Practising With a Partner: Making It Useful

Two people sit down to practise interviews for an hour. Ten minutes in, someone tells a story about a scenario they heard about. Twenty minutes in, they are comparing how many universities they have applied to. Fifty minutes in, they agree it was a really useful session and neither of them has said a single word under a clock.

An interview practice partner is genuinely valuable, mainly because a real human gives you the one thing solo work cannot: an unpredictable response and an outside opinion. But the default state of partner practice is conversation, and conversation is not practice. You have to impose a structure on it, deliberately, every time.

Choose the partner carefully

The best partner is not your closest friend, and it is not the strongest candidate you know. It is whoever will tell you the truth and turn up twice a week.

• Someone who can say that was too long without needing to soften it into oblivion first.
• Someone who is reliable. A brilliant partner who cancels is worth less than an average one who shows up.
• Someone who is not competing with you for the same single place in a way that will poison the feedback. Applicants to different universities often work better together.
• A non applicant can be excellent, particularly a parent, partner or colleague with work experience. They cannot judge content, but they can absolutely tell you whether you sounded like someone they would want treating them.

Two partners is better than one, because you get two sets of blind spots instead of one shared set.

A session that works

Sixty minutes, split cleanly, with a timer that someone actually runs. The shape matters more than the total time, and forty five focused minutes beats two undisciplined hours.

• Five minutes to agree what each of you is working on today. One target each, named out loud.
• Two stations each, run properly: prompt read outside the room, timer started, no interruptions, no restarts, no laughing in the middle.
• Three minutes of structured feedback after each station, not ten minutes of discussion.
• One repeat at the end: whoever had the weakest station runs it again with the feedback applied. This is the part everyone skips and it is where the improvement lives.
• Five minutes to write down what each of you is taking away. If it is not written, it did not happen.

Ban the debrief chat until the session is over. Talking about how hard that one was, mid session, is how an hour evaporates.

The feedback problem

That was really good is the most common sentence in partner practice and the least useful. It is not dishonesty, it is politeness, and politeness is the main reason these sessions fail. Fix it with rules rather than willpower:

• Feedback must be behavioural. Not you seemed nervous, but you looked at the ceiling for most of the second minute.
• One thing to keep, one thing to change. Exactly one of each. A list of six changes gets you no changes.
• The person who answered speaks first, for thirty seconds, on what they thought went wrong. It makes the honest version easier to say afterwards.
• No defending. If you find yourself explaining what you meant, you have stopped practising and started arguing. Write it down and move on.
• Feedback on delivery is worth more than feedback on content, because your partner is guessing at content and observing delivery directly.

Mark against something, even roughly

Impressions drift. A rough rubric keeps two people honest and comparable across weeks. Write four or five domains on a page, score each out of five, and hand the sheet over. Our breakdown of how MMI scoring works in Australia explains the usual domains, and since these vary between universities you should check the university's current admissions page for how your school describes its criteria.

You are not trying to replicate real marking, which you cannot do. You are trying to force your partner to notice specific dimensions instead of forming one warm general impression of a person they like.

Playing the other person properly

In roleplay stations your partner has the harder job, and most people do it badly in one of two directions. Too soft: they hand over the crucial information in the first thirty seconds, agree with everything, and the candidate looks brilliant while learning nothing. Too hard: they stonewall, invent absurd complications, and turn the scene into a game the candidate cannot win.

The middle is a simple rule: decide one hidden concern and one emotional temperature before you start, then only release the concern if the candidate asks a question that genuinely opens the door. Soften if they listen well, stay guarded if they lecture. That single rule makes partner roleplay feel remarkably close to the real thing.

Stay in character until the timer stops. Breaking to laugh at the awkward bit is the most common way a partner session dissolves.

Where the prompts come from

Each of you writes prompts the other has not seen, and neither of you looks at their own bank during a session. Surprise is the resource a partner provides, and you waste it by working through a shared list you both read on Sunday. Our overview of what an MMI is in Australia and New Zealand covers the station types worth writing across, so your bank does not end up being eight versions of one ethics question.

Avoid anything sold as real questions from previous years. Those lists are unreliable, universities change their stations, and practising for a specific question is the fastest way to be rattled by a different one.

If you are practising over video

Practising with a partner in another city works fine, and it is the right choice if your interview is online anyway. Keep cameras on, use the real setup rather than lying on a bed, and record the calls so feedback can point at a moment instead of a memory. Our comparison of virtual and in person MMIs covers what the online format changes about delivery.

Know what a partner cannot give you

Two candidates at the same stage share the same blind spots. You will both miss the same weaknesses, reinforce the same habits, and reassure each other that a mediocre answer was fine because it matches what you would have said. That is the ceiling of peer practice, and it is invisible from inside it.

Mix in something that does not know you and has no reason to be kind. MasterMed's live AI interviewer runs timed MMI stations and marks you against a rubric, which is a useful counterweight to a friend who wants you to feel good on a Tuesday night. The first speaking station is free on the trial, there is no card, and the trial does not convert by itself.

Used properly, a partner is still one of the best resources you have. Just hold them to the format: timer on, one target each, honest feedback in behaviours, and the repeat at the end. Do that twice a week for a month and you will both be unrecognisable by interview season.