Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain
the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in
Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles
and JavaScript.
Corrective comments posted by social media users that suggested a news story was incorrect reduced accuracy perceptions and engagement with the news posts. These corrective comments had similar effects regardless of the truthfulness of the original post.
Analysing geo-tagged Twitter data from 2010-2020, this study shows that media reports of police brutality affect public sentiment towards police in the US, while there is no statistically significant evidence that reports of local crime has similar effects.
People learn to exert more control after conflict detection, when stimuli associated with conflict are selectively reinforced, providing evidence for reinforcement learning of abstract cognitive control adaptations.
A large-scale longitudinal study analysing Ukraine-based Twitter (now X) accounts’ language use prior and during the Russian invasion of Ukraine, shows a behavioural shift away from Russian language to Ukrainian language.
A ten-minute, unmoderated chat discussion between pairs of matched Labour and Conservative party voters is followed by increased out-partisan sympathy and lasting willingness to engage in discussion.
Visual working memory models combining deep neural network features with the Target Confusability Competition model capture human memory errors on GAN-generated scenes. Layers from the DNN architectures reproduce set-size effects and response bias curves in orientation and colour.
Speech and music processing appear to involve partially distinct rhythmic timing mechanisms. A perception and a synchronization task show distinct optimal temporal rates for perception and production of music and speech.
Replicating a finding previously established in German-speaking and English-speaking cohorts, Norwegian-speaking participants likewise form a bimodal distribution of high synchronizers and low synchronizers.
When people take false news for true, they are often aware that they might be wrong. Here, Democrats and older adults were better at telling true from false news than Republicans and younger adults, but all partisan and age groups had good insight into their abilities.
Individuals vary in how often they detect changes in the emotions of others and in whether these emotional events align with what other people perceive. We find that more complex emotion vocabulary and knowledge of emotion predict this complex skill, based on a newly developed emotion segmentation paradigm.
When judging the pleasantness of a physical touch experienced by someone else, mothers are more likely to ascribe their personal feelings of pleasantness to their own children as compared to the children of others.
Across 34 countries, fear of disease and empathic concern encouraged adherence to COVID19 preventive measures. When trust in government is high, the link is weaker between preventive behaviour and fear but stronger between preventive behavior and empathic concern.
Spatial information is a primary organizing pillar of visual working memory and guides behaviour even when object locations are irrelevant and temporal information should be prioritized.
People are more adept at deriving and retaining associative inferences when the source of the information is a member of one’s own ingroup, potentially contributing to the spread of biased beliefs.
Across four studies with US American participants, hostility in tone of messages that rebut science denialism had small or trivial effects on their persuasiveness but was to the detriment of the perceived competence of the source.
Analysis of two longitudinal datasets from the USA and the Netherlands shows that people who hold extreme views are more likely to change these views than those who hold moderate views. The direction of change is towards moderation.
Across a sample of 1,600 individuals spanning 6 to 59 years of age, common saliency models of visual attention predict women’s and young adult’s fixations better than the gaze behaviour of other groups.
When people receive advice written by large language models, they downrate the competence of the source when they know the source isn’t human. Their preference to receive advice by large language models increases with positive experience.
Across two experiments, exposure to a brief informative video on Japanese American incarceration and subsequent reparations led to greater support among Asian Americans for the payment of reparations to Black Americans.
Native English and Spanish speakers performed significantly better than chance when identifying the meanings of unfamiliar foreign language words, suggesting that the sound of words holds some association with their meaning across languages.