Education & TrainingFree
Adapting a Publisher Test Bank
Send the instructor manual, the slide decks and the test bank export, and get every item mapped to your own outcomes with the coverage gaps counted.
A publisher package arrives looking like a gift. An instructor manual, slide decks per chapter, an image library, cases with solutions, and a test bank with more items than you could use in a decade. What it is actually organised around is the book's table of contents, which is the author's argument about how the subject decomposes. Your course is organised around your outcomes. Those two structures are never the same, and the difference is invisible until you try to build an exam.
ACC 201 Financial Accounting is an invented example. The bank has 1,240 items across 16 chapters. The course teaches 11 of those chapters, so 842 items are in scope and 398 are not. Map the 842 to the course's six outcomes and the distribution is nothing like the course. Recording transactions carries 20 percent of the course weight and 42.28 percent of the items. Preparing statements carries 25 percent of the weight and 14.01 percent of the items. Analysing with ratios: 20 percent, and 11.16 percent.
That skew has a hard consequence. If you draw items in proportion to your own outcome weights, the scarcest outcome sets the ceiling for the whole bank. Explaining a preparer's ethical duties is 5 percent of the course and has 12 items, so the largest outcome-proportioned bank you can assemble is 240 items. Of the 842 in scope, 602 are surplus, which is 71.50 percent. A 1,240-item bank yields 240 usable items, or 19.35 percent, and none of that is visible from the chapter counts.
The second skew is cognitive level, and it is worse
Publisher banks are built to be gradeable at scale, which pushes them toward recognition. In ACC 201's 842 in-scope items, 561 are recall or recognition, 238 are apply or compute, and 43 are analyse or evaluate. That last group is 5.11 percent of the bank. Meanwhile the two outcomes whose verbs are analyse and evaluate carry 35 percent of the course. Scale that to the 240-item proportioned bank and you need 84 analyse-level items where 23 exist, so 61 have to be written.
This is why the mapping is the work and the item count is a distraction. A bank of 1,240 items with no cognitive level tags, no item statistics and no outcome alignment is an inventory, not an assessment resource. What makes it usable is a per-item record of which of your outcomes it serves, at what level, and whether it survived your review. Everything the publisher supplies about an item describes the item's place in their book.
Two constraints on what you do next. Formats first: the Question and Test Interoperability specification exists to define an information model that can be used to represent and exchange assessment content, and a bank that ships as Word files per chapter is not that. Licensing second. Teaching is among the purposes named in the fair use section, and the statute's fourth factor is the effect of the use upon the potential market for or value of the copyrighted work. Whether a given adaptation clears that is your library's call, not a technical question.
How it works
Send the package
The test bank export, the instructor manual, the slide decks and whatever else came with it.
Send your outcomes
Your outcomes with the weight each carries, which is what the mapping is built against.
Map every item
Outcome and cognitive level per item, with out-of-scope items separated rather than deleted.
Count the gaps
Yield per outcome, the binding constraint, and the items you will have to write yourself.
What you get
- Every item mapped to one of your outcomes, or marked as out of scope for your course
- Item counts per outcome against that outcome's weight, so over-supply and shortfall both appear
- A cognitive level tag per item, because publisher banks almost never carry one
- The largest outcome-proportioned bank the package can actually yield, with the binding outcome named
- The shortfall stated as items to write, per outcome and per level, rather than as a coverage percentage
- Slides, cases and images inventoried the same way, with an adopt, adapt or discard call on each
Common questions
Why map to outcomes instead of chapters?
Because the chapter structure is the publisher's decomposition of the subject and your outcomes are yours. On the worked example the two disagree badly: recording transactions gets 42.28 percent of the items for 20 percent of the course weight. A chapter-level coverage check would call that package complete.
How can a 1,240-item bank only yield 240 items?
Because the scarcest outcome sets the ceiling. Once you draw in proportion to your own weights, the outcome with 12 items and 5 percent of the course caps the whole bank at 240. The other 602 in-scope items are real items that simply do not fit the proportions you need.
Does it tag cognitive level even when the publisher has not?
Yes, and that tagging is most of the value. The tag is derived from what the item asks the student to do rather than from its position in the book. ACC 201's bank has 43 analyse-level items out of 842, which no chapter listing would have told you.
What about the slides and the instructor manual?
Inventoried the same way, with an adopt, adapt or discard call on each asset and the reason recorded. Slides are usually the most adaptable and the most trouble, because they encode the book's sequence. Course material accessibility review is the separate pass on whether they are usable by everyone.
Can we legally rewrite publisher items?
That depends on your licence and your institution's position, and the analysis flags it rather than answering it. Fair use has a market-effect factor and a licensed bank has terms. The mapping records which items you adapted and how far, which is the record your library will ask for.
How does this connect to building an actual exam?
This produces the mapped inventory. The exam and question bank pack is where a paper gets assembled from it against a blueprint, and item analysis review is what tells you afterwards which items discriminated. Adopted publisher items rarely have prior statistics, so the first run is the first evidence.
What if the export will not open in our LMS?
Common, and it is a format problem rather than a content problem. Publisher exports are frequently vendor-specific rather than QTI. LMS migration and package conversion handles the conversion, and it is worth doing after the mapping rather than before, so you are only converting the items you kept.
Adapting a Publisher Test Bank
Fill in the form and your workspace opens with the work already underway.