Farzaneh Mozaffari; S. Mohammad Alavi; Abbasali Rezaee
Abstract
When constructing a test, an initial decision is choosing an appropriate item response format which can be classified as selected or constructed. In large-scale tests where time and ...
Read More
When constructing a test, an initial decision is choosing an appropriate item response format which can be classified as selected or constructed. In large-scale tests where time and finance are of concern, the use of response chosen known as multiple-choice items is quite widespread. This study aimed at investigating the impact of response format on the performance of structure tests. Concurrent common item equating design was used to compare multiple-choice items with their constructed response stem-equivalent in a test of grammar. The Rasch model was employed to compare item difficulties, fit statistics, ability estimates and reliabilities of the two tests. Two independent sample t-tests were also conducted to investigate whether the differences among the item difficulty estimates and ability estimates of the two tests were statistically significant. A statistically significant difference was observed in item difficulties. However, no significant difference was detected between the ability estimates, fit statistics, and reliabilities of the two tests.
