Simulation Study of Item Validity Testing and Item Discrimination Index

Abstract

In psychology or educational research, we often want to test the validity of our instrument. Usually we use the Pearson coefficient of correlation in validity testing. However, the distribution of the data are not always normal distribution so it is also possible to use the Spearman and Kendall coefficient of correlation to test the validity of items. Valid items tend to have a positive and significant coefficient of correlation. It means the item is likely to be answered correctly by people who have a high total score and tend to be answered incorrectly by people who have a low total score. Item discrimination index is an index that measures the ability of item to be able to discriminate students who have high learning outcomes and students who have low learning outcomes. Simulation studies provide a clearer description and show that the item discrimination index has a significant relationship to validity testing that is commonly used in research.