MMI Question Bank Guide for Australian Medical Training

How to build a multiple mini-interview question bank for Australian medicine: station formats, scoring domains, and checking each entry against its source.

12 October 2026 · updated 19 October 2026

You’re probably staring at a stack of interview notes, a timer on your phone, and a browser full of random practice questions that all blur together. That’s the point where a proper multiple mini-interview (MMI) question bank — a multiple mini-interview being a selection interview run as a series of short, separately scored stations — stops being a nice extra and starts becoming the thing that keeps your preparation organised, honest and usable.

For Australian medicine, a good bank isn’t a pile of prompts. It’s a working reference that shows what each station is trying to measure, how it is scored, and what the candidate needs to do under time pressure. That matters because the format varies by institution and pathway: a 2025 mapping of all 20 Australian medical programs found extensive variation in admission elements and how they are weighted, with most interviewing programs using an MMI.¹ Deakin University is one worked example. For 2027 entry to its Doctor of Medicine (MD), it interviews applicants at up to six consecutive stations of five minutes each, online in September 2026, and the MMI score makes up 50% of the ranking used for offers.² Deakin’s admission schedule, approved by its Academic Board in February 2023, also sets a minimum overall MMI score of 50%, and applicants scoring less than 2 on any station undergo review.³

Table of Contents

• What an MMI Question Bank Is For
  • Topic list, script bank, working reference
• Anatomy of an Australian MMI
  • What the panel controls and what you control
• Station Formats and Prompt Types at a Glance
• Reading One Annotated Bank Entry
  • A worked example of what should be included
• Scoring Domains and Behavioural Anchors
  • Why anchors matter
• How Assessor Judgements Are Formed
  • What an assessor listens for
• Rural, Aboriginal and Torres Strait Islander, and Cultural Safety Stations
  • What to rehearse instead of a speech
• Choosing Between Generalist and Institution-Specific Material
• Building a Two-Week Rehearsal Cycle
  • A simple rotation you can actually finish
• Verifying Source Material and Version Dates
  • A four-step check
• Quick Reference Index for the Working Bank

What an MMI Question Bank Is For

A candidate usually reaches for a question bank at the wrong moment. They’ve already collected too many prompts, and the core problem is no longer finding questions, it’s deciding which ones are worth rehearsing and why. A useful mmi question bank should answer that problem before it adds another scenario to the pile.

Topic list, script bank, working reference

A topic list gives you headings, like ethics, teamwork or rural health. That helps with breadth, but it doesn’t tell you how the station is run or how it’s scored. A script bank goes too far the other way, because polished answers can sound rehearsed and brittle when the interviewer asks a follow-up.

A working reference sits in the middle. Each entry should tell you which institution or pathway it belongs to, how long the station runs, what type of prompt it uses, which domain is being assessed, what score rule applies, and whether there’s a review condition. That is the shape you need, because the interviewer is not scoring your ability to recite a model paragraph, they are scoring whether your reasoning, structure and behaviour match the station’s aim.

Practical rule: if a practice item doesn’t tell you how it will be judged, it’s not finished enough to train on.

The rest of this guide uses that standard. The best bank entry is not the one with the most dramatic question, it’s the one that makes the scoring logic visible enough for you to rehearse against it.

Anatomy of an Australian MMI

An Australian MMI is best understood as a multi-station system. Deakin’s 2027-entry format shows that clearly: up to six five-minute stations, rolled into a single MMI score that carries half the weight of the offer ranking.² In other words, the interviewers are not judging one long interview impression, they’re collecting structured observations across separate tasks.

What the panel controls and what you control

The institution controls the station design, the rubric, the threshold, and any review trigger. You control the answer behaviour: how quickly you orient to the prompt, how you use the reading time, how you structure the response, and whether you escalate appropriately when the scenario demands it. If you mix those up, you’ll practise the wrong thing.

