Walk us through how you would scope an analysis of crime trends in a region you know nothing about.

They want to see whether you start with the data or with the decision, and whether you understand how patchy Australian crime data can be. A step-by-step walk-through: clarify the question, audit the data, do exploratory analysis, validate, then communicate. Name the real constraints at each step.

You have administrative data with a lot of missing values on the variables that matter most. What do you do?

Missing data is the norm in justice datasets. They are testing judgement, not textbook recall. Diagnose, then decide, then document. Explain the trade-off you are choosing and why.

Tell me about a time your research findings were inconvenient for a stakeholder.

Criminologists regularly produce results that contradict a program's stated success. They want to know you can hold your ground without burning the relationship. STAR. Keep the situation brief and spend most of the answer on what you actually did and what changed.

How would you handle a request from a senior official to present a finding more strongly than the evidence supports?

This is the pressure test. It comes up in policy shops and it is usually asked as a hypothetical. Judgement under pressure: acknowledge the underlying need, restate what the evidence supports, offer an alternative that meets the need, and name the escalation path if it continues.

You are presenting a complex evaluation to a community reference group with no research background. How do you approach it?

Criminologists work with police, corrections staff, councils and community groups. The panel is checking that you can communicate without condescending. Structure the answer around the audience first, then the content choices, then the interaction and how you handle challenge.

What is your approach to data privacy and ethics when working with justice records?

Linked justice, health and corrections data is powerful and tightly governed. They need to know you take the obligations seriously. A short principles statement, then a concrete example of how you have applied it, then how it shapes project design.

Home Criminologist Interview questions

Criminologist interview questions

The questions people actually get asked for this role, and what each one is really testing.

Download PDF Criminologist resume Criminologist cover letter

What to expect

Criminologist interviews test two things at once: whether you can do the analysis, and whether you can explain it to people who will never read your methods section. Panels are usually mixed, with a research or policy manager, a senior analyst, and sometimes a community or operational stakeholder.

Technical and analytical: Questions about statistical methods, data quality, study design and the tools you use. Panels want to know what you would actually do with a messy administrative dataset.
Research design and process: Walk-through questions about scoping a study, choosing a sample, getting ethics approval and managing a project to deadline.
Behavioural: Past-experience questions about influencing a decision, working with resistant stakeholders, or handling a finding that was inconvenient for someone.
Scenario and judgement: Hypotheticals about conflicting evidence, political pressure, media interest or a stakeholder asking you to overstate what the data shows.
Communication and stakeholder: How you translate findings for policymakers, police, corrections staff or community groups, including in writing.
Ethics and data governance: Privacy, consent, data sharing, Indigenous data sovereignty and working with sensitive justice records.

Most Australian government and university processes run two stages. The first is usually a panel interview of three or four people covering your research background and technical skills, sometimes with a short written or data exercise sent beforehand. The second, for shortlisted candidates, is often a deeper conversation with the hiring manager or a presentation of a piece of your work to the team. Academic and research roles may also ask for a seminar. Reference checks and a National Police Check typically follow an offer.

