The preschool years are characterized by extraordinary changes in children's causal knowledge; 3- and 4-year-olds have increasingly coherent theories about the way things work, they can reason about hidden causal forces like gravity or germs, and they can rapidly learn causal relations from new evidence. These changes are accompanied by equally remarkable developments in social knowledge -- the understanding of the motivations, emotions, and mental states that drive human behavior. Intuitively, it seems clear that what children know about people can have consequences for what children learn from people. Indeed, studies have shown that, where social information is critical (such as in learning about language or culture), children's social knowledge influences how and what they can learn. However, causal learning doesn't necessarily require social information; children can learn about causal relations from associations between events, or by simply observing the consequences of actions. Thus, children's developing social cognition may have unique and importantly different implications for causal learning than for social learning. To date, these implications have been largely unexplored. This project investigates the influence of preschoolers' developing social knowledge on their causal learning. Specifically, it focuses on their developing understanding of epistemic states -- that some people know more than others, and that people may have different areas of expertise and ability. The first four studies use lab-based experimental methods to address (1) whether, as in other learning domains, preschoolers are sensitive to others' past accuracy in causal learning; (2) whether, in contrast to other learning domains, preschoolers show selective trust of others depending on the type of knowledge and expertise they have; (3) how epistemic information (i.e., that one person knows more than another) interacts with statistical information (i.e., a strong vs. weak correlation) in causal learning; and (4) how epistemic information interacts with information from children's own active causal exploration. All four studies include measures to investigate how developmental and individual differences in social cognition contribute to the process of causal learning. Finally, a fifth classroom-based study combines experimental with observational methods to investigate how preschoolers actively engage with social sources in the process of exploratory causal play.
There is a trend in early childhood education toward increasing direct instruction, thus more emphasis on social transmission in early learning. Research is needed to better enable us to evaluate this trend. To this aim, the results of this research will have implications for both our understanding of basic learning processes in children and our ability to design effective educational environments for them. The results will also deepen our understanding of the unique processes that characterize early casual learning (the foundation for later science education) and distinguish it from learning in other domains. The final classroom-based study explicitly creates a link between basic research and early childhood education, and is expected to lead to further research targeting this link.
Project Description: The preschool years are characterized by extraordinary changes in children’s causal knowledge; 3- and 4-year-olds have increasingly coherent theories about the way things work, they can reason about hidden causal forces like gravity or germs, and they can rapidly learn causal relations from new evidence. These changes are accompanied by equally remarkable developments in social knowledge – the understanding of the motivations, emotions, and mental states that drive human behavior. This project was designed to explore some of the implications of social cognition on causal learning. Project Outcomes: The original proposal focused on how preschoolers understanding of epistemic states – that some people know more than others, and that people may have different areas of expertise and ability – impacts their causal learning from others. Attempts to measure children’s epistemic state understanding lead to additional unexpected findings, which were followed up in new studies. Finally, additional studies broadened the scope of the initial proposal to investigate how children’s learning is influences by others’ intentions to teach causal information. Key findings from all of the above are summarized below: 1) We found that 3- and 4-year-old children take past causal ability as a sign of causal knowledge and expertise. This parallels the way in which children monitor the accuracy of speakers in learning new words. 2) We found that, when given the choice between an informant with known casual abilities and one who labels objects correctly, preschoolers direct requests for new information selectively and appropriately. They ask the causal expert about causal properties and mechanisms, and ask the accurate labeler about novel words. Thus, preschoolers understand that different people know different things, and this influences from whom they learn. 3) We found an important asymmetry in the way children integrate social and physical knowledge with new evidence. Preschoolers appear to have a "social bias" about new evidence; they tend to view causal success as a sign of causal expertise, even when they are explicitly told that no such expertise exists. 4) We developed a measure to compare how children from 36-48 months monitor and infer their own knowledge and how they monitor and infer the knowledge of others. We found meaningful connections between young children’s memory for events and their ability to monitor and infer their own knowledge. The same was not true for how they monitor and remember the knowledge of others – these were found to be unrelated to memory for events. 5) We found that, though children prefer to learn from people who have a proven track record of being knowledgeable, they will also learn from people who are sometimes ignorant. 6) We found that the intention to teach causal information can sometimes be conveyed non-verbally, through different ways of performing actions. This influences whether children think the information is relevant to causal learning, relevant to social learning, or both. 7) In a theoretical review of this work and the broader literature on children’s learning from testimony, we characterize children’s learning from others as a rational process – one in which children evaluate new information in light of their existing knowledge. This framework emphasizes that children are active learners, and has important implications for future research and also for early childhood education. Intellectual merit: The proposed activities have advanced our understanding of role of social influences in young children’s learning. As such, the results of this research have implications for both our understanding of basic learning processes in children and our ability to design effective educational environments for them. Broader Impacts: This research helps us understand the process by which social understanding impacts children’s causal learning, and better enables us to evaluate this trend. This research has contributed to the training of several graduate students, lab assistants, and numerous undergraduate students who have gained hands-on laboratory experience from participating in lab activities. This includes several honors and independent study students who are contributing authors on publications resulting from this proposal. NSF support has also resulted in the development of research-intensive, writing-intensive courses for graduate and undergraduate students. The courses teach critical thinking about theories and methods in cognitive development through reading and discussion of original research articles and extensive writing assignments. The studies have been disseminated to the research community through publication top journals, presentation at scientific meetings, and invited talks at conferences and research institutions. Findings are regularly disseminated through lectures to preschool teachers, parents, and other educators and through a public website and annual newsletter which is distributed to participating parents and schools. Major research findings are also written up in Cornell’s Outreach and Extension news, which is distributed to traditionally under-served and under-resourced preschools and parenting education programs.
