After a test, most teachers look at the grade distribution. A few look at which questions the class struggled with. Almost none look systematically at which questions worked.
That's item analysis — and it's one of the highest-leverage analytical habits a teacher can develop.
What item analysis actually tells you
For each question on a test, you can ask:
- Difficulty — what percentage of students got it right?
- Discrimination — did high-scoring students get it right more often than low-scoring students?
- Distractor performance — for multiple choice, which wrong answers attracted students?
These three questions expose a lot about both the question and the students.
Difficulty: the simplest metric
If 95% of your students got a question right, the question didn't distinguish between students who learned the material and those who didn't. Every student succeeded. That's fine for a confidence-builder at the start of a test, but it's not doing assessment work.
If 10% of your students got it right, the question is either:
- Too hard (poorly written, ambiguous, out of scope)
- Testing something you didn't teach effectively
- Genuinely the hardest concept in the unit
Sweet spot for most items: 50–85% correct. That's the range where a question is informative — meaningfully separating students who understood from those who didn't.
Discrimination: the diagnostic metric
Discrimination asks: among students who got this question right, did they also do well on the rest of the test?
A high-discrimination question is one where top-scoring students reliably got it right and low-scoring students reliably got it wrong. This is what you want. The question distinguishes mastery.
A low-discrimination question (or worse, a negatively discriminating one) is a warning. If top students got it wrong and low students got it right, something is off:
- The question might be poorly worded
- The "correct" answer might actually be wrong
- There might be an unintended interpretation
The negative discrimination red flag
Any question where low-scoring students outperformed top students on that specific item deserves a second look. It almost always means the question is broken. Check the wording, the answer key, or whether the question tests something different from what you intended.
Distractor analysis: what wrong answers reveal
For multiple choice questions, look at which wrong answers students picked. If one distractor attracts a lot of students, ask why:
- Is it a common misconception you need to address?
- Is it a plausible-but-wrong reading of the question?
- Is there a lurking ambiguity that makes two options defensible?
Distractor patterns are gold for diagnosing student thinking. If 30% of your class picked the same wrong answer, you've discovered something about how they're reasoning — and what to reteach.
Where item analysis finds the most unexpected insights
1. Your best-loved question is underperforming. Teachers often have favorite questions they reuse year after year. Item analysis sometimes reveals these questions are too easy, too hard, or ambiguously worded. The data reveals what gut sense missed.
2. A unit you thought went well actually didn't. The overall grade distribution looks good, but item analysis reveals that the average is buoyed by easy items while the items actually testing the unit's key standards underperform. The class score hides the gap.
3. A lesson you spent a lot of time on didn't stick. You taught something thoroughly; the item tests it directly; 30% of students got it right. That's a signal to re-examine the lesson, not just the test question.
4. Your answer key is wrong. It happens. Item analysis surfaces it fast. If a distractor is attracting 70% of students, either your students are systematically misreading it — or your key is wrong.
Making it a habit
You don't need to do item analysis on every quiz. A reasonable cadence:
- After every unit test: review the 3–5 lowest-discrimination questions
- End of quarter: look for patterns across unit tests (always weak on this standard?)
- Before reusing a test: any question with poor performance gets reviewed or revised
This is a 10-minute habit once you're practiced. The payoff compounds: better tests, better-informed teaching, fewer "I thought this was clear" moments.
In PaperScorer
Item analysis reports are generated automatically after every test. You see:
- Per-item correct percentage
- Discrimination index for each item
- Distractor selection percentages for multiple choice
- Per-standard mastery if you've tagged items to standards
You don't have to build the analysis — you just have to look at it.
Key takeaway
Your test is a tool. Item analysis tells you whether the tool is working. Three questions per review — difficulty, discrimination, distractors — and your assessments get sharper every time you run them.
