Comparison of the Selection of Items Using Classical Test Theory and Item Response Theory Based on Sample Sizes

The study examined the comparability of item statistics of classical test theory and item response theory of the selected items for the several sample sizes (from N=200 to N=1000). A 60-item multiple choice Mathematics test which was developed by Edo State Ministry of Basic Education Board, Nigeria was used to gather data from the various randomly selected sample. Results showed that more items were selected from the 60-item Basic Education Mathematics Examination for the various sample sizes when item response theory-based item statistics estimates were used than when classical test theory-based item statistics estimates were used, it was also found that the item discrimination indices of classical test theory and item response theory were comparable for different sample sizes while the item difficulty indices of classical test theory and item response theory were not comparable only on the very small sample of 200.

Keywords: Classical Test Theory, Item Response Theory, Sample Sizes, Item Difficulty, Item Discrimination.