Evidence-Based Studying Part 3 - The Habit That Raised Scores 9 Points
I went down a rabbit hole on the research behind how we study, and built the best of it into a 5-step system. This post covers one piece. Watch the full breakdown in this video.
-Simon
The distractor you write is the distractor you spot.
Answering practice questions is the default, and writing them is where the marks are. Student-led question generation means building the whole thing yourself: a full clinical stem, the correct key, and 3 distractors plausible enough to be tempting. In one study, students who wrote as few as 10 of their own questions saw formative scores rise about 9 percentage points and were roughly twice as likely to land a B+. Ten questions.
So why does writing a bad distractor teach you more than answering a good question? Because you can’t build a tempting wrong answer without knowing exactly which feature separates it from the right one. Starting after your next study session, write 2 questions on what you just covered and swap them with a colleague.
Here’s what the evidence says:
In 2025, Sharif and colleagues studied 135 pre-clinical medical students. They compared students who authored their own multiple choice questions against those who did not, on the same formative assessment. They found scores rose from 59.2% to 68.6%, and the proportion earning a B+ went from 22% to 42%. The dose-response is the part worth noticing: the more questions students wrote, the better they scored.
WHAT MAKES A QUESTION WORTH WRITING
• A stem with a real patient in it
• One defensible key, referenced to a source
• Distractors drawn from your own past errors
• No “all of the above”
Your own wrong answers make the best distractors because they’re pre-tested traps, and you already know they work on at least one person. Skip the shortcut here. Generating 100 questions with AI in a minute gives you quantity instead of quality, and 10 questions you dug out of a guideline yourself will beat 100 you skimmed, because the digging is the part that teaches.
Researchers Bottomley and Denny saw the same pattern across 107 undergraduate biochemistry cohorts back in 2011, where authoring and answering questions tracked with better exam performance even after adjusting for prior ability.
Exam tip: write your questions on the topics where two conditions share a presentation, because that’s where a defensible key forces you to name the discriminating feature out loud. Keep a log of which distractor fooled your study partner and why. If you take one thing from this post: the best way to pass an exam is to learn how to write one.
What do attendees say about The Review Course?
"Very informative for both clinical and exam (SOO & SAMP)." - Dr. OO, The Review Course 2025 attendee.
References
Sharif et al. (2025). 135 pre-clinical medical students: write your own MCQs versus not, formative score increased from 59.2% to 68.6% (p=.010) and doubled likelihood of getting a B+, 42% versus 22% (p=.015). https://doi.org/10.2147/AMEP.S549385
Bottomley & Denny (2011). 107 undergraduate biochemistry cohorts (PeerWise) authoring, answering and rating MCQs: modest but consistent correlation with exam performance, prior-ability adjusted (all p<.05; correlational). https://doi.org/10.1002/bmb.20526
Photo by Alissa De Leva on Unsplash
Comments