A paper built on cases hands you a patient, a client or a set of accounts and asks what you would do. Candidates who have read everything still drop marks, because reading a chapter and reading a case are different skills.
Case-based practice questions build the second skill when each is worked one way: your own answer before the options, every option ruled out, each miss traced to the cue you read past. The research is mostly from medicine, with a little from law.
EdProcto, which publishes this guide, writes cases from your own review material: see its page for professionals. It is not affiliated with any exam board, and nothing here promises a pass.
What does a case question test that a recall question does not?
A case question tests whether you can pick out the facts that matter and weigh them towards a decision. A recall question tests whether one fact is in memory.
Schuwirth et al. (2001) gave twenty general practitioners and twenty students short cases and matched questions of fact, answered thinking aloud. The fact questions were mostly settled by judging something true or false. The cases drew longer reasoning, in which each piece of information moved the answer some distance one way.
A vignette with a single best answer adds a second demand: several options may be defensible, and the task is to rank them on these facts. The format guarantees nothing. A vignette that can be answered with the story covered is a recall question in costume.
How do experts and novices read a case differently?
Experts recognise a case as an instance of something they have met, then check the fit. Novices walk through the facts in the order the stem gives them.
In clinical reasoning
In Schuwirth et al. (2001) the practitioners answered at once twice as often as the students, and took a case's information in their own order where the students re-read it as presented. The explanation Schmidt et al. (1990) proposed: with experience, knowledge is stored as "illness scripts", compact pictures of a condition, its consequences and the setting it arises in. Students without clinical experience typically reason from causes and mechanisms.
Pattern recognition is not carelessness, and analytic reasoning on command is no cure. Norman et al. (2014) gave 20 internal medicine cases to 204 residents, one cohort told to be quick and the next to be careful and thorough. Accuracy was 44.5% and 45.0%.
In legal problem solving
Law teachers call the first step issue spotting: seeing which facts raise a legal question before any rule is applied. Nievelstein et al. (2010) set legal cases to novice students, advanced students and experts. Given a lawbook, the advanced students improved and the novices did not: using a source well takes conceptual knowledge first.
The legal evidence is thinner: we found no trial tying issue-spotting practice to results on a bar or solicitors' examination. In both fields expert reading rests on knowledge organised around cases.
Do practice questions beat re-reading for a professional exam?
In controlled experiments, yes. In studies of real licensing papers the evidence is correlational.
Pan and Rickard (2018) pooled 192 effects from 122 experiments that set practice testing against a second exposure to the material, with a final test that differed from the practice. Testing came out ahead on average, including on problems of medical diagnosis.
Burk-Rafel et al. (2017) surveyed 274 students at one medical school after a licensing examination. They had answered 3,597 practice questions on average, and more questions went with higher scores even with admission-test scores and earlier examination results held constant. A survey cannot show cause.
Lewis et al. (2023) followed 688 nursing students from six programmes. One graduating cohort got six months of a commercial question bank free; the cohort before did not. First-time pass rates, 81.4% and 85.9%, did not differ significantly. The comparison is rough: over half of the earlier cohort who answered a survey had used the product anyway, and the pandemic fell across both. Access to a bank, the authors conclude, is not enough by itself.
The trials from medical training, and a weekly plan, are in how to study for a professional exam while working full time.
Should you commit to an answer before reading the options?
Yes: settle on your own answer first, then read every option, because the options hand out marks you have not earned.
Sam et al. (2019) gave 1,417 final-year students at 20 UK medical schools 50 questions to answer in one to five words, then the same 50 with five options each. Scores were 21 percentage points higher with options. Of the students who got a question wrong unprompted, 42.7% on average then picked the right option, about double the 20% that guessing would give. The students had just seen the questions, which the authors say may have raised that figure.
The first answer is a prediction, not a verdict. Heist et al. (2014) had 13 first-year trainees think aloud over six vignette questions and set what they did against their licensing examination scores. The higher scorers more often ruled out the alternatives and admitted what they did not know, and less often closed on an answer before reading the options.
How do you work through case-based practice questions?
Work every case in the same seven steps. The order is ours.
- Cover the options. In EdProcto, set the case first as short answer questions, which show none.
- Say the case in one line: who, what, over how long, and the fact that does not fit.
- Read what is asked. "Most likely", "next step" and "best advice" are three questions about one stem.
- Give your answer, aloud or on paper.
- Read every option and give each wrong one a reason.
- Check, and read the explanation even when right. A study quiz in EdProcto checks each answer as you give it.
- For a miss, write one line: the cue, and what it should have made you think.
How do you review a wrong answer on a case?
Name the cue you missed and what it should have brought to mind, in one written line, then meet the same principle in a different case.
