Item Analysis: Difficulty and Discrimination Index

Item analysis tells you which test questions worked: how hard each one was (difficulty index) and whether it separated learners who know the lesson from those who don't (discrimination index).

Step 1: Form the upper and lower groups

Rank the learners from highest to lowest total score. Take the top third as the upper group and the bottom third as the lower group. Each group has n learners.

n = number of learners ÷ 3 (rounded up)

Step 2: Count correct answers per item

For each item, count how many in the upper group got it right (U) and how many in the lower group got it right (L).

Step 3: Compute the two indexes

Difficulty index    P = (U + L) ÷ (2 × n)
Discrimination index D = (U − L) ÷ n

Step 4: Interpret them

PDifficulty
0.00 – 0.14Very difficult
0.15 – 0.39Difficult
0.40 – 0.70Moderate
0.71 – 0.85Easy
0.86 – 1.00Very easy
DDiscrimination
0.19 and belowPoor, reject
0.20 – 0.29Marginal, needs improvement
0.30 – 0.39Reasonably good, may improve
0.40 and aboveVery good

Combine the two to decide. A moderate, very good item is kept as it is. An item with poor or marginal discrimination is discarded when it is also easy or difficult, and revised (mostly its choices) when it is moderate. Items with reasonably good or very good discrimination are kept, revised as needed.

Worked example

45 learners took the test, so each group has n = 45 ÷ 3 = 15 learners. On item 7, 12 in the upper group and 6 in the lower group got it right.

  1. P = (12 + 6) ÷ (2 × 15) = 18 ÷ 30 = 0.60 → moderate
  2. D = (12 − 6) ÷ 15 = 6 ÷ 15 = 0.40 → very good
  3. Decision: accept the item as it is. It's a good item to reuse.

Tips

Let ClassToolsPH compute this for you

ClassToolsPH runs the full item analysis (P, D, interpretation and decision for every item) right after the exam, and lets you save your best items to a reusable item bank.

More guides for teachers