That’s why the station metadata matters. A bank entry should preserve the source institution, the intake year, and a last-checked date, because a change in timing or weighting can make an older prompt misleading. Station length differs even between programs that both use an MMI: Bond University’s medical program, for example, used one minute of reading and seven minutes of response per station in 2022 and 2023, against Deakin’s five-minute stations for 2027 entry.⁴²

The number of stations doesn’t matter on its own. The timing, threshold and domain weighting are what shape your answer.

A candidate who reads the admissions page well can spot the levers that matter. If the station is short, they need crisp framing. If the rubric uses a threshold, they need to know what counts as unsafe or incomplete. If a score below a set point triggers review, as it does at Deakin, they need to avoid a half-answer that sounds thoughtful but omits the key action.

Station Formats and Prompt Types at a Glance

The quickest way to stop a question bank feeling random is to sort stations by format, not just by topic. Ethical stations ask for judgement. Role-plays ask for communication under pressure. Prioritisation tasks ask what you would do first and why. That’s a very different rehearsal problem from a values station.

A table like this is more useful than a long list of sample questions because it tells you what the station is measuring. If a bank doesn’t let you filter that way, you end up revising the wrong skill for the wrong prompt.

You stop asking, “What’s the answer?” and start asking, “What behaviour would earn marks here?” For extra structure, it helps to keep a separate prep resource open, such as this MMI interview prep guide, while you sort prompts by format.

Reading One Annotated Bank Entry

A strong bank entry reads more like a flight plan than a sample essay. Start with the institution, because that tells you which format and scoring rules apply. Then attach the timing, the prompt type, the domain, the rubric and the review rule. Without that metadata, the question is just a topic.

A worked example of what should be included

Take a rural-context scenario station. The entry should say which pathway it belongs to, how long you get, what you must do, and which domain is being scored. It should then spell out the observable behaviours that earn marks, such as acknowledging limits, escalating appropriately, naming continuity of care, and avoiding vague reassurance.

A good model response is not a polished speech. It’s a model of observable conduct. The best annotation points to things an assessor can hear, such as “I’d clarify the risk first”, “I’d involve the supervisor now”, or “I’d explain the follow-up plan to keep care continuous”. A borderline response often sounds calm but misses the action, for example by offering comfort without escalation or by giving a general ethical statement with no pathway through the problem.

Practical rule: if the model answer can’t be mapped back to a rubric line, it’s too vague to teach from.

That structure lets you audit any resource. If it has a prompt but no task, it’s thin. If it has a task but no rubric, it’s not defensible. If it has a rubric but no intake year or source institution, it’s hard to trust across admission cycles.

Scoring Domains and Behavioural Anchors

MMI scoring works better when it is anchored to behaviour, not impression. Bond University’s medical program publishes one worked example. Its behaviourally anchored rating scale (BARS — a rating scale where each score point is described by specific, observable behaviour) was developed for the school through a three-round modified Delphi process (a structured method of reaching expert consensus over repeated rounds) with 13 experts, and it rates seven attributes: communication skills, self-regulation, effective team member, adaptability, analytical or critical thinking, empathy and cultural capability.⁴ Each station is rated independently by two trained assessors (one for the group station), and from 2023 the scale moved from four points to five, running from “not performed / unsatisfactory” to “above expectations”.⁴

That is one school’s rubric. Other programs name different qualities: Deakin lists communication skills, commitment to rural and regional practice, evidence-based practice, self-directed learning, teamwork, motivation for a career in medicine, commitment to social justice, professionalism, effective use of resources and health promotion as qualities that may be assessed for 2027 entry.²

Why anchors matter

A vague model answer often tells candidates to “be professional” or “show empathy” without saying what that looks like under pressure. Behavioural anchors solve that. They let the assessor distinguish between a response that notices the risk and one that drifts into generic reassurance, and they help two assessors reach similar judgements.

For preparation, the practical lesson is even more direct. Each answer should be rehearsed against a named behaviour. Did you name the risk? Did you sequence the next step? Did you avoid dismissing the other person’s concern?

