Generating Question Banks for Financial Literacy With AI
A financial-literacy question bank needs one tag beyond the standard five (content standard, Bloom's level, format, difficulty, last-used): claim shape — whether an item is categorical, a magnitude comparison, or a decision. Format alone can't substitute, because two of the subject's four claim shapes both route to multiple choice, collapsing into one indistinguishable tag.
That's a specific, testable claim, not a restatement of "financial literacy needs its own bank guide." This series already built the pieces: a general Bloom's-aligned bank architecture, a distractor-and-dominance-test system for financial-literacy quiz items, and a four-claim-shape routing framework connecting every format this subject uses. None of them asked what a bank built from those items should tag by. This guide does.
Quick Answer: A financial-literacy question bank inherits the standard five-tag architecture — content standard, Bloom's level, item format, difficulty, last-used — and needs exactly one new dimension: claim shape (categorical, magnitude, or decision). The gap is real because magnitude items (compound interest) and well-structured decision items (a clean opportunity-cost choice) both get scored as multiple choice, so the format tag alone can't tell them apart later.
This guide covers:
- What a financial-literacy bank already inherits from general bank architecture, including one candidate tag (domain) that turns out not to be new
- Where item format already separates claim shapes, and the one place it stops working
- The claim-shape tag itself, tested against a concrete retrieval scenario
- A full six-tag set and a step-by-step build workflow, with a worked classroom example
- What's known, and not known, about AI's ability to tag claim shape automatically
What a Financial-Literacy Bank Inherits From General Bank Architecture
A financial-literacy question bank starts from the same five tags any Bloom's-aligned bank needs, and four of them carry over untouched. AI Question Banks Aligned to Bloom's Taxonomy set the baseline: content standard, Bloom's level, item format, difficulty, and last-used/exposure, checked against a living table-of-specifications blueprint (Notar, Zuelke, Wilson, & Yunker, 2004).
The Parts That Transfer Without Modification
Difficulty, last-used, and Bloom's level need no financial-literacy-specific adjustment. An item is easy or hard, seen or unseen, Remember or Analyze, the same way in any subject. That leaves content standard and item format, where the subject's real shape starts to matter.
Domain Isn't a Missing Tag — It's Already the Standard Tag
It's tempting to reach for the Jump$tart Coalition and Council for Economic Education's six domains — Earning Income, Spending, Saving, Investing, Managing Credit, and Managing Risk (Jump$tart Coalition & Council for Economic Education, 2021) — as a new axis a financial-literacy bank needs on top of the general five.
It isn't one. Those six domains are this subject's content standard, doing the exact job "content standard" already does in any Bloom's-aligned bank: letting a teacher pull items by topic. Filing them under a second tag alongside "content standard" would just duplicate the same information under a new name.
- Content standard, for financial literacy: one of the six Jump$tart/CEE domains, the same role Common Core's W.1/W.2/W.3 plays for writing
- What would make it a genuinely new tag: information the standard tag doesn't already capture
- What domain actually captures: exactly what "content standard" already captures, just named for this subject
The honest finding here is a negative one, worth stating before going further: domain is not the missing tag. Writing's own question-bank guide reached the identical non-finding about genre — Common Core's writing standards already served as that subject's content-standard tag, so genre didn't need a second one either. Financial literacy's domain scaffold plays the same role here.
Where Item Format Already Separates Claim Shapes — And Where It Stops
Item format does real work sorting financial-literacy claim shapes, just not all the work. The routing guide connecting this subject's seven formats established four claim shapes: categorical, magnitude, well-structured decision, and ill-structured decision, drawing the well-structured/ill-structured line from educational-technology researcher David Jonassen's problem-solving framework (Jonassen, 1997). Two of those four are already format-transparent.
Categorical and Ill-Structured Decisions Already Have Their Own Format
A categorical claim — true for every household or false for every household, regardless of unstated assumptions — routes to true/false, using the invariance test the true/false guide built. "Carrying a balance builds your credit score" is false the same way for everyone; nothing about it needs a claim-shape tag once you already know the item is filed as true/false.
