Fake news has permeated online media, presenting consumers with the challenge of detecting it. At what age are we capable of undertaking this challenge? And what factors predict success? We explored these questions with elementary-school-aged children (n = 86), who were asked to judge the veracity of ten news stories, five fake and five real. Children also completed a developmental version of the cognitive reflection test (CRT-D; Young & Shtulman, 2020a). As a group, children were at chance at differentiating fake news from real news, and their individual performance did not vary by age or cognitive reflection. Adults (n = 271) given the same materials succeeded at detecting fake news, especially those high in cognitive reflection. These results suggest that children lack the knowledge or skill needed to evaluate news credibility and that cognitive reflection predicts fake news detection only after we have attained some baseline level of information literacy.

Research suggests children readily treat robots as social actorsand sources of information for learning. Here we ask if childrenuse depictions of gender-counterstereotypic robots (e.g., afemale construction worker robot) and gender-stereotypicrobots (e.g., a female secretary robot) as sources of informationabout cultural gender stereotypes. Forty-five 6- to 8-year-oldchildren participated in a short counterstereotyping task.Children in the counterstereotypical condition viewed videosof cartoon female gendered robots with culturally stereotypedmasculine occupations, interests in activities, and traits.Children in the stereotypical condition viewed videos ofcartoon female gendered robots with culturally stereotypedfeminine attributes. Children completed a measure of genderstereotyping before and after viewing the intervention videos.From pretest to posttest, children’s gender stereotypingdecreased in the counterstereotypical condition and increasedin the stereotypical condition. These finding suggest childrenmay learn from robots as models of cultural gender stereotypes.

The cognitive reflection test (CRT; Frederick, 2005) assesseshow well adults can reflect on the validity of their ownthinking, and it has been shown to predict several measures ofnormative reasoning. Here, we sought to create a version ofthe cognitive reflection test suitable for elementary-school-aged children, which could be used to study the emergence ofcognitive reflection as well as its role in the development ofother forms of higher-order cognition. We identified eightchild-friendly questions that elicit an incorrect, intuitiveresponse that must be inhibited in order to provide a correct,analytic response. We compared children’s and adults’performance on these questions (dubbed the CRT-D) toseveral measures of rational thinking (denominator neglect,base rate sensitivity, syllogistic reasoning, otherside thinking)and thinking dispositions (actively open-minded thinking,need for cognition). The CRT-D was a significant predictor ofrational thinking and normative thinking dispositions in bothchildren and adults. Moreover, performance on the CRT-Dcorrelated with performance on the original CRT in adults,and in children, it predicted rational thinking and normativethinking dispositions above and beyond age. These resultssuggest that the CRT-D is a valid measure of children’scognitive reflection and pave the way for future investigationsof its development and its developmental consequences.

Adults with extensive science education exhibit cognitiveconflict when reasoning about counterintuitive scientificideas, such as whether clouds have weight or whether bacterianeed nutrients. Here, we investigated whether elementary-school-aged children show the same conflict and whether thatconflict can be reduced by targeted instruction. Seventy-eight5- to 12-year-olds verified, as quickly as possible, statementsabout life and matter before and after a tutorial on thescientific properties of life or matter. Half the statements wereconsistent with intuitive theories of the domain (e.g., “frogsreproduce”) and half were inconsistent (e.g., “cactusesreproduce”). Participants verified the latter less accurately andmore slowly than the former, both before instruction andafter. Instruction increased the accuracy of participants’verifications for counterintuitive statements within thedomain of instruction but not their speed. These resultsindicate that children experience conflict between scientificand intuitive conceptions of a domain in the earliest stages ofacquiring scientific knowledge but can learn to resolve thatconflict in favor of scientific conceptions.

Scientific ideas can be difficult to affirm if they contradictearlier-developed intuitive theories. Here, we investigatedhow instruction on counterintuitive scientific ideas affects theaccessibility of those ideas under time pressure. Participants(138 college undergraduates) verified, as quickly as possible,statements about life and matter before and after a tutorial onthe scientific properties of life or matter. Half the statementswere consistent with intuitive theories of the domain (e.g.,“zebras reproduce”) and half were inconsistent (e.g.,“mushrooms reproduce”). Participants verified the latter lessaccurately and more slowly than the former, both beforeinstruction and after. Instruction did, however, increaseaccuracy for counterintuitive statements within the domain ofinstruction, but changes in accuracy were not accompanied bychanges in speed. These results confirm the conclusion drawnfrom studies with professional scientists that scientific ideascan be prioritized over intuitive ones but the conflict betweenscience and intuition cannot be eliminated altogether.