Item-Analysis Calculator

Free · no signup · your data stays in your browser

Item-Analysis Calculator

Paste your exam's response matrix and get the full classical-test-theory item analysis — difficulty, discrimination, point-biserial, distractor efficiency, KR-20 — instantly. Everything is computed in your browser: exam and student data never upload anywhere.

Load your responses

Paste from Excel / Sheets (students in rows, items in columns)

Your data is parsed and analyzed locally — open your browser's network tab to verify nothing is uploaded.

Methodology

Every statistic follows the standard classical-test-theory definitions used in medical and higher education, and mirrors the implementation inside the StudyDrome assessment engine. Citations per metric:

Difficulty index (p)
Proportion of students answering correctly. Items at p ≥ 0.8 are flagged too easy, p ≤ 0.3 too hard. NBME Item-Writing Guide
Discrimination index
Correct-rate difference between the top and bottom 27% of students ranked by total score; the 27% split maximizes the contrast between groups. Kelley 1939, J. Educational Psychology
Point-biserial correlation
Correlation between getting the item right and the total score, computed with the population SD and clamped to [−1, 1]. Crocker & Algina, Introduction to Classical and Modern Test Theory
Non-functioning distractors & efficiency
A distractor chosen by fewer than 5% of examinees is non-functioning; efficiency is the share of distractors that function. Tarrant, Ware & Mohammed 2009, BMC Medical Education
KR-20 reliability & SEM
Internal-consistency reliability for dichotomous items; SEM = SD·√(1−r). High-stakes exams should aim for ≥ 0.85–0.90. Kuder & Richardson 1937, Psychometrika

These are the same formulas StudyDrome runs automatically on every exam.

Share this tool

Frequently asked questions

Does my exam data get uploaded anywhere?

No. Parsing and every statistic run in your browser. The only network calls this page makes are anonymous aggregate counters (which tool, which action) and — only if you use it — the export form, which sends your email address and nothing else. You can verify this in your browser's network tab.

What input formats work?

Paste straight from Excel or Google Sheets (tab-separated), or upload CSV/XLSX. Two layouts are auto-detected: scored (cells are 0/1, blank = 0) and raw responses (cells are option letters A–E, first data row is the answer key). Raw responses additionally unlock distractor analysis. Header rows and name/ID columns are detected automatically, with manual toggles if detection guesses wrong.

How many students do I need for reliable statistics?

The tool computes everything from 2 students up, but below 10 students it suppresses per-item flags and point-biserial values because they are too unstable to act on — the same threshold the StudyDrome engine applies. Discrimination indices become meaningful from roughly 30 students; reliability estimates stabilize with more items and more students.

What does KR-20 mean?

KR-20 estimates how consistently your exam measures — from 0 (noise) to 1 (perfectly consistent). Around 0.7 is generally acceptable for classroom exams; high-stakes decisions call for 0.85–0.90+. A low value usually means too few items or several flawed ones; fix the flagged items first.

Is this really free?

Yes. Analysis is free with no account; the XLSX/PDF export asks only for an email address. The tool exists so you can see the kind of item analytics StudyDrome produces automatically on every exam.