An ill-structured decision — one that fails the quiz guide's dominance test because more than one option is defensible — routes to short-answer scored with a Claim-Evidence-Reasoning rubric, adapted from science-assessment research (McNeill & Krajcik, 2012). In this series, nothing else currently lands in short-answer, so that format tag is claim-shape-transparent too.
| Claim Shape | Routes To | Format Tag Alone Identifies It? |
|---|---|---|
| Categorical | True/false | Yes — no other claim shape uses this format |
| Ill-structured decision | Short-answer + CER rubric | Yes — no other claim shape uses this format |
| Magnitude | Multiple choice | No — shares this format with well-structured decisions |
| Well-structured decision | Multiple choice | No — shares this format with magnitude items |
The Ambiguity Format Can't Resolve
Two of the four claim shapes — magnitude and well-structured decision — both get scored as multiple choice. The quiz guide built different design machinery for each: a three-surface distractor set (linear-growth, compounding-period, arithmetic-slip), grounded in documented exponential-growth-bias research (Stango & Zinman, 2009), for magnitude items, and a dominance test for decision items. Once filed, "multiple choice" is the only format information either one carries.
That's the real gap. A bank filtered by format=multiple-choice returns both kinds mixed together, with no tag distinguishing a compound-interest computation from a clean opportunity-cost choice.
The Tag That's Actually Missing: Claim Shape
What neither the standard tag nor the format tag captures is which of financial literacy's three practically distinct claim shapes — categorical, magnitude, or decision — a multiple-choice item actually tests. That gap matters because it sits exactly where a teacher's real retrieval need lives: pulling a set of items that share a reasoning demand, not just a grading mechanic.
Why Bloom's Level Clusters Instead of Separating Them
Bloom's level doesn't rescue the ambiguity either, though the failure looks different here than in other subjects. A magnitude item like a compound-interest computation typically sits at Apply, the kind of reasoning Lusardi and Mitchell's widely used "Big Three" financial-literacy instrument was built to test (Lusardi & Mitchell, 2011). A well-structured decision item, working through an opportunity-cost trade-off, also typically sits at Apply, sometimes Analyze.
Both claim shapes cluster at the same one or two Bloom's levels most of the time. A tag that doesn't discriminate between them can't be the tag a teacher filters or audits by — the same problem the general bank-architecture guide warned about for an unmanaged bank generally, just showing up here for a different reason: clustering, not saturation.
Testing Claim Shape Against a Concrete Retrieval Scenario
Say a bank holds items tagged Managing Credit, multiple choice, Apply level, medium difficulty. Two items fit equally well:
- "A cardholder carries a $500 balance at 22% APR, compounded monthly, for six months with no payments. How much interest accrues?" — a magnitude item, testing whether a student models compounding correctly.
- "A shopper can pay with a rewards credit card offering 2% cash back and no fees, paid in full monthly, or a debit card with no cash back and no fees. Which is the better choice?" — a well-structured decision item; one option dominates on every stated criterion, so it passes the dominance test cleanly.
Domain, format, Bloom's level, and difficulty tag both items identically. A teacher who wants five items rehearsing opportunity-cost reasoning specifically, not compounding arithmetic, has no tag left to filter on until claim shape gets added.
Domain answers which topic. Format answers how it's graded. Neither answers what kind of reasoning the item actually rehearses — that's what claim shape is for.
The Tag Set a Financial-Literacy Question Bank Actually Needs
A workable financial-literacy bank keeps the general five tags and adds exactly one: claim shape. That's a targeted, single addition, not a rebuild of the general architecture the bank-architecture guide already established.
| Tag | Carries Over From a General Bank? | Notes for Financial Literacy |
|---|---|---|
| Content standard | Yes, unchanged | Already maps to the six Jump$tart/CEE domains (2021) |
| Item format | Yes, unchanged | Fully separates categorical and ill-structured decisions; doesn't separate magnitude from well-structured decisions |
| Bloom's level | Yes, but weaker signal for one split | Magnitude and well-structured decision items cluster at the same one or two levels |
| Difficulty | Yes, unchanged | Easy/medium/hard, or a tracked pass rate |
| Last used/exposure | Yes, unchanged | Same purpose as any bank, any subject |
| Claim shape | No — new for financial literacy | Categorical, magnitude, or decision; format already splits decision into well- vs. ill-structured once claim shape is known |
Recording Claim Shape as Three Values, Not Four
The routing guide's four claim shapes collapse to three practical tag values once format is already known. Categorical and ill-structured-decision items don't need the claim-shape tag to be found, since their format already identifies them. Only "decision" needs recording as its own value, because format alone can't tell a magnitude item from a well-structured decision item.
