> For the complete documentation index, see [llms.txt](https://docs.lamsfoundation.org/lams/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.lamsfoundation.org/lams/question-bank.md).

# Question Bank

## Introduction to the Question Bank

The **LAMS Question Bank** is an organised digital library of assessment questions. You can use it to **create, store, organise, share, analyse, revise and reuse questions** across lessons and courses.

Instead of writing every quiz or exam from the beginning, you can select questions from the bank and add them to LAMS activities such as [**Assessment**](/lams/tools/assessment.md)**,** [**Scratchie**](/lams/tools/scratchie.md) **and supported** [**Q\&A**](/lams/tools/q-and-a.md) **workflows**. This makes routine assessment authoring faster while preserving the **history and evidence** associated with each question.

The Question Bank is **more than a repository**. It supports the **complete lifecycle of a question**:

1. Create the question manually, generate a draft with AI or import it.
2. Organise it in a collection.
3. Tag the learning outcome or outcomes it assesses.
4. Use it in one or more LAMS lessons.
5. Review its performance and student responses.
6. Improve it by creating a new version.
7. Reuse the most appropriate version in a future assessment.

This cycle helps you develop a stronger body of assessment evidence over time rather than treating each test as a separate, disposable document.

### Why use the Question Bank?

The Question Bank helps you:

* find and reuse questions efficiently;
* organise questions by course, topic, purpose or assessment programme;
* collaborate with other lecturers through shared collections;
* generate draft questions with AI from a topic and optional source material;
* map questions to learning outcomes;
* review item performance across lessons and semesters;
* preserve earlier versions while improving a question;
* import and export assessment content; and
* build a more coherent programme of assessment.

**Centralisation does not guarantee quality.** A reused question may no longer match the teaching context, current evidence, student cohort or intended learning outcome. **Review every question** before adding it to a new assessment.

### Question collections

Every question belongs to a **question collection**. Collections help you **organise and control access** to assessment content.

You can use collections for:

* a subject or course;
* a topic or curriculum area;
* a year or programme stage;
* formative or summative assessment;
* a team of lecturers; or
* a controlled item-development project.

#### Private collections

A **private collection** is available only to you. Use one for draft questions, personal resources or items that are not ready to share.

**Private does not necessarily mean suitable for secure high-stakes content** under every institutional policy. Follow your organisation's requirements for assessment security, retention and access.

#### Shared collections

Use a **shared collection** when you want to develop and review questions collaboratively with lecturers from the courses you teach.

A shared collection can support:

* co-authoring;
* peer review and moderation;
* consistent terminology and standards;
* shared coverage of learning outcomes;
* coordinated assessment across subjects or course instances; and
* continuity when teaching responsibilities change.

Agree how your team will **name, review, approve and revise questions**. Sharing access without a common workflow can produce duplicates, inconsistent versions or unreviewed changes.

### Generate questions with AI

Within a collection, you can use **generative AI to create draft questions**. Describe the topic, optionally provide source material, select a target cognitive level and request between **1 and 10 questions**.

Source material can include **PDFs, presentations and other supported media**. Providing relevant material can help ground the generated questions in the content students were expected to learn.

#### Prepare the generation request

Give the AI enough information to produce useful items. Include:

* the topic and its boundaries;
* the course or learner level;
* the knowledge or skill you want to assess;
* any terminology, context or constraints that matter;
* the desired cognitive level; and
* suitable source material, when available.

Request a **small, purposeful set** when you intend to review questions carefully. Generating ten questions is useful for exploring alternatives, but it **does not guarantee ten usable items**.

#### Select a cognitive level based on Bloom’s Taxonomy

Choose the **level of cognitive thinking** you want the questions to target, such as recall, understanding, application or analysis. Use this selection to **guide generation, not as proof of alignment**.

These cognitive levels are:

* **Remember:** Recalling facts, terms, or basic concepts.
* **Understand:** Explaining ideas or the meaning of data.
* **Apply:** Using information in new or practical situations.
* **Analyse:** Breaking information down to explore relationships and patterns.
* **Evaluate:** Making judgments, critiques, or decisions based on evidence.
* **Create:** Putting parts together to build or pitch something original

Review the **task students must actually perform**. A question may mention an authentic scenario while still asking only for factual recall. Tag a learning outcome only after confirming that the generated question assesses the action and level required by that outcome.

