Developmental changes in the consideration of sample diversity in inductive reasoning

Determining whether a sample provides a good basis for broader generalizations is a basic challenge of inductive reasoning. Adults apply a diversity-based strategy to this challenge, expecting diverse samples to be a better basis for generalization than homogeneous samples. For example, adults expect that a property shared by two diverse mammals (e.g., a lion and a mouse) is more likely to be shared by all mammals than a property that is shared by two more similar mammals (e.g., a lion and a tiger). Across four studies, we document a developmental progression in children's understanding that diverse samples provide a strong basis for generalizations, such that young children (grade 1) consistently failed to consider sample diversity within their inductive reasoning, but older children (grade 5) preferred to create diverse samples on which to base inferences about basic-level categories. These results suggest that recognizing the value of a diverse sample for inductive reasoning emerges slowly across the elementary school years.