- Categorical — format already identifies it (true/false); the tag is informational, not load-bearing
- Magnitude — the value that actually disambiguates a multiple-choice item from a decision item
- Decision — the other value that disambiguates; format then tells you whether it was well- or ill-structured (multiple choice vs. short-answer)
Building a Claim-Shape-Tagged Bank: A Step-by-Step Workflow
Building a financial-literacy bank follows the general bank workflow with one extra decision inserted at the moment an item is written: naming its claim shape before it ever gets filed.
- Set a blueprint across two axes — the six Jump$tart/CEE domains down one side, claim shape across the other — instead of the single domain-by-Bloom's-level grid a general bank uses.
- Run the shape-appropriate check while writing the item: the invariance test for a categorical claim, the three-surface distractor design for a magnitude claim, or the dominance test for a decision claim.
- Let the check's outcome set the tag. An item that passes the invariance test is categorical; an item built around named distractor surfaces is magnitude; a decision item that passes the dominance test is a well-structured decision, tagged for multiple choice.
- Route a decision item that fails the dominance test to short-answer with a Claim-Evidence-Reasoning rubric, tagged ill-structured decision.
- Tag every item on all six dimensions at once — standard, Bloom's level, format, difficulty, claim shape, last-used — the moment it's written, not in a separate pass.
- Audit claim-shape distribution alongside domain distribution. A bank can look domain-balanced across all six Jump$tart/CEE areas and still be almost entirely magnitude items, with well-structured decisions barely represented.
- Check format and claim shape together during an audit. A "decision" tag with no matching short-answer or clean-dominance multiple-choice item behind it signals a mistagged item.
A Worked Example: One Unit, Tagged by Claim Shape
Say you teach a grade 8 personal-finance elective and you're building a standing bank for a unit spanning Saving and Managing Credit, one you plan to reuse across sections and future years.
Tagged only by the general five dimensions, a filtered pull for "Managing Credit, multiple choice, Apply level" returns compound-interest computations and opportunity-cost decisions mixed into one list, indistinguishable without opening each item. Add the claim-shape tag, and the same pull splits cleanly: five magnitude items for a compounding-focused review day, or five well-structured decision items for a lesson built around trade-off reasoning.
Can AI Tag Claim Shape Accurately?
No dedicated accuracy research exists yet for classifying financial-literacy items by claim shape, which is an honest gap rather than a small caveat. The tagging-speed guide found real, published accuracy figures for automated Bloom's-level classification: roughly 72-73% for general AI tools used with no special training, climbing to the low-to-mid 90s for classifiers built specifically for the task (Kumar, Gulwani, & Singh, 2025). Nothing with that kind of dedicated research base has been published yet for claim-shape classification specifically.
Tag-at-Creation Is Cheaper Here Than It Sounds
That gap matters less for claim shape than for a purely subjective label, for one reason: claim shape isn't a judgment call layered onto an already-written item. It falls directly out of a check the quiz guide already requires — the invariance test, the distractor design, or the dominance test — before the item is finished at all.
- Writing a categorical item already means confirming it passes the invariance test; the categorical tag is a byproduct, not extra work.
- Writing a magnitude item already means naming three distractor surfaces; the magnitude tag falls out of that design step.
- Writing a decision item already means running the dominance test to decide its format; the well-structured/ill-structured split falls out of that same test.