#### Review every generated question

Treat **generated questions as drafts requiring lecturer judgement**. Before you approve or use one:

* verify every factual claim against an authoritative source;
* check that the answer key is correct and unambiguous;
* ensure distractors are plausible but incorrect;
* remove clues, stereotypes and unnecessary complexity;
* confirm that the cognitive demand matches the selected level;
* check alignment with the intended learning outcome;
* review scoring, feedback and accepted responses;
* check accessibility and inclusive language;
* preview the question in its receiving LAMS activity; and
* ask a colleague to review consequential items.

**Do not assume that grounding material prevents errors.** Generated content can misinterpret a source, combine unrelated passages or invent unsupported details. **Check the question against the original material** rather than against the AI's explanation.

#### Use source material responsibly

{% hint style="warning" %}
**Responsible AI use**
{% endhint %}

Despite the fact that **LAMS AI does not use any resources of file uploaded to training any LLM models, upload only material that you are permitted to use** and that your institution allows you to process with generative AI. 

Do not provide **confidential assessment content, personal data, restricted student work or third-party material** without the required rights and safeguards.

Check generated wording for excessive reproduction of source material. A Question Bank item should assess learning rather than reproduce a passage simply because it appeared in an uploaded document.

#### Preserve academic responsibility

**The lecturer remains responsible** for the validity, accuracy, fairness and security of every generated question. AI can accelerate drafting and suggest alternatives; it cannot decide whether an item is appropriate for your students or defensible in a high-stakes assessment.

Once reviewed, manage an AI-generated question like any other banked item: **tag its learning outcomes, record its provenance, use versioning and review its analytics**.

### Reuse questions across LAMS activities

You can **add banked questions to supported LAMS activity tools** instead of recreating them. Depending on your learning design, the same underlying question may contribute to:

* a low-stakes practice activity;
* a diagnostic quiz;
* an individual readiness test;
* a team assessment such as Scratchie;
* an application exercise; or
* a formal examination.

**Not every LAMS activity supports every question type or setting.** The receiving activity also controls aspects of delivery such as attempts, feedback timing, answer disclosure, question order and scoring behavior. **Preview the complete student experience** each time you reuse a question in a different activity.

Reuse the content when it remains valid, but **reconsider the assessment interpretation**. A question suitable for immediate-feedback practice may require different security, weighting and feedback settings in a summative examination.

### Available question types

The Question Bank supports several ways of collecting evidence from students:

* [Multiple-choice questions](/lams/question-bank/multiple-choice-questions.md)
* [Matching pairs](/lams/question-bank/matching-pairs.md)
* [Very short answers](/lams/question-bank/short-answers.md)
* [Numerical questions](/lams/question-bank/numerical-questions.md)
* [True/False questions](/lams/question-bank/true-false.md)
* [Essay questions](/lams/question-bank/essay.md)
* [Ordering questions](/lams/question-bank/ordering.md)
* [Mark distribution questions](/lams/question-bank/mark-distribution.md)

Choose the question type that **captures the evidence required by the learning outcome**. Do not default to the easiest format to mark when students need to construct, explain or perform something.

### Track a question throughout its lifecycle

When you reuse a question, the Question Bank can retain information about the lessons in which it has appeared and its performance in each context. This allows you to **examine an item across assessments and semesters** rather than judging it from a single use.

Depending on the question type and available response data, item analysis can include:

* **Difficulty:** the proportion of students who answered correctly. A higher value normally indicates an easier question.
* **Discrimination:** how well the question differentiates between stronger and weaker overall performance.
* **Point-biserial correlation:** the relationship between performance on the question and performance on the assessment as a whole.
* **Answer distribution:** how students used the available answers, including which distractors they selected.

LAMS provides Question Bank analytics across uses of a question ([LAMS analytics guidance](https://docs.lamsfoundation.org/tbl/tbl-for-teachers/analytics)). Use these statistics as prompts for academic review, not automatic rules.

Interpret analytics with the:

* number of students;
* purpose and stakes of the assessment;
* characteristics of the cohort;
* teaching and learning opportunities provided;
* position and context of the question;
* version of the question used; and
* construct measured by the assessment.

A difficult question is not necessarily a good question, and an easy question is not necessarily weak. Unexpected results may indicate ambiguity, an incorrect key, inaccessible content, inadequate teaching coverage or a genuine misconception.

### Create and manage question versions

When analytics or academic review identify a problem, you do not need to discard the question and lose its history. **Create a new version** and improve it while preserving the earlier version and its evidence.

You might create a new version to:

* clarify an ambiguous prompt;
* replace a non-functioning distractor;
* correct factual or technical content;
* improve feedback;
* change scoring or accepted responses;
* improve accessibility;
* align the question more closely with a learning outcome; or
* adapt it to a new context without overwriting the original.

