What Is a Pilot Study?
A pilot study in research is a small-scale trial run of your study, carried out before the main data collection begins. Think of it as a rehearsal: you test your questionnaire, your interview guide, your procedures, and your analysis plan on a small group of participants so you can find — and fix — the problems while they are still cheap to fix.
Every experienced researcher has a story about the question nobody understood, the survey link that broke on phones, or the interview that ran twice as long as planned. A pilot study is how you make sure those stories happen in the rehearsal, not on opening night.
Why Run a Pilot Study?
A pilot study answers practical questions that no amount of planning on paper can settle:
- Do participants understand your questions? Wording that seems obvious to you may confuse the people answering it.
- Does your instrument work? A questionnaire might have a broken skip pattern, a scale might produce almost no variation, or an interview guide might miss the topic entirely.
- How long does everything take? If your survey takes 45 minutes instead of the promised 15, your response rate in the main study will collapse.
- Can you actually recruit these participants? The pilot tests your recruitment channels, your consent process, and your inclusion criteria.
- Does your analysis plan work? Running your planned analysis on pilot data reveals coding problems, missing variables, and formatting errors before the real dataset arrives.
In short, a pilot study protects the main study. The few weeks it costs can save months of rework — or a dataset that cannot answer your research question at all.
Pilot Study vs Feasibility Study: What’s the Difference?
Students often use these two terms as if they mean the same thing. They are related, but they are not identical.
| Feature | Pilot Study | Feasibility Study |
|---|---|---|
| Main question | Does this specific study design work as planned? | Can this study be done at all? |
| Focus | The procedures, instruments, and logistics of one planned study | Broader questions: recruitment rates, resources, willingness to participate, funding |
| Typical scale | A small version of the main study, run the same way | May not resemble the main study’s design at all |
| Example | Testing your 30-item questionnaire with 20 students before surveying 300 | Checking whether enough clinics will take part before planning a trial |
In practice, a pilot study is often one part of a larger feasibility assessment. The distinction matters most when you write it up: examiners and reviewers expect you to use the right label for what you actually did.
How to Conduct a Pilot Study in 6 Steps
Step 1: Decide what you are testing
List exactly what the pilot needs to answer. Is it the wording of your questionnaire? The timing of your interviews? The recruitment process? The analysis? Write these as questions before you start, so the pilot has a clear purpose instead of becoming a vague “try it out”.
Step 2: Choose your pilot participants
Pick people who resemble your target population. Piloting a student questionnaire on your classmates is fine if your main study also targets students; piloting it on your family is not. As a rule of thumb, many researchers use around 10 to 30 participants, or roughly 10% of the planned main-study sample. A commonly cited suggestion is 12 participants per group (Julious, 2005) — enough to spot problems, small enough to stay cheap. The exact number matters less than the match: the pilot group should look like the real group.
Step 3: Run it exactly like the real study
Use the same invitation, the same consent form, the same instructions, and the same instrument. If you shortcut the process, you will not discover the problems hidden in the parts you skipped.
Step 4: Collect feedback, not just data
Ask pilot participants what confused them, what felt too long, and what they almost answered differently. A short debrief — even three open questions at the end — is often the most valuable part of the whole pilot.
Step 5: Analyse the pilot data as planned
Run your planned analysis on the pilot data. Check that your variables are coded correctly, your scoring works, and your software produces the output you expect. If you planned reliability checks, run them — but remember that reliability estimates from a tiny sample are unstable, so treat them as a warning system, not a final verdict.
Step 6: Decide what changes — and record it
Write down every change you made because of the pilot and why. This becomes a paragraph in your methodology chapter and a ready-made answer for the examiner who asks, “How did you validate your instrument?”
A Worked Example: Piloting a Questionnaire
Imagine you are studying study habits among undergraduates. Your questionnaire has 28 items, and you plan to survey 300 students. Before that, you pilot it with 20 students from the same university.
