Question Bank Template and Blueprint Check
Four documents and four sheets that work out how many items of your blueprint the bank can actually fill, and which cells it cannot.
Free download · No account needed
Blueprint Coverage
What the paper needs, against what the bank can give
One row per outcome and cognitive band. Demand comes from the blueprint. Available is what survives all four tests.
| Outcome | Band | Demand | In bank | Available | Short | Action |
|---|---|---|---|---|---|---|
| CO-1 | Recall | — | — | — | — | — |
| CO-1 | Apply | — | — | — | — | — |
| CO-2 | Recall | — | — | — | — | — |
| CO-2 | Apply | — | — | — | — | — |
| CO-3 | Recall | — | — | — | — | — |
| CO-3 | Apply | — | — | — | — | — |
| CO-4 | Recall | — | — | — | — | — |
| CO-4 | Apply | — | — | — | — | — |
| CO-5 | Recall | — | — | — | — | — |
| CO-5 | Apply | — | — | — | — | — |
Totals
| Row | What it counts | Value |
|---|---|---|
| T-1 | Unique items in the bank | — |
| T-2 | Available after the four tests | — |
| T-3 | Blueprint length | — |
| T-4 | Items that must be written | — |
Available is the column an item bank has no room for, and it is the one that says whether the next paper can be built.
Every guide to organising a question bank describes the same repository: structured metadata, tags for blueprint domain and cognitive level, version history, a review workflow, exposure monitoring, and a lifecycle from draft to retired. All of it is correct and none of it answers the question you have three weeks before a sitting, which is whether the bank you hold can build the paper you need. This pack works that out, per blueprint cell.
NURS 218 Pharmacology at Kelbridge College: 412 items from prior exam files, a publisher bank and a QTI export, with four sittings of history and a 50-item midterm to build. 37 duplicate stems come out, leaving 375. Four tests then stand between an item existing and an item going on a paper, and 32 are retired on performance, 6 have drifted, 78 fail the writing standard and 24 are resting. That leaves 235 available.
Then the join. 42 of the 50 items can be found and 8 have to be written, in four cells, none of them recall. The bank is 62% recall because the publisher import is 81% recall, while the blueprint asks for 16%. At analyse or above the bank holds 42 items and only 17 are available, because those are the ones reused every term. The item analysis report for a single sitting cannot see any of this, and work marked against a rubric is counted somewhere else entirely.
What's in the pack
Blueprint Coverage
One row per outcome and cognitive band, including the cells with no demand: what the blueprint asks for, what the bank holds, what is available, and the shortfall.
Item Register
One row per item with the outcome its stem serves next to the outcome the file claimed, the band, the writing defects, the status and the action.
Exposure Tracking
Which items appeared on which sittings, how many times each has been used, and the sitting it becomes eligible again under the exposure rule.
Performance History
One row per item per sitting, with difficulty and discrimination and how far each has moved since the item's first use.
Item Writing Standard
Six defects a second reader can point at, each with the fix, and the published prevalence figures that put your own rate in context.
Bank Structure Note
The tag vocabulary, the identifier convention, and the retirement, drift and exposure rules written down so the next instructor inherits them.
Bank Status Note
The findings for one bank: what is available, which cells are short, which items to write first and why the shortage is about shape rather than size.
How the Bank Is Checked
The two joins and the four tests written out, with the arithmetic and the three things the method deliberately refuses to settle.
How to use it
- 1
Open it in River, or download it
Open the pack in River and let the agent read your bank and run both joins, or download the eight blank Word and CSV files with no account.
- 2
Send every source at once
Old papers, the publisher bank, a platform export, any item analysis reports. Duplicates across sources are expected and come out on the first pass.
- 3
Give it the outcomes and the blueprint
How many items the next paper takes from each outcome at each band. Never inferred from the last paper, because the bank already reflects that.
- 4
Read the shortfall by cell
Not the total. Eight items in four cells is a different afternoon from eight items spread over eight, and the emptiest cell is where to start.
Frequently asked questions
Is this template free?
Yes. Download all eight files as Word documents and CSV sheets with no signup and no card. Edit with AI is the optional path, where the agent reads your own bank, places every item and runs the blueprint join on your own blueprint. The rest of the packs sit in the template library.
What format are the downloaded files?
Word documents (.docx) for the writing standard, the structure note, the status note and the method explainer, and CSV (.csv) for the four sheets, zipped together. They open in Word, Pages, Google Docs, Excel, Numbers and Sheets. In River the same content opens as live Docs and Sheets.
How is this different from an item banking platform?
A platform stores items and tracks the fields you fill in. This runs the two calculations a platform leaves to you: what the item is against what its tag claims, and blueprint demand against the items that survive four tests. Both run on stated course outcomes, which come from a pack of their own.
Do I need item analysis reports for this to work?
No, and it will tell you exactly what changes without them. Without performance data every item is unproven rather than acceptable, so retirement and drift cannot run, and the join reports availability from the writing standard and exposure alone. Reading a single sitting's report is a separate job.
Why would a 375-item bank come up short on a 50-item paper?
Shape rather than size. A publisher import is mostly recall, so the bank at Kelbridge holds 232 recall items and 42 at analyse or above against a blueprint wanting 8 and 17. Buying another 200 recall items makes the surplus worse and the shortfall identical.
How do you know an item has been over-exposed?
By two movements together across three or more sittings: difficulty rising by 0.20 or more while discrimination falls by 0.15 or more. Neither alone is enough, since an item taught better also gets easier. Testing programmes treat that pattern as item parameter drift and flag it for review.
What counts as a badly written item?
Six things a second reader can point at, from an all-of-the-above option to a key that is the longest by a clear margin, all drawn from NBME's item-writing guide. On real examinations the rates run high: 47% of items on ten nursing papers carried a flaw, so 42% is normal rather than negligent.
Find out how many items of your next paper you actually have
Download the blank pack as Word and CSV files, or open it in River and have every item placed and the blueprint join run on your own bank.
Edit with AI