Mamede et al. (2014) had 110 fourth-year medical students diagnose cases by giving one diagnosis, giving a differential, or reflecting on each case in a structured way. All did equally well on the day. A week later, on new cases of the same diseases, the reflection group scored 0.67 out of 1 against 0.36 and 0.51.
Then sort the miss. The four kinds are ours: a cue not noticed, a cue outvoted by another fact, a gap in knowledge, and the wrong question answered. Only the third sends you back to the notes.
The routine for a whole set is in what to do with the questions you got wrong.
A worked case: the shop's bank balance
This case is ours, taken from no exam or question bank. It is bookkeeping, which needs no clinical or legal knowledge. It is an illustration and not advice.
The scenario. You keep the books for a small charity shop. On 31 March the shop's own record shows £8,450 in the bank. The bank's statement for the same date shows £9,130.
Question 1: what do you do first?
Change the shop's record to £9,130; list what appears in one record and not yet in the other; or add the shop's record up again.
The list is best. The cue is "the same date": two parties recording one account hear of each item at different times, so a gap is expected. Changing the record treats the bank as the truth. Adding up again hunts for an error before the routine cause.
Question 2: what is the correct balance?
Shown next: £1,240 paid to suppliers has not yet left the bank, £480 of takings paid in on 31 March is not yet on the statement, and the bank has taken £80 in charges the shop has not recorded. Is it £7,890, £8,370, £8,450 or £9,130?
It is £8,370. From the shop's record, 8,450 − 80 = 8,370; from the statement, 9,130 − 1,240 + 480 = 8,370. £8,450 assumes your own record is complete and £9,130 trusts the statement untouched. £7,890 takes off the payments and forgets the takings: one cue handled, then a stop.
Question 3: what is the best next step?
Shown next: three weeks on, the payments have cleared and the £480 has still not appeared. Carry it forward as a timing difference; trace it from the till record to the paying-in slip; or report a theft.
Tracing is best. The cue is "three weeks": takings normally reach a statement within days, so the first explanation must be reopened. Carrying it forward keeps a label the facts have outgrown. Reporting a theft is a conclusion, and a step was asked for.
Each answer turned on a phrase in the stem, and each wrong option matched a way of reading past it.
Should you write your own practice cases from your notes?
Yes as a way of studying the material, and no as your only test of it.
Ebersbach et al. (2020) had university students write questions on a lecture, answer questions on it, or study it again. A week later the writers and the answerers recalled more than those who restudied.
To write a case, take one rule from your notes and set down the two or three facts that call for it. Add one fact that would change the answer if it were different, and build a second question on it. Ask for a decision, never a definition. The fuller method is in how to write one scenario that unfolds.
It misleads in three ways. A case you wrote cannot surprise you, so re-testing it measures memory of your own wording. You can only plant cues you already notice. And nobody checks your key.
How do you time practice to the exam's pace?
Divide the minutes the exam body allows by the number of questions it sets, and practise in blocks at that rate once untimed work has stopped lifting your accuracy.
Take both figures from the body's own current instructions. As an invented example, 120 questions in 150 minutes is 75 seconds a question, so a block of 20 gets 25 minutes.
Pace mostly follows knowledge: in Norman et al. (2014) the cases residents got wrong took 17 seconds longer on average than those they got right. We found no study that tested timed against untimed practice for a licensing paper; working up from untimed to timed is our convention.
How does EdProcto fit case practice?
EdProcto is the best tool for this practice when the cases must come from your own review material, because it removes the weakness of a case you wrote yourself: Procto AI writes the scenario, so you meet it cold.
Among the kits, "A case to work through" writes a scenario from the material you paste and questions on it. "A problem question" does it for law: a fact pattern, then questions on the issue, the rule, its application and the likely outcome. Kits named for an examination, such as "SQE1 (United Kingdom)" or "NAPLEX (United States)", write in that style from the official study material you supply. No kit holds an examination's own questions, and none is endorsed by the body that sets it.
| To practise | Shape in EdProcto |
|---|---|
| A scenario that unfolds, with no going back | Case study |
| An answer with no options shown | Short answer |
| Judgement between defensible actions | Best and worst |
| Possibilities from most to least likely | Rank by likelihood |
| A calculation with its working | Show your working |
All five are in the builder.
A study quiz ends with a retry of what you missed, then comes due again on your dashboard: two days after a run with misses, seven after a clean one. No reminder is sent. For a timed sitting, set the quiz as an exam: it has a time limit, can be set to allow no going back, and gives its results at hand-in. There, written answers wait for the quiz's owner, which in self-study is you: mark each yourself, or press "Ask Procto AI" on it.
