Why your own bank wins
| Commercial bank | Your own bank | |
|---|---|---|
| Coverage | The subject as generally taught | Exactly your syllabus, including your lecturer's pet topics |
| Emphasis | Even across the field | Weighted the way your marks are |
| Notation and conventions | The textbook's | Your department's, which is what's marked |
| Building cost | None | Real, and it's also the revision |
| Best used | Volume practice once yours is exhausted | Primary revision resource |
This isn't an argument against commercial banks — for high-volume pattern practice in subjects like NCLEX or USMLE prep they're essential. It's an argument that they're a supplement to a bank shaped by your course, not a replacement for one.
What makes a question worth keeping
- One defensible answer. If two answers are arguable, it's a discussion prompt, not a test item.
- Answerable without the source open. "What does the diagram on slide 12 show?" fails; it tests the deck, not the subject.
- Not free. If elimination or grammar gets you there without knowing anything, it inflates your score and teaches nothing.
- Aimed at application where possible. "Given this scenario, what happens?" beats "define X" on almost every real paper.
- Worth remembering. If you wouldn't care about forgetting it, it doesn't earn a place in a queue you'll clear daily.
The four question types to write for each topic
Writing one question per fact produces a bank that tests recall and nothing else. Four per topic, in these shapes, covers the range a real paper uses.
- 1
A recall question — the arbitrary fact
The value, the name, the date, the classification. Genuinely necessary and the easiest to write, which is why most banks are entirely made of these.
- 2
An application question — a scenario
Give a specific case and ask what happens or what you'd do. This is where most exam marks are and where most self-made banks are thinnest.
- 3
A discrimination question — this or that?
Present the topic alongside the thing it's most confusable with. Examiners target confusions, so find yours and card them together.
- 4
A failure question — what if this breaks?
What happens if this step is blocked, this assumption relaxed, this component missing? These are the highest-yield questions in mechanistic subjects and they generate the pathology or edge-case content for free.
- 5
Optionally, an error-spotting question
Here's a wrong answer — what's wrong with it? Trains the checking skill you'll use on your own work under exam pressure.
Build it weekly, not in revision week
Written the same week as the lecture, questions cost ten minutes and capture what the lecturer emphasised. Written in revision week, they cost hours and come from a textbook, because the lecture is gone — which means the bank you build at the end is the generic one you were trying to avoid.
Ten to fifteen questions per lecture, produced in the processing slot of your study system. Across a twelve-week module that's around 150 questions, which is a substantial revision resource acquired at two hours a week that you were spending on note-making anyway.
Store them so they get used
- Question and answer separated, so you can't see the answer while attempting. A document with both visible becomes reading.
- Tagged by topic, so you can pull a mixed set across the whole module — which is the interleaved practice the exam actually resembles.
- In something that schedules them. A bank you have to remember to open won't get opened; a queue that surfaces due items will.
- With a source reference. When a question looks wrong months later, you need to check it in seconds rather than re-derive it.
- Editable. You'll rewrite perhaps a fifth of your questions after attempting them, because writing a good question is harder than it looks.
Use it as a test, not as a document
The failure mode is reading your question bank. It has answers in it, it's well-organised, and reading it feels efficient — and it's rereading with extra steps.
Attempt cold, write the answer down rather than thinking "I know this", check, and mark honestly. The gap between "I know this" and a written answer is exactly the gap that shows up in an exam, and thinking-through is how people arrive at a paper surprised.
Sharing with your cohort
A shared bank is one of the few genuinely efficient forms of group study: five people writing fifteen questions per lecture produces a bank nobody could build alone, from the same lectures, with the same emphasis.
Two rules make it work. Everyone writes their share before anyone gets access — otherwise you get four readers and one author. And every question carries its writer's name, so a wrong answer can be queried rather than silently learned. Study groups covers the wider structure; the question bank is usually the highest-value thing a group produces.