A dissertation pilot study is a small trial run of your procedures before the main data collection. It tests whether your recruitment, instruments and analysis plan work in practice, so problems surface while they are still cheap to fix.
This guide explains what a pilot can genuinely tell you, how to design one for different methods, how many participants are typical, how to set feasibility criteria and how to report the pilot in your methodology chapter.
What a Pilot Study Is For
A pilot tests feasibility: can you recruit, do participants understand the questions, how long does it take, does the data come out in a usable form and does your analysis plan run? It is not a small version of the main study for testing hypotheses.
Set clear criteria before you start, record what you change and why, and report the pilot briefly in your methodology chapter. Most universities expect ethics approval to cover the pilot as well as the main study.
What a Pilot Can and Cannot Tell You
The most common misunderstanding is treating a pilot as a mini experiment. With a handful of participants, statistical tests have very little power and effect estimates are highly unstable.
| A pilot can tell you | A pilot cannot reliably tell you |
|---|---|
| Whether recruitment and consent procedures work | Whether your hypothesis is supported |
| How long the survey, interview or task takes | A trustworthy effect size for a power calculation |
| Which questions confuse participants | Final reliability of a scale |
| Whether data can be exported, coded and cleaned | Whether the main study will reach significance |
| Whether your analysis plan runs on real data | What the findings of the main study will be |
Methodologists have long warned against using pilot effect sizes to set the main sample size, because a small pilot can easily overestimate or underestimate the true effect. Base your power analysis on published studies or a smallest effect of practical interest instead.
Pilot Study, Pretest or Expert Review?
These terms overlap, and departments use them loosely. Being precise about which one you did helps examiners judge it.
| Activity | What it involves | Typical size |
|---|---|---|
| Expert review | Subject experts judge relevance and coverage of items | Often 3 to 10 experts |
| Cognitive interviewing | Participants think aloud while answering questions | Often 5 to 15 people |
| Pretest | A short trial of an instrument for wording and timing | Small, varies widely |
| Pilot study | A trial of the full procedure, from recruitment to analysis | Depends on purpose; see below |
| Feasibility study | Tests whether a larger study can be done at all | Often larger, common in health research |
How Many Participants a Pilot Needs
There is no single correct number, and published guidance varies. What matters is that the size fits the purpose of the pilot and that you justify it.
- Checking wording and timing of a questionnaire can often be done with a small group drawn from the target population.
- Estimating a standard deviation or recruitment rate needs more. Rules of thumb in the literature range from about 12 per group to 30 or more participants.
- Testing interview guides often uses one to three practice interviews with people similar to your participants.
- Early scale reliability can be explored, but Cronbach's alpha from a small pilot sample is imprecise and should be treated as provisional.
Choose pilot participants who resemble your main sample. Where the population is small, such as a single organization, check whether pilot participants can still take part in the main study; often they cannot.
Setting Feasibility Criteria in Advance
A pilot only helps if you decide in advance what counts as success. Write progression criteria before collecting any data, so you are not tempted to explain away problems afterwards.
| Area | Example criterion | If not met |
|---|---|---|
| Recruitment | At least half of those invited agree to take part | Change the invitation or widen the sampling frame |
| Completion | Most participants finish all sections | Shorten the instrument or reorder sections |
| Timing | Survey takes under 15 minutes | Remove low-value items |
| Comprehension | No item misunderstood by more than one participant | Reword the item and retest |
| Data handling | Data exports cleanly and the analysis script runs | Fix coding and variable names |
The thresholds above are illustrations. Set yours according to your context, your supervisor's advice and what the main study needs.
Designing a Quantitative Pilot
For surveys and experiments, run the whole sequence exactly as the main study will: invitation, consent, instrument, data export and analysis. Shortcuts in the pilot hide the problems you are looking for.
- Recruit from the same population, using the same channel.
- Record time taken and drop-off points.
- Add a few open questions at the end asking which items were unclear.
- Run your planned analysis on the pilot data to check that variables are coded correctly.
- Inspect distributions, missing data and items with almost no variation.
Report descriptive information from the pilot, such as completion rates and timings. Avoid reporting significance tests as if they were findings.
Designing a Qualitative Pilot
In qualitative research a pilot usually means one or more practice interviews or focus groups. The aim is to refine your guide, your prompts and your own interviewing technique.
- Check whether questions produce rich answers or one-word replies.
- Notice where you lead participants or talk too much.
- Test the recording equipment and the transcription process.
- Try an early round of coding to see whether the data address your research questions.
Some qualitative researchers argue that interview guides naturally evolve during data collection, so a formal pilot is less essential. Even so, a practice interview is widely recommended, especially for first-time researchers.
Reporting the Pilot in Your Methodology Chapter
The pilot usually earns a short subsection in Chapter 3, often a few paragraphs. Report what you tested, what you found and what you changed.
Example pilot write-up. The questionnaire was piloted with eight staff nurses from a ward not included in the main study. Median completion time was 13 minutes, within the 15-minute target. Two participants misread item 9, which asked about "shift length" without specifying whether handover time counted; the item was reworded to "hours from start of handover to end of handover".
One item on the workload subscale showed no variation and was retained, since the validated scale does not permit removal. The data exported correctly and the planned regression script ran without errors. Pilot data were not included in the main analysis because item 9 changed.
All figures in this example are invented for illustration. Notice that it reports changes with reasons and states clearly whether pilot data were used.
Should Pilot Data Go into the Main Analysis?
Usually not. If the instrument or procedure changed after the pilot, combining datasets mixes two different measurements.
Some researchers include pilot data when nothing changed and the participants met all inclusion criteria, but this should be decided in advance, approved by your supervisor and stated in your methodology. In qualitative work, a pilot interview is sometimes included if it was conducted with the final guide and full consent; again, say so explicitly.
How DissertationHorse Helps with Pilot Studies
Send your research design, draft instruments, ethics requirements and any notes from your pilot. Your writer drafts a custom pilot protocol or the pilot section of your methodology chapter, with feasibility criteria and a clear record of what changed.
We never invent pilot data or participants. If you have collected pilot data, a writer can help you describe it accurately. Free revisions within the scope of your original request are included.
Want Your Pilot Planned and Written up Properly?
Send your design, instruments and any pilot notes. Your writer drafts a custom pilot plan or write-up that reports feasibility honestly and justifies every change.
Order Your Pilot Write-upFree revisions · Full refund if late · Written from scratch for your order
Frequently Asked Questions
Not always. Many programs recommend one, especially for new instruments or interview guides, but requirements vary. Check your handbook and ask your supervisor.
It depends on the purpose. Testing wording may need only a small group, while estimating variability needs more. Published rules of thumb range from about 12 per group to 30 or more.
Usually yes, if it involves human participants. Most universities expect the pilot to be covered by your ethics application.
Generally not, especially if anything changed after the pilot. If you plan to, agree it with your supervisor in advance and explain it in your methodology.
No. Pilot samples are too small for meaningful hypothesis testing. Use the pilot to test procedures, not outcomes.
It is not recommended, because small pilots give unstable estimates. Use effect sizes from published research or the smallest effect that would matter in practice.
Usually as a subsection of the methodology chapter, with any revised instruments in the appendices.
The terms overlap. A feasibility study asks whether a larger study can be done at all; a pilot is a small rehearsal of the planned study. Health research often treats pilots as a subtype of feasibility study.