Retroactive Tagging Still Needs a Human Read
For an old backlog with no claim-shape recorded, the verification-sampling method built for Bloom's-level tagging is the safest fallback: auto-tag a first pass, then read a sample cold. That matters because a tag determining which items reach a lesson or exam deserves a human check, in line with general guidance to keep a person in the loop for AI-assisted instructional decisions (ISTE, 2023).
Treat an AI-suggested claim-shape tag on an existing item as a first draft. There's no accuracy study yet to fall back on if a later guess turns out wrong, which is exactly why naming the shape while writing the item — when the design check has already been run — is worth the small habit change.
Tools for Building a Financial-Literacy Question Bank
No AI tool ships with a claim-shape field the way some flag Bloom's level automatically; the practical difference between tools is how easily a teacher can name a target claim shape up front and get back an item that actually reflects it.
| Tool | Claim Shape Named at Creation | Distractor/Dominance Checks Supported | Format Range | Starting Cost |
|---|---|---|---|---|
| EduGenius | If stated in the prompt or class profile | Yes, if all three distractor surfaces or the dominance criteria are requested by name | True/false through case-study and exam formats | 25 welcome credits free; Starter $7.99/month (500 credits); Professional $15.99/month (1,000 credits) |
| General AI chat tools | Only if typed out each time | Only if explicitly requested | Yes, if prompted correctly | Varies by provider |
| Publisher financial-literacy item banks | Rarely tagged this specifically | Editorially reviewed, not shape-tagged | Reliable within that program's bank | Included with curriculum adoption |
| Manual, teacher-built bank | Fully teacher-controlled | Fully teacher-controlled | Fully teacher-controlled | Teacher's own prep time |
You could use EduGenius's class profile to hold a grade level and domain emphasis, then generate items across a unit, naming the target claim shape directly in the prompt: "a magnitude item with three named distractor surfaces," or "a decision item that passes the dominance test." Its Bloom's Taxonomy-aligned generation already helps with the level side of tagging; claim shape still depends on what a teacher names in the request.
For the wider category, The Ultimate Guide to AI-Powered Assessment and Quiz Generation covers quiz and grading automation across subjects, and AI Worksheet Generators: The Complete 2026 Guide covers classroom materials beyond quizzes and banks.
Pro Tips for a Claim-Shape-Balanced Financial-Literacy Bank
- Tag claim shape the same moment you run the design check, not in a separate pass — the tag is nearly free once the invariance test, distractor design, or dominance test is already done.
- Audit claim-shape distribution the same way you'd audit Bloom's-level distribution. A bank can look perfectly balanced by domain and still be almost entirely magnitude items.
- Don't let format substitute for claim shape inside multiple choice. "Multiple choice" confirms one scored key, not whether an item rehearses compounding arithmetic or trade-off reasoning.
- Reserve the heaviest verification for decision items feeding a graded exam, the same priority the tagging-speed guide recommends for Bloom's-level risk generally.
- Keep the dominance test as a live check, not a one-time filter. A decision item that passed it when written can stop passing once a lesson changes what "best" depends on.
What to Avoid When Building a Financial-Literacy Question Bank
Treating the Six Domains as a Second New Tag
The Jump$tart/CEE domains do the same job "content standard" already does in any Bloom's-aligned bank. Filing them as a separate axis on top of content standard duplicates information instead of adding it.
Assuming Multiple Choice Means the Same Reasoning Demand Every Time
A "multiple choice, Apply level, Managing Credit" filter can return a compounding computation and an opportunity-cost decision with no way to tell them apart. Tag claim shape, or accept the filter will mix two different skills.
Skipping the Dominance Test Because an Item "Looks" Well-Structured
The quiz guide found that a cost-benefit item can look clean on the page while still lacking one stated criterion that would make it well-structured. Run the test before tagging, not after a confused discussion reveals two defensible answers.
Trusting an AI Claim-Shape Tag the Way You'd Trust a Bloom's-Level Tag
The accuracy research behind Bloom's-level classification (Kumar, Gulwani, & Singh, 2025) has no published equivalent for claim-shape classification. Treat any retroactive AI claim-shape tag as unverified until a human reads the item against the relevant check.
Key Takeaways
- A financial-literacy question bank inherits general bank architecture almost entirely intact — content standard, item format, Bloom's level, difficulty, and last-used all carry over unchanged.
