Itembank — Quiz Item Analysis & Question Bank

Itembank — Quiz Item Analysis & Question Bank — a premium, 100% offline digital tool by Aceplore. Runs entirely in your browser: no accounts, no subscriptions, your data never leaves your device.

A private, offline question bank with classical item analysis — difficulty, discrimination, point-biserial and KR-20, every figure printed with the number of responses behind it and withheld outright when that number is too small to interpret — 31 tools in one, on your own machine.

31 built-in tools in one app. Question banks with a detail page, an item list, a statistics page per item, an item writer with a live candidate preview and pre-save checks, topics, cognitive levels against a target mix you set, and a topic-by-level blueprint that names every thin cell; papers with a full analysis page, an assembly screen whose blueprint fills in as you pick, a response grid you can paste into, and candidate scores with the group each one fell into; the item analysis itself — difficulty p as correct over attempts, discrimination D as p(upper) minus p(lower) at an editable group fraction, point-biserial with M₁, M₀, p and the population SD all printed, distractor counts option by option, dead distractors, a key check against the upper group's most-chosen answer, KR-20 with k, Σpq and the variance shown in full, and an item-flag list against thresholds you set; retirement that keeps the item and everything it recorded, and a revision log; plus topic performance, candidate performance, a side-by-side paper comparison, a written report, CSV export and a Method page whose worked examples run live.

Difficulty (p). 18 correct of 30 attempts is p = 0.60, and the division is printed next to the result. A higher p means an easier item. An item with no attempts has no p-value rather than a zero, and any item below your minimum response count is withheld rather than estimated.

Reliability (KR-20). k ÷ (k − 1) × (1 − Σpq ÷ variance). With k = 4, p values of 0.8, 0.6, 0.5 and 0.3 and a variance of 1.2, Σpq is 0.86 and KR-20 is 0.377778. A variance of zero is refused with a reason rather than divided by, and the page states what the number is not — not a pass rate, not a grade, not comparable across papers of different lengths.

Blueprint. A grid of active items with a minimum per cell you set, so the gaps in coverage are named rather than discovered by a reviewer. It is a coverage check against your plan and says nothing about whether the items are any good — that is what the statistics views are for.

What's included:

• Itembank.html
• Quick-Start-Guide.html
• How-To-Use-Guide.html
• README.txt
• LICENSE.txt

Includes a personalized welcome wizard, an interactive guided tour, a ⌘K command palette, light/dark themes with 8 accents, and one-click JSON backup & restore.

Itembank is a planning, scheduling and record-keeping tool for people who study, teach or sell what they know. Every figure it reports is computed by a stated formula from the times, counts, prices and thresholds YOU enter, and those thresholds are editable because they are conventions rather than laws. Where it implements a published scheme, it names the scheme and exposes its parameters: a review schedule orders your practice, it does not predict what you will remember, and a statistic computed from a small number of responses is reported with that number beside it because it cannot be read as fact. Nothing here is a guarantee of a grade, a pass, a completion, a sale or a job, and none of it is academic, financial or legal advice. You are solely responsible for your own study, teaching and trading decisions.