1
Walk us through how you would scope an analysis of crime trends in a region you know nothing about.
Why they ask: They want to see whether you start with the data or with the decision, and whether you understand how patchy Australian crime data can be.
How to structure your answer: A step-by-step walk-through: clarify the question, audit the data, do exploratory analysis, validate, then communicate. Name the real constraints at each step.
Example answer
“I start with the decision, not the dataset. First I would meet the person who asked for this and find out what they are going to do with the answer, because a briefing for a police commander about where to deploy this month is a different job from a three-year prevention strategy. Then I would map what data exists: state police recorded crime data, ABS recorded crime series, court and corrections administrative data, and any local council or health data with a geography I can match. I would check the data dictionary and the collection rules before touching the numbers, because changes in reporting practice or a new online reporting portal can look exactly like a crime wave. After that I would run basic exploratory analysis in R, look at rates rather than counts so population shifts do not mislead, and map the results in ArcGIS to see whether any pattern holds up spatially. Before I write anything I would test the finding against a second data source and put the caveats in the first paragraph rather than the appendix.”
2
You have administrative data with a lot of missing values on the variables that matter most. What do you do?
Why they ask: Missing data is the norm in justice datasets. They are testing judgement, not textbook recall.
How to structure your answer: Diagnose, then decide, then document. Explain the trade-off you are choosing and why.
Example answer
“First I would work out whether the missingness is random or structured, because those lead to different answers. If missingness is concentrated in particular courts, regions or time periods, dropping those cases could bias the whole analysis. I would report the extent of it to the project lead early rather than at the end. Depending on the pattern I might restrict the analysis to a subset where the variables are reliable, use multiple imputation where the assumptions hold, or reframe the question so it can be answered with what we have. What I would not do is quietly run the model on complete cases and present the result as if it covered everyone. If the missingness is severe enough to limit the conclusions, that goes in the write-up, and I would flag it to whoever is going to act on the findings so they know the limits.”
3
Tell me about a time your research findings were inconvenient for a stakeholder.
Why they ask: Criminologists regularly produce results that contradict a program's stated success. They want to know you can hold your ground without burning the relationship.
How to structure your answer: STAR. Keep the situation brief and spend most of the answer on what you actually did and what changed.
Example answer
“A few years ago I evaluated a supervision program that the agency was keen to expand. The early results suggested it was working, but once I split the sample by whether participants had stable accommodation, the effect disappeared for the group without it. The program manager had already briefed upward on the positive numbers. I asked for a meeting before the report went anywhere and showed him the breakdown myself, along with the option of reporting it as a program that works for one group and needs a housing referral for the other. He was not happy initially, but he took it to his executive as an implementation problem rather than a failure, which was a much better outcome than the expansion going ahead and the results collapsing a year later. The final report named both findings.”
4
How would you handle a request from a senior official to present a finding more strongly than the evidence supports?
Why they ask: This is the pressure test. It comes up in policy shops and it is usually asked as a hypothetical.
How to structure your answer: Judgement under pressure: acknowledge the underlying need, restate what the evidence supports, offer an alternative that meets the need, and name the escalation path if it continues.
Example answer
“I would try to understand what they actually need, because the request usually comes from a deadline or a decision that has already been made. Then I would be clear about what the data can support and what it cannot, and offer wording that is confident but accurate, for example that the evidence points in this direction and further analysis would confirm it. I would rather give them a sentence they can defend in a Senate estimates hearing than a stronger one that falls apart under scrutiny. If the pressure continued after that conversation, I would raise it with my manager and put my position in writing, because the report carries my name and the organisation's. I have never had to go further than that, but I would not sign off on something the data does not show.”
5
You are presenting a complex evaluation to a community reference group with no research background. How do you approach it?
Why they ask: Criminologists work with police, corrections staff, councils and community groups. The panel is checking that you can communicate without condescending.
How to structure your answer: Structure the answer around the audience first, then the content choices, then the interaction and how you handle challenge.
Example answer
“I would start by cutting the content down to the two or three things the group can act on or challenge, and put everything else in a one-page handout for anyone who wants it. I would lead with what we found and what it means for their area, then explain how we got there in plain terms, using maps or simple charts rather than regression tables. I would avoid describing the community as a dataset. I would also leave real time for questions, because people in these groups often hold knowledge that explains a result better than any variable in the model, and I have changed my interpretation of findings after those conversations more than once. If someone disputes the numbers, I would take it seriously, explain the source and the limits, and follow up in writing afterwards.”
6
What is your approach to data privacy and ethics when working with justice records?
Why they ask: Linked justice, health and corrections data is powerful and tightly governed. They need to know you take the obligations seriously.
How to structure your answer: A short principles statement, then a concrete example of how you have applied it, then how it shapes project design.
Example answer
“My starting position is that people in these records did not consent to being studied, so the burden is on me to justify what I am accessing and to use the minimum needed to answer the question. In practice that means going through a human research ethics committee even where it is not strictly required, putting a data-sharing agreement in place before any file moves, keeping identifiable data on secure systems with access limited to the people who need it, and reporting only in aggregate where the numbers allow. I have worked with linked justice and health records, and I spent more time than I expected on governance before the analysis started. On projects involving Aboriginal and Torres Strait Islander communities I would follow the relevant data sovereignty principles and make sure the community has a say in how findings are used, not just how they are collected.”