Eighty first-grade childen were pretested on a variety of conservation tasks. Subjects who were either nonconservers or intermediate conservers were randomly assigned to one of three conditions: social interaction, social observation, and individual control. Subjects in the social interaction condition worked collaboratively on conservation tasks with a same sex partner. Subjects in the social observation condition individually observed pairs of subjects working together and control subjects worked individually on conservation tasks. The purpose of the social observation condition was to control for the effects of task relevant information that was expressed during dyadic interactions. All subjects were individually post-tested on conservation tasks that were the same form but different content than the pretest items.
Subjects in the social interaction condition had significantly greater cogntive change scores (post-test less pretest) than subjects in the social observation and control conditions. There were no significant differences between change scores of subjects in the latter two conditions. Also, subjects in the social interaction condition gave significantly more novel explanations for conservation judgments than subjects in the social observation condition. These findings supported a socio-cognitive conflict model of cognitive development in young children.