The signal columns are our own rehearsal shorthand, not Bond’s published descriptors. If you’ve used generic score sheets before, this feels stricter. It is stricter, and that’s useful: the clearer the anchor, the easier it is to improve the answer in a measurable way. A domain score guide can help keep that scoring language consistent.

How Assessor Judgements Are Formed

A station score is not pulled out of thin air. A 2025 observational study of an MMI used for Australian general practice training selection (data collected in 2015 at a national selection centre) describes interviewer judgements as socially constructed: interviewers balanced institutional rules against their own professional judgement, took part in moderation and shared meaning-making with fellow interviewers, and reflected on their own practice.⁵ That means two interviewers can hear the same answer and still reason about it differently if their calibration differs.

That study describes one specialty’s selection process as it ran a decade ago. Whether a given specialty interviews at all now depends on the college and the intake year, and the guide to specialty selection interviews in Australia sets out which specialties interview, which do not, and who runs it for the 2027 intake.

What an assessor listens for

The question most candidates ask is, “What’s the one correct answer?” That’s the wrong frame. The better question is, “What would the assessor need to hear before feeling safe giving a higher mark?” In practice, that usually means you have shown clarifying uncertainty, prioritising safety, acknowledging limits, escalating appropriately and, where relevant, cultural safety.

A rehearsed script often sounds polished but empty because it jumps straight to a conclusion. A stronger answer sounds like thinking under supervision. It shows the listener how you got from the prompt to the action, so the assessor can see the reasoning rather than just hearing a memorised finish.

The interviewer isn’t looking for a single perfect line. They are looking for a defensible chain of judgement.

Many banks fall short here. They give you a sample answer but not the scoring logic behind it. Once you see that, rehearsal changes: you practise the reasoning out loud, then compare your response with what an assessor would need to hear, rather than chasing a polished monologue.

Rural, Aboriginal and Torres Strait Islander, and Cultural Safety Stations

A lot of candidates talk about rural commitment as if it were a personal slogan. That misses the point. Rurality, equity and cultural safety are better treated as specific competencies grounded in local context, referral pathways and the realities of care delivery.

The 2025 mapping of Australian medical school admissions found that 18 of the 20 programs reported a separate pathway for Indigenous applicants and seven a separate pathway for rural applicants, and that all but one school adjusted entry requirements for rural applicants in some way.¹ Within the interview itself, rurality is assessed less often: an earlier study of nine graduate-entry medical schools, summarised in the same review, found only three assessed rurality during the interview, and two included rurally focused MMI stations.¹ Evidence on whether rural applicants score differently at interview is inconsistent across studies, and in one Australian MMI that included rural-specific questions, applicants mostly reported that those questions gave rural applicants an advantage.¹

What to rehearse instead of a speech

Candidates often rehearse a persuasive story about why they “want rural medicine”. Where a station is set in a rural or Aboriginal and Torres Strait Islander health context, a story is not enough: your answer has to deal with referral limitations, continuity of care, telehealth, workforce scarcity, local community governance and culturally safe escalation.

The clearest answers separate cultural safety from cultural knowledge. You do not need to pretend to know everything about a community to show respect. You do need to show that you’d ask, listen, avoid stereotyping, and involve the right people. That applies just as much to Aboriginal and Torres Strait Islander health as it does to remote service design or follow-up after discharge.

Rural commitment should sound like reasoning, not a performance.

One useful way to practise is to answer the same prompt twice. First, as a general ethics problem. Then, as a rural or Aboriginal and Torres Strait Islander health problem with service constraints. If the second answer sounds no different from the first, it’s too generic.

Choosing Between Generalist and Institution-Specific Material

A candidate who uses the wrong material often practises the wrong habit. Generalist question lists can still build repetition and fluency, but they cannot tell you how a particular school or selection body scores. Material built from the institution’s own published format — station count, timing, the qualities it says it assesses, and its thresholds — is narrower but closer to what you will actually face.

