Suppose we pilot a first draft of the test on a sample of 300. From the data we will calculate Discrimination value, CVI etc.. and we finalise the items
After that should we have to take data for reliability and validity ? Or we can use the same data?