Construction of Geometric Analogy Problems by Young Children in a Computer-Based Test

Abstract

The two purposes of this study were a) to explore how young children structure and solve geometric analogy problems in a computer-based test, and b) to compare this performance to the same item stems in a paper-and-pencil form. Twenty-four kindergarten and twenty-four second-grade children were required to construct solutions to analogy problems that were presented on a microcomputer. Half of the sample was academically gifted. These kindergartners were also tested with a paper-and-pencil version of the test to examine for possible differences between the two testing methods. Second graders were only tested with the computer version. Statistical analysis for the computer-based test focused on how these young children constructed responses to geometric analogy problems, and the error patterns committed by each age group. Analyses showed a strong linear decrease in performance as the item information load increased, but this effect was more pronounced for the kindergartners than for the second graders. Analyses also provided some evidence that changing the items from a paper-and-pencil response selection format to a constructed response format alters which component of the item carries the greatest weight in determining overall item performance. This study also investigated the crossover effects of computer-based testing on paper-and-pencil testing and vice versa in the kindergarten subjects. It was found that there was a positive effect on the post-test computer-based performance when pretested with the paper-form, but a negative effect on paper-form post-test performance when pretested with the computer form.