A successful import preserves the teaching decisions inside your questions, not just their text. Start with a small reviewed sample, agree the target format and verify scoring before moving the full bank.
Separate content preparation from the upload
A question bank is more than a folder of worksheets. Each question needs a stable reference, a clear answer key, useful feedback and enough context to find it again. Converting a file does not verify the subject matter.
Keep the original files unchanged. Make a working copy for cleanup and record which questions you own or have permission to use. Do not include real learner responses or personal information in a content-conversion sample.
Build a small editorial inventory
Start with one topic and roughly ten representative questions. Include the difficult cases, not just plain text: a diagram, a calculation, multiple correct answers or a question with detailed feedback, where your course uses them.
For each item, record:
- Source reference: a stable ID and the original file or page.
- Question: the full prompt, including units and necessary context.
- Answers: all options and an unambiguous correct-answer key.
- Explanation: the reasoning, not merely a repetition of the answer.
- Classification: topic, intended question type and review status.
- Media: the relevant file and any caption or attribution.
- Review owner: the person who will check the converted result.
This inventory is for editorial work. It is not a universal upload schema.
Work through one example
Suppose your question asks: “A cart travels at a constant 3 m/s for 4 s. What distance does it travel?” The correct answer is 12 m because distance equals speed multiplied by time.
An inventory could identify it as PHY-MOTION-001, with topic constant speed, one correct option and a reviewed explanation. Keep the units with both the prompt and answer. If you randomise the answer options, verify that the key still points to 12 m rather than an old option position.
This is an illustrative question, not an import-ready file. Your bank’s actual schema must match the receiving system.
Agree the conversion before sending the whole bank
Ask the implementation team which formats and question types are supported for your project. Specify how source IDs, tags, images, equations, answer keys and explanations will map.
For PrepOS, confirm the import or content-setup route for your course before preparing a bulk JSON file. Do not assume a PDF, spreadsheet or another LMS’s export can be uploaded unchanged. Some source types need conversion or manual rebuilding.
Decide how duplicate IDs and revised questions will be handled. Ask what happens to an existing question if the same sample is processed twice.
Validate the sample in the learner view
Compare every converted question with the source. Submit a correct answer, a wrong answer and, where applicable, a partially correct answer. Check scoring and feedback timing separately from appearance.
Open diagrams and mathematical notation on a phone. Ask a second subject reviewer to check the sample independently. Record rejected questions and their reasons instead of silently dropping them.
Approve a pilot only when the item count, answer keys, explanations and topic mapping agree. For the full bank, repeat validation on every unsupported or transformed format and use a documented review sample for the rest. Keep an exception list and a rollback copy.
Make the bank maintainable
Name an owner for corrections after launch. Reuse a question in a second practice set and establish whether an edit updates the shared question or a copy. Test a revision before making it visible to a live cohort.
Explore PrepOS exams and question practice, the content editor and the academy migration checklist. Bring the reviewed sample to the implementation discussion so the scope is based on real material.
