This project developed new strategies to facilitate efficient scientific reasoning in young children. From early in development, children use categories to learn about the world around them. For example, upon learning something about an individual animal, children can generalize this information to the category as a whole and use this information to guide behavior. In this way, if a child learns that a particular spider will bite, the child might infer that all spiders bite, and use that information to avoid spiders in the future. A critical challenge of this type of reasoning is determining whether a limited sample of evidence--for example, a specific spider observed on a particular day--provides information likely to be true of other category members as well. Young children often evaluate samples of evidence much less efficiently than older children or adults, which can impede their biological, social, and scientific reasoning. This project enhanced basic understanding of how and why inductive learning changes across childhood. Through a series of experiments, this project revealed the cognitive and developmental mechanisms that underlie age-related changes in children's reasoning strategies. This project also used this new knowledge about cognitive development to devise and test new strategies for facilitating efficient reasoning in children. By developing new educational strategies, partnering with a major informal educational institution, and providing training opportunities for undergraduate and graduate students, this project maximized opportunities for this research to influence educational practices and benefit educators, children, and future researchers.

Young children often fail to consider the extent to which a sample provides diverse representation of a category when deciding if the sample provides generalizable information. For example, whereas adults view relatively diverse samples (e.g., a robin, penguin, and eagle) as more informative than less diverse samples (e.g., three robins), children below the age of eight years view such samples as equally informative. The present studies examined why this is the case, by testing the hypothesis that instead of trying to obtain diverse samples of evidence, young children seek out what they view as the best possible evidence. From this perspective, children have a particular view of what a category should be like, and they seek samples of evidence that best match that view, rather than seeking samples that contain all of the variation in the category. A series of experiments confirmed this hypothesis: when asked to generalize information obtained from small samples of evidence, young children generalized from more extreme examples that best matched what they thought category members should be like (even if such ideal exemplars were rare), with a shift across age towards generalizing from more average exemplars. Subsequent experiments confirmed the developmental and cognitive mechanisms underlying these age-related changes via a combination of experiments with children and computational modeling. The final set of experiments used this new information about child development to design a new educational approach that used a specific language-based strategy to help children reason more efficiently about biological categories. This project addressed its Intellectual Merit aims by answering key theoretical questions in cognitive and developmental psychology about the mechanisms that underlie knowledge acquisition and category-based induction. Further, this project met its Broader Impacts goals by providing training in developmental science, broadly disseminating the findings to communities of interest, and developing a new approach to exhibit design that uses the knowledge gained from this program of research to create more effective environments for informal science learning.

Last Modified: 11/13/2020
Modified by: Marjorie Rhodes