Skip to content
Omnikalc

Item analysis calculator: how hard is a test question, and does it separate strong from weak students?

Enter the correct answers in the upper and lower groups to get the difficulty index, the discrimination index and the ideal difficulty for the answer format.

Updated on
Your inputs

Value out of range

Value out of range

Value out of range

Value out of range

Results

Discrimination index D

0.519

Excellent discrimination
Difficulty index (proportion correct)
0.630
Chance level of guessing
0.250
Ideal difficulty for this format
0.625

Item analysis checks how well each question on a test works. The item difficulty and discrimination calculator takes the size of the upper and lower groups, the correct answers in each and the number of answer options, and gives the difficulty index, the discrimination index, the chance level and the ideal difficulty.

How the calculation works

  • Rank the test takers by total score and take the upper group and the lower group, often the top and bottom 27 %, with the same size n.
  • The difficulty index is the proportion who answered the item correctly across both groups: p = (correct upper + correct lower) ÷ (2 × n). A high p means an easy item.
  • The discrimination index is the gap between the groups: D = (correct upper − correct lower) ÷ n. It runs from −1 to +1.
  • The ideal difficulty for a multiple-choice item is halfway between chance, which is 1 ÷ options, and 1: (1 + 1 ÷ options) ÷ 2.

Example

An item is given to the top 27 and the bottom 27 test takers, with 4 options. 24 of the top group and 10 of the bottom group answer correctly. The difficulty is 34 ÷ 54 = 0.630, close to the ideal of 0.625. The discrimination is 14 ÷ 27 = 0.519, which is excellent. Chance level is 0.25.

Why it matters

A test is only as good as its questions. Item analysis finds the items that are too easy, too hard, miskeyed or unable to tell strong students from weak ones, so they can be fixed before the next use.

Good to know

The indices come from small groups and are noisy, so use them as flags and not as verdicts. They also assume a test that measures one main skill, and an item in a mixed test can look weak while measuring something that matters. Review flagged items before deleting them.

The Cronbach’s alpha calculator measures the consistency of the whole test, and the percentile calculator ranks a single score within a group.

Frequently asked questions

What does the discrimination index tell me?

It tells you whether students who do well on the whole test also do well on this item. A positive value means the item separates strong from weak performers. A value near zero means it does not help, and a negative value means weak students do better than strong ones, which usually points to an error in the key or a misleading question.

What is a good difficulty?

Items answered correctly by about half to three quarters of the test takers tend to work best, because they separate people the most. Very easy items, above about 0.9, and very hard items, below about 0.2, give little information, although a few of each can be useful for fairness and motivation.

Why use the top and bottom 27 percent?

Kelley showed in 1939 that using the extreme 27 percent of a normal distribution gives the most stable estimate of discrimination, because the groups are far enough apart to differ and large enough to be reliable. With small classes you can use halves instead.

What should I do with a poor item?

Check the answer key first, then the wording, and then the distractors. An item that the strong group misses may have two defensible answers. An item that nobody misses may be too easy, and a distractor that nobody picks does no work and can be replaced.

Was this calculator helpful?

Thanks for your feedback!

Results are estimates provided for information purposes. They do not replace professional advice.