The pilot reveals three problems in one afternoon:
- Two items are misunderstood. The item “I review my notes regularly” means “daily” to some students and “before exams” to others. You rewrite it as “I review my lecture notes at least three times a week.”
- The survey takes 22 minutes, not 10. You cut six redundant items and shorten the instructions, bringing it down to 12 minutes.
- One response scale is broken. A five-point scale on the motivation items has no neutral middle option, forcing uncertain students to pick a side. You fix the scale before the main study.
Without the pilot, those three problems would have contaminated all 300 responses. With it, they cost one afternoon and 20 volunteers.
What a Pilot Study Cannot Do
A pilot study is a rehearsal, not a miniature version of the findings. It cannot:
- Test your hypotheses. The sample is far too small, and the pilot was never designed to detect effects.
- Estimate your effect size. Effect sizes from tiny samples bounce around wildly; using them to plan your main sample size can seriously mislead you.
- Prove your instrument is valid. A pilot can reveal obvious problems, but real validity evidence comes from the full study and from the wider literature.
- Stand in for the main study. Pilot results are preliminary by definition. Report them as lessons learned, not as findings.
Common Mistakes to Avoid
- Skipping the pilot because of time pressure. This is the most common — and most expensive — mistake. The time a pilot costs is almost always less than the time a broken main study costs.
- Piloting on the wrong people. Testing a teacher questionnaire on fellow researchers tells you little about how teachers will read it.
- Treating pilot data as real findings. With 15 or 20 participants, a “significant” result is usually noise. Keep pilot results out of your findings chapter.
- Changing nothing after the pilot. If your pilot finds no problems at all, you probably did not look hard enough. A pilot that changes nothing was a formality, not a test.
- Forgetting to document it. Examiners expect a short account of the pilot in your methodology chapter: who took part, what you found, and what you changed.
Do You Always Need a Pilot Study?
Almost always, yes — especially when you have designed your own questionnaire, interview guide, or intervention. The exceptions are narrow: if you are using a well-established, published instrument exactly as its authors intended, with a population very similar to the ones it was validated on, a full pilot may be less essential — though even then, a quick check of timing and instructions is wise.
For theses and dissertations, the answer is simpler: your examiner will almost certainly ask whether you piloted your instrument. Having a clear, honest answer is worth the effort on its own.
Frequently Asked Questions
How many participants do I need for a pilot study?
There is no single magic number. A common rule of thumb is 10 to 30 participants, or about 10% of your planned main-study sample. Julious (2005) suggested 12 per group as a practical minimum for estimating averages and variation. Match the pilot group to your target population — that matters more than the exact count.
Can I include pilot participants in my main study?
Usually not. If the pilot led you to change the instrument or the procedure, the pilot participants experienced a different study from the main participants, so their data should not be mixed in. If nothing changed at all, some researchers include them — but check your institution’s rules and say so clearly in your write-up.
Is a pilot study the same as a pre-test?
Not quite. A pre-test usually means trying out one instrument — for example, checking whether questionnaire items are clear. A pilot study is broader: it tests the whole study process, including recruitment, consent, timing, data collection, and analysis. A pre-test can be one part of a pilot study.
How do I write up a pilot study in my thesis?
A short subsection in your methodology chapter is enough: who took part, how many, what you tested, what problems you found, and what you changed as a result. One or two paragraphs, plus a sentence in your limitations if anything remained unresolved.
Does a pilot study need ethical approval?
Yes, in most institutions. Pilot participants are still human participants: they need the same information sheet, consent process, and data protection as main-study participants. Get approval before you pilot, not after.
Key Takeaways
- A pilot study is a small-scale rehearsal of your full study, run to find problems while they are still cheap to fix.
- It tests your instrument, your procedures, your timing, your recruitment, and your analysis plan — not your hypotheses.
- Ten to 30 participants who resemble your target population are usually enough.
- Always collect feedback from pilot participants, analyse the data as planned, and record every change you make.
- Never report pilot results as findings — and never skip the pilot because you are short on time. That is exactly when you need it most.