The purpose of this study was to investigate the development of the concept of family in young children. Comparisons were made relative to cognitive developmental level (conservers vs. nonconservers), race (black children and white children), and gender (males and females). Subjects were 32 kindergarten and 36 first grade students enrolled in two university laboratory schools. Thirty-four children were black and 34 were white; 34 male and 34 female. Two instruments were used A conservation of number task was administered to assess (...) cognitive developmental level. A three-part assessment task was devised to obtain information on the subjects’ concepts of family. Subjects were first asked to draw their family. They were then asked to construct a family selecting members from a multiethnic group of paperdoll people. Finally, the concept of family was assessed using a family configuration task. Twelve different groups of paperdoll people were presented. Subjects were asked if each group was a family and then asked to give a rationale for their response. The results confirmed a moderate positive relationship between number of different configurations selected as families and cognitive developmental level for females. Some other developmental, gender and ethnic group differences were identified in choice of configurations of paperdoll people as families or not families and in selection of rationales to support these choices. The majority of the children drew an accurate picture of their families but when given the group of paper people, most constructed families that had more members. The results hold implications for our current Social Studies instructional practices and leave us with a number of questions for further research. (shrink)