Children's use of sample size and variability to make social inferences

Document Type


Publication Date





Learning to generalize from instances is an important part of the development of inductive reasoning skills. This study examined developmental trends in elementary school children's use of sample size and variability to make inferences about commonly encountered biological and behavioral characteristics. Children in grades 1, 3, and 5, and college students made population estimates based on samples in which both sample size and variability were manipulated. Subjects of all ages used sample size and variability to make their population estimates. Even the youngest children gave significantly lower population estimates for low-variability than for high-variability samples, and for smaller (one or three) than for larger (30) samples. Although the use of sample size depended on variability, this relationship did not change by age group. Variability and domain, however, each independently interacted with age group. Most of the differences in population estimates appeared to be due to sample variability rather than to sample size.