The proposed research investigates the hypothesis that socio-pedagogical context affects learning. Previous research show that both children and adults draw different inferences from the same data depending on the socio-pedagogical context---whether the informant is knowledgeable, helpful, and intends to teach. Rational analysis of pedagogical learning explains these differences and generates a number of predictions about how inferences should differ in novel pedagogical and non-pedagogical situations.
The proposed work specifically investigates predictions about how pedagogically-sampled (or non-pedagogically-sampled) examples affect trust in the teacher and affect subsequent learning, and how these social-pedagogical inferences can be leveraged to facilitate learning in richly-structured STEM domains.
University of Louisville
The project is using a cross-sectional research design and will generate evidence that is causal [experimental, computational modeling] and synthetic [computational modeling]. Lab experiments are being used to collect data on young children and college undergraduates. The study is comparing a pedagogical sampling of data with a non-pedagogical sampling of data. Analysis plans include standard statistical methods such as binomial tests, chi-squared, and t-tests, as well as computational modeling.
Findings will be posted as they become available.
All publications resulting from this grant will be posted on the University of Louisville Computational Cognitive Science lab website at http://louisville.edu/psychology/shafto.