- Practice
- About 14 min
- 5 practice questions
Item analysis and test statistics: facility value and simple calculations
Calculate and interpret the facility value and discrimination index of a test item, plan a test with a blueprint, and separate the ideas of validity and reliability.
If something looks wrong, please tell us.
A test is only as good as its items. After a class test, a careful teacher asks of each item: was it too easy or too hard, and did it separate pupils who know from pupils who do not? Competitive exams ask for these ideas in calculation form. This lesson gives the formulae, how to read the results and how a test blueprint is planned.
Facility value (difficulty index)
The facility value of an item is the fraction of pupils who answered it correctly: facility value = number answering correctly divided by number attempting. It runs from 0 to 1, or from 0 to 100 percent. A high value means an easy item; a low value means a hard one. Many textbooks prefer a spread of items, and prefer to avoid items that are answered correctly by nearly everybody or almost nobody, since they tell the teacher little.
Discrimination index
A good item should be answered correctly more often by high scorers than by low scorers. To measure this, rank pupils by total score and take equal-sized upper and lower groups, each of n pupils. The discrimination index is D = (U - L) / n, where U and L are the numbers answering the item correctly in the upper and lower groups. D lies between -1 and +1. A clearly positive D means the item separates well. A D near zero means it does not separate. A negative D means low scorers did better than high scorers, so the key may be wrong or the item misleading, and it should be checked.
Blueprint (table of specification)
A blueprint plans the test before it is written, setting how many marks go to each topic and each type of objective, so that the test covers the syllabus in the intended proportion. This improves validity, which is the extent to which a test measures what it is meant to measure. Reliability is the consistency of its results, such as giving similar scores if repeated under similar conditions. A test can be reliable without being valid, but a test cannot be valid if it is not reliable.
Worked example 1: facility value and discrimination
Item A was attempted by 40 pupils and 28 answered correctly. Facility value: 28 / 40 = 0.70, or 70 percent, an item of medium ease. Now take the top 10 and bottom 10 pupils by total score. In the upper group 9 got item A right and in the lower group 3 did. Discrimination index: D = (9 - 3) / 10 = 0.6, a clearly positive value, so the item separates well. Item B: upper group 4 correct, lower group 7 correct. D = (4 - 7) / 10 = -0.3. This is negative, so the teacher checks whether the answer key is wrong or the question misleads strong pupils.
Worked example 2: planning a blueprint
A 40-mark test is planned with these weights: number 30 percent, geometry 25 percent, algebra 20 percent, mensuration 15 percent and data 10 percent. Check the weights: 30 + 25 + 20 + 15 + 10 = 100 percent. Marks: number 30 percent of 40 = 12; geometry 25 percent of 40 = 10; algebra 20 percent of 40 = 8; mensuration 15 percent of 40 = 6; data 10 percent of 40 = 4. Check: 12 + 10 + 8 + 6 + 4 = 40. The weights are illustrative; a teacher sets them from the time given to each topic in the syllabus.
| Statistic | Formula | Reading the result |
|---|---|---|
| Facility value | Number correct / number attempting | Near 1: easy; near 0: hard |
| Discrimination index | (U - L) / n | Positive: separates; zero: does not; negative: check the item |
| Class mean | Sum of scores / number of pupils | Average performance of the class |
Facility value tells how easy an item was; the discrimination index tells whether it separates strong from weak pupils. A negative D is a warning.
Key takeaways
- Facility value is the proportion correct; discrimination is (U - L) / n.
- A negative discrimination index means the item should be checked.
- A blueprint spreads marks across topics and objectives before the test is written.
- Validity asks whether the test measures what it should; reliability asks whether results are consistent.
Check your understanding
-
1. In a class of 50 pupils, 35 answer an item correctly. What is its facility value?
- 0.70
- 0.35
- 0.50
- 1.43
Show answer
Answer: A. Facility value = 35 / 50 = 0.70.
-
2. In groups of 15, the upper group has 12 correct and the lower group 6 correct on an item. What is D?
- 0.6
- 0.8
- 0.4
- 0.2
Show answer
Answer: C. D = (12 - 6) / 15 = 6 / 15 = 0.4.
-
3. An item has D = -0.2. What should the teacher do first?
- Keep the item because it is difficult
- Check the answer key and wording
- Give extra marks to the lower group
- Make the item much longer and harder
Show answer
Answer: B. A negative D means weaker pupils did better than stronger ones, which suggests a faulty key or a misleading item.
-
4. A test measures reading speed rather than the mathematics it was meant to test. Which quality is weak?
- Reliability of the scores
- Objectivity of marking
- Length of the test
- Validity
Show answer
Answer: D. A test is valid if it measures what it intends to measure.
-
5. A 50-mark test gives 20 percent of marks to algebra. How many marks are for algebra?
- 5
- 10
- 20
- 15
Show answer
Answer: B. 20 percent of 50 = 0.2 x 50 = 10 marks.
Sign in to keep track of the lessons you finish.