Questions to Ask a Scientist
Questions for a conversation with a working researcher, covering what they are doing this week, how they judge a result, what never makes it into the paper, funding pressure, peer review and the claims in their field they doubt. For students, writers and anyone curious about how research actually runs.
The questions
Open any question for the note
What are you actually working on this week?
Why ask it
Better than asking what they research, which invites the funding-application summary. The weekly answer is concrete, a sample prep, a stuck analysis, a rewrite, and it gives you something specific to follow.
How would you explain it to a friend who does not work in science?
Why ask it
You will find out fast whether they have ever had to. Scientists who teach or work near clinicians usually have a clean version ready, and the analogy they choose tells you what they think the core of the problem is.
What made you think this question was worth years of your life?
Why ask it
Often the answer is not noble: a supervisor had funding, a technique was available, a result did not make sense. That honesty is more interesting than a story about wanting to help people, and it tells you how research careers really form.
What result would change your mind about your own hypothesis?
Why ask it
The most revealing question you can ask. Someone who can name the specific observation that would sink their idea is doing science, and someone who cannot name any is defending a position.
What part of this work never appears in the paper?
Why ask it
Months of failed conditions, the equipment that broke, the pilot study that went nowhere. Published papers are a tidy narrative built after the fact, and this question is how you get the untidy version.
How often does an experiment work the first time?
Why ask it
Expect a low number, and expect it to depend on the field. It usefully corrects the impression, common among students, that repeated failure means you are bad at this.
What did you get wrong that took the longest to notice?
Why ask it
Ask for the mechanism of the mistake: a contaminated batch, a mislabelled column, an assumption inherited from an old paper. The way they found it is the transferable lesson.
How do you decide a result is real rather than noise?
Why ask it
Listen for whether the answer is only statistical or also practical, replication in a second system, a different method, a blinded analysis. A p-value alone as an answer is worth following up on.
How much of your week is research, and how much is grants, admin and teaching?
Why ask it
The ratio surprises most people outside universities, and for senior researchers bench time can be close to zero. This is also the honest answer to why science moves more slowly than the press suggests.
What happens to this project if the next grant does not come through?
Why ask it
Answers get specific quickly: a technician's contract, a student's project, a freezer full of samples. It shows how much of research continuity depends on funding cycles rather than on the science.
What do you do when a reviewer asks for an experiment you think is unnecessary?
Why ask it
The practical answer is often that they do it anyway, which reveals a lot about the incentives around publishing. Ask how much time the last round of review added to the paper.
What widely repeated claim in your field do you doubt?
Why ask it
Every field has one, and specialists discuss them privately far more than publicly. Ask what evidence would settle it, which keeps the answer from being simply contrarian.
Who else in the world works on this, and do you talk to them?
Why ask it
Small fields can be a handful of groups who either collaborate or compete. Which one it is shapes what gets shared before publication and how fast the field moves.
What can you do now that would have been impossible ten years ago?
Why ask it
Gets you the real technical change rather than a trend: cheaper sequencing, better imaging, computing that used to need a week. It also shows what still cannot be done, which is usually the more interesting half.
What would you need to answer your question that does not exist yet?
Why ask it
A tool, a dataset, a model organism, a way to measure something without disturbing it. Researchers rarely get asked this and often have a very precise answer, because it is the thing blocking them.
What happens when your work gets reported in a way you would not have written?
Why ask it
Most researchers have an example of a headline that overstated their findings, and often they had little control over the press release. Ask what they do about it now, since the answer is usually resigned rather than dramatic.
What ethical question in your field would surprise people outside it?
Why ask it
Rarely the one the public expects. It may be about data reuse, consent for old samples, authorship order or who benefits from a finding, and it is a better question than asking whether the work is ethical in general.
Who does work in your group that does not get credited?
Why ask it
Technicians, animal care staff, the person who maintains the instrument. A researcher who answers this readily is telling you something true about how the lab runs, and the answer often names the person who actually knows how everything works.
What would you tell someone thinking about a doctorate in this field?
Why ask it
Push past the encouragement for a specific caution: the job market, the supervisor mattering more than the topic, the years of low pay. Vague enthusiasm here usually means they have not thought about their own students' prospects.
What part of your work do you expect to still hold in fifty years?
Why ask it
A good closing question because it forces a distinction between what they hope and what they think. Many will say the method rather than the finding, which is itself a useful thing to understand about science.
Interviewing a researcher
Practical guidance for the conversation itself
Before the conversation
Read one abstract, properly
Find their most recent paper and read the abstract and the final paragraph of the discussion, which is usually where the limitations sit. Ten minutes of that changes the whole conversation, because you can ask about their work rather than their field.
Ask for the limitation early
What would you not conclude from this is a question researchers answer generously and journalists rarely ask. It also protects you from repeating a claim more strongly than the data supports.
Agree what is on the record
If you are writing anything up, settle this at the start, and offer to check quotes about technical detail. Unpublished results and colleagues' work are the two areas where scientists most need that assurance.
Ask for a number and a unit
Significantly increased means nothing without the size of the effect. Ask how much, out of how many, compared with what, and keep asking until the sentence has quantities in it.
What frustrates scientists
So when will there be a cure
Most research is several steps removed from any application, and the honest answer is unsatisfying. Ask instead what has to be true for this to matter to someone outside the lab, which they can answer accurately.
Treating uncertainty as weakness
Hedged language is precision, not evasion. If a researcher says the effect holds in one cell line, do not translate that into it works. Ask what would need to happen to say it more strongly.
Asking them to speak for all of science
A geochemist has no special authority on vaccines. Staying inside their expertise gets you better answers and avoids putting them in a position they will regret in print.
Assuming they own their time
Teaching loads, grant deadlines and animal or instrument schedules are inflexible. Offer a short, specific slot rather than an open request for a chat.
Follow-ups that keep an answer moving
When the answer is technical
- 1What would I see if I were standing next to you while you did that?
- 2What is the step that most often fails?
- 3If it worked perfectly, what would you know that you do not know now?
When the answer is a claim
- 1How big was the effect, and in how many samples?
- 2Has anyone outside your group reproduced it?
- 3What is the strongest argument against it?