Specialty-training selection interviews — how they really work, and how to prepare

First: which interview are you actually facing?

People lose time preparing for “the interview” as if it were one thing. It isn’t. Australian registrar selection is fragmented by design, and the building blocks are mixed and re-weighted differently for every specialty, in every state, every year. Before anything else, pin down your process:

Preparing for medical-school admission interviews (UCAT/GAMSAT-era MMIs) instead? That’s a different process with different rules — this page is for doctors competing for training jobs.

How selection actually works in Australia

The most useful mental model is that two things happen — sometimes as one step, often as two:

• Selection — are you suitable, and how do you rank against other applicants?
• Allocation / employment — which specific job or site do you actually get? Colleges mostly don’t employ you; a health service does, so you often must satisfy both a college and an employer, and many processes end in a preference-based match, not a simple offer.

Accredited vs unaccredited — know which job you’re chasing

State merit scaffolding

Every state runs medical recruitment under a government merit-selection framework that mandates structured, criterion-based, bias-aware panels — not free-flowing chats. In NSW, for example, policy directive PD2023_024 requires the same assessment method and a common set of questions for every applicant, a minimum two-person panel, and structured referee checks. How the layer is organised differs sharply by jurisdiction:

Snapshot of the current cycle — states re-publish campaign rules, dates and criteria annually. Confirm against the live campaign for your intake.

The five building blocks — and how to prepare for each

Almost every Australian registrar process is assembled from five components. Which ones apply, and how each is weighted, is specialty-and-year-specific — but the way you prepare for each is stable.

What the evidence really says — and doesn’t

You asked good questions about the research, so here’s the honest read, in three parts:

1 · The general selection science is solid — but it was just revised. The famous Schmidt & Hunter (1998) figures — structured interviews and cognitive ability both around .51 — are out of date. Sackett and colleagues (2022) showed decades of meta-analyses systematically over-corrected for range restriction; the corrected numbers are lower across the board (structured interviews ~.42, cognitive ability ~.31) Sackett 2022Schmidt & Hunter 1998. The durable takeaway that survives all the debate is the structured-vs-unstructured gap: rubric-scored, criterion-anchored interviews clearly beat free-flowing chats. So don’t trust the naive .51 — but do trust that structure matters.

2 · The medicine-specific evidence for MMIs and SJTs is real, moderate, and thin. SJTs predict at about r≈.32 in medicine, stronger in postgraduate settings Webster 2020. The one large Australian postgraduate predictive study (GP registrars) found the MMI and SJT both predicted end-of-training performance and complemented each other — but it’s one college, one 2010–11 cohort, with a disclosed commercial conflict, and it predates the tools now in use Patterson 2016. Crucially, SJTs carry the largest adverse-impact gaps of any component, with documented differential attainment for international medical graduates Tiffin 2024.

3 · Australian registrar interviews specifically are lightly studied — and the process is admittedly imperfect. Most of what anyone can state confidently is documentary (what the rules say), not empirical (whether it picks better doctors). Only a couple of genuine Australian predictive-validity studies exist Post 2024Patterson 2016. When surveyed, physician-training directors described selection as heterogeneous, interview-heavy, subjective and opaque — literally calling it “an inexact process” Grosse & Thomas 2024. So the honest posture is: we can tell you the format to expect and why these formats are used — we can’t promise any given college’s process is a validated meritocracy, and every figure here is a snapshot of one intake year.

The STAR method — how to answer a behavioural question

Behavioural questions (“tell me about a time…”) are the backbone of structured panels and many MMI stations, and past-behaviour questions out-predict hypotheticals Taylor & Small 2002. STAR is the practitioner framework for structuring a clear answer (STAR itself is a heuristic, not a trialled intervention — what the evidence backs is the behavioural question format).

The question types — and how to attack each

Best-practice tactics (below) rather than trial data. Map your answers to the college’s published selection domains.

• Behavioural (“tell me about a time…”) — answer in STAR with a real example; keep Action first-person.
• Situational (“what would you do if…”) — state your principle, walk through your steps, name when you’d escalate or seek help.
• Motivation (“why this specialty / why here”) — specific, honest reasons tied to what you’ve actually done; avoid generic passion.
• Clinical governance & safety — quality improvement, audit, incident reporting, open disclosure; have one QI/audit story.
• Ethics & professionalism — structure it: gather facts, weigh options/stakeholders, act, seek senior/ethics input; show you know your limits.
• Teamwork, leadership & error — a time you led, followed, or handled a mistake or complaint honestly — reflection matters more than a flawless story.
• CV / portfolio — know every line; be ready to give the STAR behind an achievement, not just recite it.

Prepare like it’s an exam

The evidence says the highest-yield prep is mock interviews, studying the selection content, and organising your answers — coaching and preparation raised structured-interview scores in controlled studies, including a randomised trial, mainly by helping candidates match answers to what’s marked (without undermining validity) Tross & Maurer 2008Maurer 2001. A workable loop:

Managing interview nerves

Interview anxiety is real and it costs marks — more anxious candidates get lower interview ratings McCarthy & Goffin 2004Powell 2018. Two evidence-informed moves:

• Reframe, don’t suppress. In experiments, people who told themselves “I’m excited” performed better than those trying to calm down — the arousal is the same, the label helps Brooks 2014.
• Rehearse under realistic conditions. A single peer-led mock run significantly reduced medical students’ pre-assessment anxiety in a randomised trial Lima 2026 — familiarity is the antidote, and it doubles as answer practice.

On the day — and the mistakes to avoid

Best-practice tactics:

• Signpost your structure — “there are two things I’d prioritise…”. It helps the assessor mark you and keeps you on track.
• Answer the question asked — a brilliant answer to a different question scores nothing on a rubric.
• Be specific and first-person — “I did” not “we did”; the panel is scoring you.
• Treat each MMI station as a fresh start — a weak station is recoverable; don’t carry it into the next.
• Don’t wing motivation questions — “why this specialty” with a generic answer is a common, avoidable stumble.
• Reflect, don’t defend — for error/complaint questions, insight beats a spotless record.

FAQ