Writing Your Own Practice Questions

Writing your own questions is not preparation for study — it IS study. To write a good question you must decide what actually matters in the topic, what a wrong answer would look like, how hard it should be, and what counts as a full answer. Those are the examiner's decisions, and making them forces exactly the kind of thinking that builds understanding. There's a catch though: your question quality caps your learning quality. Write only 'define X' questions and you'll only ever practise defining — then the exam asks you to apply, and it feels like a different subject. So this lesson gives you a ladder of five levels, and a recipe for mixing them.

Five levels, each with its own stems. LEVEL 1 RECALL: 'What is…?', 'Define…', 'List the…' — quick to write (twenty in ten minutes), objective to mark, good as warm-ups, but dangerous alone because they build false confidence. LEVEL 2 COMPREHENSION: 'Explain how… works', 'Describe… in your own words' — the rule here is that a good comprehension question can never be answered in one word. LEVEL 3 APPLICATION: 'What would happen if…?' — and critically, the scenario must NOT come straight from your textbook; it has to be genuinely new. LEVEL 4 ANALYSIS: 'Why this way and not another?', 'Compare X and Y' — generated by asking why about everything. LEVEL 5 SYNTHESIS: design, evaluate, connect across topics — save it until the lower levels are solid.

Here's the recipe for a balanced ten-question practice test: two to three Level 1 recall questions to establish a baseline, three to four Level 2 comprehension questions to check real understanding, two to three Level 3 application questions to test flexibility, and one or two at Level 4–5 to push toward mastery. Then sequence them easy to hard — early wins build the confidence to keep going when the difficult ones arrive. That mix matters because a test made only of recall questions will tell you you're ready when you aren't, and a test made only of synthesis questions will crush you before you've built a foundation.

Making your own test should take minutes, not hours, and you already own the raw material. HEADING HARVEST (10 minutes): list every heading and subheading in the chapter and turn each into a question — 'Light-Dependent Reactions' becomes 'What are the light-dependent reactions and how do they work?' CORNELL CONVERSION (5 minutes): if your notes have a question column down the left, your test already exists — cover the notes, answer the questions, check. OBJECTIVE HARVEST (15 minutes): take the learning objectives from your syllabus or textbook and turn each into a question; these are literally the list of what you're expected to know, which makes them the highest-value questions available.

Five ways a self-made question goes wrong. TOO EASY: recognition-based, answer obvious from the options — fix it by requiring you to generate and explain. TOO VAGUE: 'Explain photosynthesis' is unmarkable; specify the aspect. TRICK QUESTIONS: testing an obscure technicality trains trap-spotting rather than understanding. THREE QUESTIONS IN ONE: 'Explain X including where, when, why and how it links to Y' — split it up, or you'll never know which part you failed. NO SUCCESS CRITERIA: if you haven't written what a full answer must contain, you'll mark yourself generously every time. The habit that prevents all five: write the answer key as you write the question — if you can't, the question isn't finished.