- The six Jump$tart/CEE domains aren't a missing tag. They already do the job "content standard" does in any Bloom's-aligned bank, the same non-finding writing's own question-bank guide reached about genre.
- Item format already separates two of the four claim shapes — categorical routes to true/false, ill-structured decisions route to short-answer — with no ambiguity left to resolve for either.
- The genuinely missing tag is claim shape, because magnitude items and well-structured decision items both get scored as multiple choice, collapsing into one indistinguishable format tag.
- Bloom's level doesn't rescue the gap either. Both claim shapes typically cluster at Apply, sometimes Analyze, with no discrimination left over for a teacher to filter on.
- Claim shape is nearly free to tag at creation, since it falls directly out of a check — the invariance test, the distractor design, or the dominance test — the item's design already requires.
- No published accuracy research yet covers automated claim-shape classification, which makes tag-at-creation more valuable here than a retroactive fix.
Frequently Asked Questions
Is a financial-literacy question bank different from a general Bloom's-aligned question bank?
Mostly the same, with one real addition. A financial-literacy bank keeps the same content-standard, item-format, Bloom's-level, difficulty, and last-used tags a general Bloom's-aligned bank uses, and adds a claim-shape tag — a dimension this subject needs because two of its four claim shapes both get scored as multiple choice.
Why isn't the Jump$tart/CEE domain list a separate tag from content standard?
Because it already plays that role. Earning Income, Spending, Saving, Investing, Managing Credit, and Managing Risk (Jump$tart Coalition & Council for Economic Education, 2021) function as this subject's content-standard tag, the same way Common Core's W.1/W.2/W.3 standards already serve as writing's content-standard tag.
Why can't item format alone organize a financial-literacy question bank?
Because two of the subject's four claim shapes — a compound-interest computation and a clean opportunity-cost decision — both get scored as multiple choice. A bank filtered by format=multiple-choice returns both mixed together, with no tag left to separate a growth item from a decision item.
Can AI accurately tag which claim shape a financial-literacy item targets?
There's no published accuracy research on this specifically, unlike Bloom's-level classification, where studies report roughly 72-73% accuracy for general AI tools and the low-to-mid 90s for purpose-built classifiers (Kumar, Gulwani, & Singh, 2025). Naming the shape while writing the item, when the invariance, distractor, or dominance check has already been run, sidesteps needing that research to exist yet.
Sources
- Bloom, B. S. (Ed.). (1956). Taxonomy of Educational Objectives: The Classification of Educational Goals.
- Anderson, L. W., & Krathwohl, D. R. (Eds.). (2001). A Taxonomy for Learning, Teaching, and Assessing: A Revision of Bloom's Taxonomy of Educational Objectives.
- Notar, C. E., Zuelke, D. C., Wilson, J. D., & Yunker, B. D. (2004). The table of specification: Insuring accountability in teacher made tests. Journal of Instructional Psychology, 31(2), 115-129.
- Jonassen, D. H. (1997). Instructional design models for well-structured and ill-structured problem-solving learning outcomes. Educational Technology Research and Development, 45(1), 65–94.
- Lusardi, A., & Mitchell, O. S. (2011). Financial literacy around the world: An overview. Journal of Pension Economics & Finance, 10(4), 497–508.
- Stango, V., & Zinman, J. (2009). Exponential growth bias and household finance. The Journal of Finance, 64(6), 2807–2849.
- McNeill, K. L., & Krajcik, J. (2012). Supporting Grade 5-8 Students in Constructing Explanations in Science: The Claim, Evidence, and Reasoning Framework for Talk and Writing. Pearson.
- Jump$tart Coalition for Personal Financial Literacy & Council for Economic Education. (2021). National Standards for Personal Financial Literacy.
- Kumar, R., Gulwani, D., & Singh, S. (2025). Automated analysis of learning outcomes and exam questions based on Bloom's taxonomy. arXiv preprint arXiv:2511.10903.
- International Society for Technology in Education (ISTE). (2023). Guidance on artificial intelligence in education.