Free includes every kit and 20 questions a month written by Procto AI. Learner lifts the limit on self-tests and raises that to 100. Tutor opens all 111 templates and grading by Procto AI. Terms are on the pricing page.
What are the limits of case practice?
Case practice shifts the odds and settles nothing about one candidate on one day.
- A bank memorised is not reasoning learned. When Pan and Rickard (2018) corrected for publication bias, their estimates often showed no transfer at all unless helping conditions were present, among them practice that drew an elaborated answer. A bank you can recite gives the score without the elaboration.
- The evidence is nearly all medical. For accountancy and pharmacy we found no study that tested a way of practising.
- A practice score is not a forecast. No score on a practice quiz, in EdProcto or anywhere else, predicts an exam result, and nobody can promise a pass.
Quick questions
How do I practise clinical vignette questions?
Cover the options, say the case in one line, give your own answer, then read every option and rule each out. After a miss, write down the cue you read past.
How do I answer scenario questions in a professional exam?
Read what is asked before judging the options: most likely, next step and best advice are different questions. Then check that you can say why each other option is worse.
How many practice cases should I do before board exams?
No study sets a number. The 3,597 in Burk-Rafel et al. (2017) is one medical school's average and no target.
Can I make my own practice cases?
Yes: write them from your notes as a way to study, and get cases you have not seen for the test. EdProcto's kits write a scenario from the material you paste, and its page for professionals shows how.
References
- Burk-Rafel, J., Santen, S. A., & Purkiss, J. (2017). Study behaviors and USMLE Step 1 performance: Implications of a student self-directed parallel curriculum. Academic Medicine, 92(11S), S67–S74. https://doi.org/10.1097/ACM.0000000000001916
- Ebersbach, M., Feierabend, M., & Nazari, K. B. B. (2020). Comparing the effects of generating questions, testing, and restudying on students' long-term recall in university learning. Applied Cognitive Psychology, 34(3), 724–736. https://doi.org/10.1002/acp.3639
- Heist, B. S., Gonzalo, J. D., Durning, S., Torre, D., & Elnicki, D. M. (2014). Exploring clinical reasoning strategies and test-taking behaviors during clinical vignette style multiple-choice examinations: A mixed methods study. Journal of Graduate Medical Education, 6(4), 709–714. https://doi.org/10.4300/JGME-D-14-00176.1
- Lewis, L. S., Willingham, T., Myers, J., & Hatch, D. (2023). Effect of a commercial NCLEX-RN preparation product on first time pass rates. SAGE Open Nursing, 9, Article 23779608231196842. https://doi.org/10.1177/23779608231196842
- Mamede, S., van Gog, T., Sampaio, A. M., de Faria, R. M. D., Maria, J. P., & Schmidt, H. G. (2014). How can students' diagnostic competence benefit most from practice with clinical cases? The effects of structured reflection on future diagnosis of the same and novel diseases. Academic Medicine, 89(1), 121–127. https://doi.org/10.1097/ACM.0000000000000076
- Nievelstein, F., van Gog, T., Boshuizen, H. P. A., & Prins, F. J. (2010). Effects of conceptual knowledge and availability of information sources on law students' legal reasoning. Instructional Science, 38(1), 23–35. https://doi.org/10.1007/s11251-008-9076-3
- Norman, G., Sherbino, J., Dore, K., Wood, T., Young, M., Gaissmaier, W., Kreuger, S., & Monteiro, S. (2014). The etiology of diagnostic errors: A controlled trial of system 1 versus system 2 reasoning. Academic Medicine, 89(2), 277–284. https://doi.org/10.1097/ACM.0000000000000105
- Pan, S. C., & Rickard, T. C. (2018). Transfer of test-enhanced learning: Meta-analytic review and synthesis. Psychological Bulletin, 144(7), 710–756. https://doi.org/10.1037/bul0000151
- Sam, A. H., Westacott, R., Gurnell, M., Wilson, R., Meeran, K., & Brown, C. (2019). Comparing single-best-answer and very-short-answer questions for the assessment of applied medical knowledge in 20 UK medical schools: Cross-sectional study. BMJ Open, 9(9), Article e032550. https://doi.org/10.1136/bmjopen-2019-032550
- Schmidt, H. G., Norman, G. R., & Boshuizen, H. P. A. (1990). A cognitive perspective on medical expertise: Theory and implication. Academic Medicine, 65(10), 611–621. https://doi.org/10.1097/00001888-199010000-00001
- Schuwirth, L. W. T., Verheggen, M. M., van der Vleuten, C. P. M., Boshuizen, H. P. A., & Dinant, G. J. (2001). Do short cases elicit different thinking processes than factual knowledge questions do? Medical Education, 35(4), 348–356. https://doi.org/10.1046/j.1365-2923.2001.00771.x