalent look like? Does the word choice apply consistently across comparable groups or events?
One of the most studied forms of media bias. Classic examples: 'estate tax' vs. 'death tax,' 'enhanced interrogation' vs. 'torture,' 'pro-choice' vs. 'pro-abortion,' 'freedom fighter' vs. 'terrorist.' Research shows these choices significantly affect audience attitudes on the underlying policy.
Use these tools to detect, analyze, or train this aspect.