A Brief Discussion of Test Question Performance Analysis (P and R values)
P and R Values in Test Questions That Have Been Used Previously On An Exam
Indicates the difficulty of the question. The number demonstrates the percentage of candidates who answered the item correctly (e.g., a p-value of .60 means that 60% of candidates got the question right). An ideal p-value falls in the range of .30-.80.
Indicates the discrimination value of the question. It measures the correlation between high scoring candidates getting the item correct and low scoring candidates getting the item wrong. If the value is negative, it means that lower scoring candidates did better on the item than high scoring candidates, and is a red flag. If the value is 0 or low positive, it suggests that the item did little to discriminate and therefore has little value. The ideal r-value is around 0.3.