AI Insight
This study investigated how social authority labels affect people's confidence in linguistic judgments and subsequent translation revision behavior. Researchers found that when translation segments were labeled as produced by an "expert," participants showed overconfidence in erroneous translations and were more likely to leave mistakes uncorrected, while "peer" labels led to excessive skepticism and unnecessary edits to acceptable translations. The miscalibration between confidence and accuracy partially explained these biased revision patterns, suggesting that authority cues disrupt people's ability to accurately monitor the reliability of their own judgments.
Why it matters
These findings have practical implications for professional translation, editing, and any collaborative writing process where source credibility influences decision-making. Understanding how authority labels bias metacognitive monitoring could help develop training interventions or workflow designs that reduce under-editing of expert-attributed content and over-editing of peer-attributed work.
Understand the Science
IntroductionLinguistic judgment rarely occurs in a social vacuum. This study examined whether social authority cues reshape the alignment between confidence and decision accuracy, known as metacognitive calibration, and subsequently influence linguistic revision behavior.MethodsUsing an English-Chinese revision paradigm, participants evaluated translation segments labeled as produced by an “expert,” a “peer,” or an anonymous source. We examined perceived source credibility, linguistic quality judgment, metacognitive calibration, and revision behavior, including under-editing of erroneous segments and over-editing of acceptable segments.ResultsAuthority cues significantly influenced perceived credibility and linguistic quality judgments. Expert labels increased overconfidence when participants evaluated erroneous segments and were associated with a higher likelihood of under-editing. Conversely, peer labels induced skepticism and were associated with unnecessary revisions to acceptable segments. Mediation analysis indicated that calibration bias partly accounted for the associations between authority labels and revision behavior. However, because confidence and revision decisions were measured within the same segment-level task, these indirect effects should be interpreted as statistically consistent explanatory pathways rather than definitive evidence of temporally ordered causal mechanisms.DiscussionSocial authority cues may affect not only linguistic judgments but also individuals’ monitoring of the reliability of those judgments. These findings demonstrate the vulnerability of metacognitive calibration to source information and provide a psychological framework for understanding biased intervention in professional communication.