Abstract: Positive diversity effects among adults have been reported in a number of studies using a variety of methods (e.g., Heit, 2000; Kim, Keil, 2003; Osherson, Smith, Wilkie et al., 1990; Lopez, 1995; Sloman, 1993). However, there has been considerable debate about whether children also prefer diverse samples as the basis for induction. Several initial developmental studies, in which children were asked to make inductive inferences about animal categories based on a limited set of examples, suggested that children before 9 years old do not consider the degree of sample diversity when determining whether a sample provides a good basis for generalization (Lopez et al., 1992; Gutheil & Gelman, 1997; Rhodes, Gelman, & Brickman, 2008). In contrast to these null findings, some studies argue that children as young as 5 years old and even younger (4 years old) value sample diversity under certain circumstances (Heit, Hahn, 2001; Lo, Sides, Rozelle, 2002; Shipley, Shepperson, 2006; Wu X, Li H, 2008). Although multiple-exemplar samples were presented in the previous studies, some differences in the methods and materials should be realized, such as property category and conceptual category. In the present study, we further examine diversity-based reasoning in children as well as directly investigating the possible effects of property category and conceptual category on diversity-based reasoning.
5-6 and 8-9 years old children were recruited and each was asked to perform on sixteen tasks in order to analyze the effect of different property categories and conceptual categories on diversity-based reasoning. The experiment used a 2(property category: hidden properties vs. visible properties) ×2 (conceptual category: biological category vs. abiological category) within-subject design.
Across four types of tasks, we documented a developmental progression in children’s understanding of diverse samples as a strong basis for generalizations, showing that young children (5~6 years old) consistently failed to consider sample diversity in their inductive reasoning while older children (8~9 years old) preferred to create diverse samples based on which they made inferences about abiological category and hidden properties. What’s more, the inductive strength on diversity-based reasoning is remarkablely affected by property category and conceptual category: in conceptual category, the inductive strength on abiological category is higher than biological category; in property category, the inductive strength on hidden properties is higher than visible properties.
These results suggest that the capability of recognizing the value of a diverse sample for inductive reasoning emerges at a relatively late age and is enhanced slowly across young children. However, why the youngest children in our studies performed more poorly than those in other studies is not known yet, and more researches about the impact factors of diversity-based reasoning are required.

Key words: inductive reasoning, diversity effect, conceptual category, property category