Skip to content
Uncategorised

How to Run a Dissertation Pilot Study

dissertation pilot study

Run a dissertation pilot study by rehearsing your planned methods with a deliberately small sample, recording problems, and revising procedures before full data collection. Its main purpose is to test whether your research plan works, rather than answer the dissertation question conclusively.

A confusing survey item or unusable recording can undermine weeks of research. Start by deciding which uncertainties your pilot needs to resolve.

Key Takeaways

  • Define the pilot’s purpose before recruiting; a small convenience sample doesn’t automatically make a study a pilot.
  • Rehearse recruitment, consent, data collection and analysis, rather than testing only your questions.
  • Set decision criteria beforehand, document revisions and keep conclusions proportionate to the evidence.

Understand What a Pilot Study Can Test

A pilot study tests the safety, feasibility, acceptability or preliminary efficacy of a design, method, intervention or concept before larger-scale testing. Its small sample is intentional and justified by those aims.

Calling an under-recruited study a pilot afterwards doesn’t repair its design. The defining feature is a planned contribution to subsequent research, including changes to the study design.

Clinical research commonly distinguishes four overlapping types.

Pilot TypeMain QuestionWhat It Examines
SafetyCould the procedure cause harm?Adverse events and tolerability
FeasibilityCan the proposed study work?Recruitment, procedures and usable data
AcceptabilityWill people take part?Burden, delivery and attendance
EfficacyIs there an initial indication of benefit?Preliminary responses requiring stronger testing

For example, in neonatal science, a safety pilot in a clinical setting might test a feeding intervention with well newborns before involving preterm infants, subject to appropriate safeguards. An acceptability pilot might examine whether families can attend sessions or find measurements burdensome.

Feasibility also concerns whether the data collected can address the intended question, given the proposed procedures. Successful recruitment alone doesn’t establish that.

An illustrated scientist in a lab coat observes a glass apparatus on a large wooden lab table.

In laboratory science, related preliminary work is often called proof of concept. For dissertations, the same principle applies to questionnaires, interview guides, equipment, coding frameworks and data-processing procedures.

Plan Your Dissertation Pilot Study

Define the Uncertainty You Need to Resolve

Write a pilot aim that tests a practical uncertainty in your study design, not the dissertation’s substantive research question. Useful targets include whether participants understand eligibility criteria, whether interviews generate relevant material, or whether an instrument records the required measurements.

Your research design should answer the main question; the pilot checks whether its methods can collect suitable evidence. For example, your main study might investigate student experiences, while the pilot tests whether interview questions elicit relevant accounts.

If that connection remains unclear, revisit our guidance on developing a workable research question. A pilot can’t rescue a question that demands evidence your design cannot collect.

Set Criteria for Proceeding or Revising

Decide what would justify proceeding, revising or stopping before seeing the results. Base those decisions on recruitment needs, available time, acceptable participant burden and essential data quality.

Record these criteria in a short pilot protocol. The Queen Mary pilot study resources include methodological material on progression decisions.

Also schedule visible outputs: completed recruitment materials, a tested survey, an exported dataset and a revision log. Allow separate time for permissions, recruitment and retesting rather than treating the pilot as one diary entry.

A researcher in a study room sorting colored cards on a desk next to a large tray.

Choose Participants and Address Ethical Risks

Justify the Sample Against the Pilot Aim

Choose participants who resemble the intended study population in characteristics relevant to the procedure. Testing specialist terminology only with friends outside the field may reveal little about your intended respondents.

Your sample must provide enough information to refine the design without placing people, particularly those from a high-risk population, under unnecessary burden. In a clinical setting, assess whether any participant faces high risk and whether appropriate safeguards are in place. Effect-size estimation may contribute to planning, but small-sample estimates are unstable.

The Duke pilot and feasibility guide provides further methodological context. Justify your sample size against the pilot objective, rather than copying another dissertation’s participant count.

Describe What Taking Part Involves

A pilot involving participants still raises questions about consent, confidentiality, recording, withdrawal and potential distress. Small scale doesn’t remove those responsibilities.

