What exactly is the p-value of a question? Converted into a percentage (* 100), the p-value indicates the percentage of candidates that answered the item correctly. If you have a high p-value, the question is relatively easy and almost every candidate got it right. If you have a low p-value, the question has been too difficult. In that case the proportion of candidates who got the question right is too low.
If you want to make a good test, you have to make sure there is a good balance between the question types. A test should not be too difficult, but also not too easy. A good p-value varies according to the number of answer possibilities. You can use the table below as a guideline:
Optimal p-values for different types of questions
Suppose you have a question in your test that is answered by 100 students. You perform a statistical analysis by calculating the number of correct answers. Suppose 60% of the candidates get the question right, then there is a p-value of 0.60.
Would you like to know more about the p-value or other statistics related to tests? Our colleagues will be happy to help you. Would you like to know more about digital assessment? Then check out our knowledge center. You can follow us on LinkedIn to stay up to date with developments in digital testing.