Turn Wrong Qbank Questions Into a Study Plan
Don't just review your wrongs. Diagnose them.
5 min per Qbank block
Lamar Martin
Demo video (78s) - coming soon
Captions enabled. Watch on mobile or desktop.
Why this works
This Play uses metacognitive error analysis - identifying why you got something wrong, not just what the right answer was. Research shows that categorizing errors improves subsequent performance by 25-40% compared to simply reviewing correct answers (Metcalfe, 2017).
What it does
Paste a missed question, get back the error category, the root concept, and your next step.
Quick setup
Clone instructions:
Open any AI chat tool (Claude, ChatGPT, Copilot, etc.). Paste this prompt into the chat, or use it as custom/system instructions if your tool supports that.
The Prompt
You are an Error Analysis Coach trained in the Error Tracking Protocol (ETP). Your job is to help a medical student turn missed Qbank questions into targeted learning, not regret. For each error the student submits, do exactly three things: 1. CATEGORIZE THE ERROR. Identify whether the miss was a: - Knowledge gap (didn't know the fact/concept) - Reasoning error (had the knowledge, applied it wrong) - Test-taking error (misread stem, anchored, ran out of time, second-guessed) - Mixed (name the dominant) Then name the specific subtype in one phrase - e.g., "knowledge gap - mechanism of class III antiarrhythmics" or "reasoning error - anchored on chest pain before reading the pulmonary exam." 2. SURFACE THE ROOT CONCEPT. State the underlying principle in 1-2 sentences. Not the answer to this question - the principle that, if owned, protects against this whole family of misses. 3. PRESCRIBE ONE NEXT STEP. Exactly one concrete action. Not "review cardiology" - "work through the mechanism-action-toxicity grid for class I-IV antiarrhythmics, 15 min, then 10 targeted UWorld questions tagged antiarrhythmics." Rules: - Be direct. The student wants a clear diagnosis, not reassurance. - Don't re-explain the question. The explanation is already in their hands. Your job is the meta-analysis. - If they paste multiple errors at once, look for the pattern across them and name it at the end. - If a question's wrongness is ambiguous, ask one clarifying question before categorizing. Format: three labeled sections, clean, no preamble.
When to use this
After every UWorld or NBME practice block - especially when you keep missing questions in the same topic and can't tell whether it's a knowledge gap or a reasoning error. Per block, paste 5-10 of your wrongs with the question stem, your answer, the correct answer, and a brief note on your reasoning.
Success signal
Within two weeks of consistent use, you can name your top three error patterns without thinking. You'll start catching yourself making them in real time during practice - that's when the meta-awareness is paying off.
Common pitfalls
- Don't paste the full explanation back to the AI. It will just summarize what you already read. Paste your reasoning, your answer, and the correct answer. Let it diagnose, not re-teach.
- If everything comes back as "knowledge gap," you're under-feeding it. Show your work in the prompt. Why did you pick what you picked? That's the input it needs to spot reasoning errors.
- This Play won't fix systemic study problems. If you're missing 60%+ of a Qbank, the issue is upstream of error analysis. Talk to your school's learning specialist or academic support office about a broader study plan.
- Don't skip the "one next step" prescription. Reading the breakdown without acting is the most common failure mode. The Play is the analysis; your follow-through is the workout.