Medical school entry and specialty selection are also different stages with different processes. A bank built for one does not transfer cleanly to the other, and for specialty selection there may be no MMI at all.

The useful question is simple. Does the material show why a station scores the way it does, or does it only hand you prompts? A working reference should help you rehearse reasoning, not memorise scripts. Use institution-specific material for alignment, then add generalist prompts only where you need extra repetition.

Building a Two-Week Rehearsal Cycle

A short rehearsal cycle works best when it repeats the same three actions: read, answer, review. Day one is not about collecting more prompts, it’s about forcing structure. If you can run the cycle twice before interview day, you’ll have seen enough format variation to notice your own weak spots.

A simple rotation you can actually finish

Start with timed reading. Cover the prompt, then give yourself the station’s reading window before speaking. That makes you manage the first few seconds properly, which is where many candidates waste time.

Then move to recorded response. Speak once, out loud, with the timer running. Don’t stop to tidy the answer halfway through, because that hides the core problem. A second person, if available, should score it against the same behavioural anchors used in the bank entry.

Finish with review against the rubric. Ask whether the response showed the right domain, whether it addressed the actual task, and whether it reached the threshold behaviour cleanly. If the answer only sounded polished, it probably needs more work.

A workable 14-day pattern looks like this:

• Days 1 to 4: Rotate ethical, communication, and values stations.
• Days 5 to 8: Add role-play, prioritisation, and teamwork.
• Days 9 to 11: Use rural or cultural safety prompts with stricter timing.
• Days 12 to 14: Run a full dry rehearsal with threshold-style scoring and no interruptions.

The point isn’t volume for its own sake. The point is to train the exact habits the assessor is listening for, under the same time pressure the station creates.

Verifying Source Material and Version Dates

A serious bank should be current enough to trust. At minimum, every entry should show the source institution, the intake year, a last-checked date, the station count, timing, weighting, minimum threshold and any review condition. Deakin’s published rules show why: the station count and timing sit on the admissions page for the intake, while the 50% minimum and the station-level review rule sit in a separate admission schedule.²³ An entry that captures only one of those is incomplete.

A four-step check

1. Check the institution name. Make sure the entry matches the actual pathway or school you’re targeting.
2. Check the timing and station count. If either has changed, the old practice material may no longer fit.
3. Check the scoring rule. Look for thresholds, review triggers and any station-specific weighting.
4. Check the intake year. If it’s an earlier year, treat the entry as historical, not current.

If any of those details are missing, pause. A prompt without version control can still be useful for general rehearsal, but it shouldn’t be treated as a current representation of the interview. That habit saves you from practising for a format that no longer exists.

Quick Reference Index for the Working Bank

If you need the fast version, use this index.

• One published attribute set (Bond University): communication skills, self-regulation, effective team member, adaptability, analytical or critical thinking, empathy and cultural capability.
• Six station categories: ethical scenario, role-play, prioritisation, communication or breaking bad news, teamwork or group task, motivation or values.
• Two kinds of material: generalist prompts and institution-specific material.
• Two-week rehearsal cycle: timed reading, recorded response, rubric review, repeated twice.
• Four-step verification routine: institution, timing, scoring rule, intake year.

Glossary

• Pathway: the specific admission or selection route.
• Station: one timed assessment task.
• Threshold: the minimum standard required.
• Review condition: a result, such as a low station score, that triggers a further look at the application.
• Moderation: the process of aligning assessor judgement.

A bank that uses those labels consistently is much easier to revise from. If you can find an entry, read the rubric, check the intake year and rehearse the behaviour quickly, you’ve got a usable reference instead of a dumping ground for questions.

CliniRoo is building preparation material for the Australian medical pathway. Nothing is on sale yet. If you want to be told when there is something to use, CliniRoo takes an email address and does nothing else with it.