The Development of Thinking: Inductive Inference in Young Children

Ohio State University, The, Columbus OH

Investigators

Abstract

Generalization from known to the unknown, or inductive inference, is an important component of human thinking. For example, if one learned that robins have hollow bones, would she think that eagles also have hollow bones? Is it reasonable to infer that bats have hollow bones too? Which attributes of robins and eagles are important for inferring that eagles do (or do not) have hollow bones? Do these attributes differ in their importance? And how do children learn which attributes are most predictive for inferring biological properties? Our prior research indicated that young children induce biological properties based on the overall similarity between compared entities, whereas preadolescents rely on a single most predictive source of information, such as biological kind or inheritance information. In this project, we will examine how young children combine different sources of information when inferring biological properties, and how they acquire the ability to rely on a single most predictive source of information. Specifically, we will estimate the importance of different sources of information across different age levels. We will also examine the transition from young children's induction (based primarily on overall similarity) to preadolescents' induction (based primarily on most predictive attributes). Finally, we will examine changes in inductive inference as a function of learning. In the course of the project, children and preadolescents (3-4-, 5-6-, 7-8- and 11-12 year olds) will participate in inductive inference and learning studies. The former will consist of a single session where participants will solve induction problems, whereas the latter will consist of 6 learning sessions spaced over a two-week period. At each session participants will be presented with induction problems and provided with feedback as to whether or not their inference is correct. The goal of these learning studies is to establish how children develop understanding of which attributes are most predictive for inferring biological properties. The project will contribute to our understanding of the development of human thinking, particularly of inductive inference and concept formation. In addition, the project may have important implications for the instructional theory and for learning and teaching science because it will specify how children think when they do not have knowledge of most predictive attributes, and how knowledge of these attributes changes their thinking.