Abstract
The present study compared the use of two methods to compare multiple ideas, in children's mathematical problem solving. Fifth grade children (N＝46) were assigned to one of two method conditions and shown four ideas based on the mean, mode, maximum, and minimum in an intervention problem. Children assigned to Condition 1 found similarities and differences between the four ideas, and children assigned Condition 2 selected the best idea from the four ideas. After the intervention, all the children solved two problems: (A) examining whether they could estimate a number in ways other than the mean when the data contained an outlier, and (B) examining whether they could refer to a multiple representing value without a prompt. Only among children who gave a correct answer to problem A did more children assigned to condition 1 refer to multiple representing values. The analysis of answers when children referred to multiple representative values in problem B showed that there was a marginally significant tendency for children assigned to Condition 1 to respond that the mean is not always the appropriate value, or that the mean is affected by the maximum/minimum.