**Versioning lets you compare the performance of a revised item with its predecessor.** It also gives you a deliberate choice: use the revised version in a new activity or retain an earlier version when that version is more appropriate.

#### When to create a new version

Create a new version when a change could affect how students **understand, answer or score** on the question. This includes changes to the prompt, options, answer key, accepted error, unit conversion, feedback shown before another attempt or grading rule.

Minor metadata corrections may not alter the construct, but follow your local versioning policy. Preserve a clear record of what changed and why.

**Do not compare analytics from materially different versions as though they came from one unchanged question.** Improvement requires both continuity of history and honest separation of changed evidence.

### Import and export questions

#### IMS Question and Test Interoperability

LAMS supports **IMS Question and Test Interoperability (QTI)** for exchanging assessment content. QTI is an open 1EdTech specification designed to move questions and tests between authoring tools, item banks, learning platforms and assessment systems ([1EdTech QTI](https://www.1edtech.org/standards/qti)).

Use QTI when you want to:

* move supported questions between systems;
* preserve assessment content in a standardised package;
* reduce manual re-entry; or
* support migration and institutional ownership of assessment material.

**QTI support varies by system, specification version and question type.** After importing, verify the prompt, formatting, media, answer key, scoring, feedback, metadata and accessibility. Successful file import does not prove that the question will behave identically in another platform.

#### Microsoft Word templates

You can also import questions using the **LAMS Microsoft Word template**. This is useful when lecturers prefer to author or review questions in a familiar document format or need to prepare several questions in bulk.  <i class="fa-file-word" style="color:purple;">:file-word:</i> [Template file](https://raw.githubusercontent.com/lamsfoundation/lams/v4.0/lams_www/web/public/MSWord-question-import.docx)

The template can support:

* formatted question and answer text;
* images and tables;
* feedback; and
* learning-outcome metadata.

**Follow the template structure exactly.** After upload, review the questions that LAMS detects and select the ones you want to import. Then check all formatting, media, scoring and metadata in the Question Bank. LAMS documents both Word and QTI import workflows in its [question-import guidance](https://docs.lamsfoundation.org/tbl/tbl-for-teachers/getting-started/tbl-design-wizard-1/readiness-assurance-questions).

### Map learning outcomes to questions

You can **tag each question with one or more learning outcomes**. This connects curriculum intent with the assessment evidence collected from students.

Outcome mapping can help you answer:

* Where is this outcome assessed?
* Which question types provide evidence for it?
* How much assessment weight does it receive?
* How have students performed on questions associated with it?
* Is the outcome assessed at the intended cognitive or performance level?
* Where are the gaps or unnecessary repetitions in the assessment programme?

**Tag only outcomes for which the question provides meaningful evidence.** A question about the topic of an outcome is not necessarily aligned with the action required by that outcome. For example, recalling a definition does not demonstrate an outcome that requires students to evaluate, design or perform.

**Outcome tags support curriculum mapping; they do not prove validity by themselves.** Review the prompt, required response, scoring and teaching context together.

### Use the Question Bank within a programme of assessment

The Question Bank can support **programmatic assessment** by connecting many assessment data points across time. Individual assessment events should provide useful feedback and learning opportunities, while **higher-stakes decisions draw on a broad body of evidence** rather than one test or question ([van der Vleuten et al., 2012](https://pubmed.ncbi.nlm.nih.gov/22364452/)).

Use the bank to plan purposeful sampling across:

* learning outcomes;
* topics and contexts;
* cognitive demands;
* question and response formats;
* formative and summative purposes; and
* stages of the course or programme.

**Do not allow the availability of existing questions to determine the curriculum.** Start with the evidence you need, then select or create questions that supply it.

### A practical quality workflow

Before first use:

1. Align the question with a learning outcome and assessment purpose.
2. Select the question type that captures the required evidence.
3. Review the content, answer key, scoring, feedback and accessibility.
4. Ask a colleague to review consequential questions.
5. Preview the question in the receiving LAMS activity.

After use:

1. Review student responses and item analytics.
2. Investigate unexpected patterns and student challenges.
3. Decide whether the issue lies in the question, scoring, teaching or context.
4. Retain the version when it remains sound, or create an improved version.
5. Record the reason for the change.
6. Reassess the revised question after its next use.

### Key principle

**Reuse questions, but do not reuse them uncritically.** The value of the Question Bank comes from combining **efficient authoring, outcome alignment, collaborative review, version history and evidence-informed improvement**.


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:

```
GET https://docs.lamsfoundation.org/lams/question-bank.md?ask=<question>&goal=<endgoal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.