Include the pilot in the relevant ethics documentation and obtain required permissions before participant-facing work. Our guide to applying for dissertation ethics approval explains the supporting documents.

Keep the protocol, information sheet, consent form and research materials consistent. Also explain how you will store pilot data and whether you intend to retain them for later analysis. Substantial revisions may require an amendment before further recruitment.

Rehearse the Complete Data Collection Process

Test Surveys Beyond the Question Wording

Run the questionnaire as participants will encounter it, including the invitation, information sheet, consent process, screening questions and submission page. Test mobile access where relevant.

Ask participants to explain selected questions in their own words and describe how they selected an answer. This reveals interpretation problems that a simple “Was everything clear?” can miss.

Next, inspect routing, completion times, skipped items and exported responses. Our guide to testing dissertation survey questions covers ambiguous wording and response options.

A completed submission is only useful if the resulting variables support your analysis.

Test Interviews, Equipment and Analysis

For interviews, rehearse the opening explanation, question order, neutral prompts, recording and closing procedure. Notice whether questions invite detailed accounts or repeatedly produce brief answers.

Then transcribe a section and try your intended analytical approach. This tests whether the material addresses the research question and whether your workload is realistic.

For laboratory work, examine equipment performance and measurement consistency. For document-based research, trial the search, selection and coding procedures.

Across designs, record what happened without changing several procedures unnoticed. Keep dated versions of materials so you can distinguish the original method from the revised one.

Analyse Feasibility Before Testing Outcomes

Start with the outcomes your pilot was designed to assess. Feasibility findings and preliminary outcome signals from a pilot trial inform whether to proceed to a later full-scale trial. Report numbers approached, eligible, consenting and completing where available, and distinguish recruitment targets from actual invitation counts.

Examine missing responses, technical failures, participant feedback and the time required for each task. Use these findings to decide whether to proceed or revise.

Pilot studies are generally underpowered for reliable efficacy conclusions. Small samples produce uncertain estimates, so a non-significant result doesn’t show that no effect exists. Overinterpreting uncertain findings carries a high risk of error.

Statistical significance in a pilot doesn’t establish clinical significance or prove that an intervention works in its intended population.

Where appropriate, report effect size and confidence intervals, but acknowledge their uncertainty. Don’t use an unstable pilot effect as the sole basis for planning a definitive study.

In their Advances in Neonatal Care editorial, Jacqueline McGrath and Debra Brandon describe findings in neonatal science as a low level of evidence for practice. Positive findings support a promising concept, rather than a treatment recommendation.

The introduction to the CONSORT pilot extension explains the distinct purpose of preliminary trials. Their contribution is to prepare better research, not to support premature clinical conclusions.

Report What Changed and Why

In your methodology chapter, explain the pilot’s purpose, participants, recruitment, procedures, analysis and decision criteria. Then state what actually happened, including departures from the plan.

Describe the connection between each finding and the resulting revision. Separate changes to wording or navigation from changes that alter the meaning of a measure.

Your revision record should identify the problem, supporting evidence, action taken and whether you retested it. Attach relevant materials in an appendix where your course permits.

Our dissertation methodology examples show how procedural detail connects with methodological justification. For randomised pilot trials preparing a definitive trial, the CONSORT pilot reporting extension offers tailored guidance.

Discuss limitations, including sample composition, uncertainty and limited generalizability. Explain what the next steps in the research need to establish.

Some journals may decline to publish pilot studies because readers may overinterpret small-sample findings. Clear reporting needs to state both the pilot’s useful contribution and the claims it cannot support.

Frequently Asked Questions About Dissertation Pilot Studies

Can I Include Pilot Data in My Final Dataset?

Sometimes, but the decision needs methodological and ethical justification. Inclusion is more defensible when procedures remain unchanged and consent covers the intended use. Substantial changes to questions, eligibility or delivery can make the datasets incompatible. Explain your decision transparently rather than merging data simply to increase the sample.

Does a Successful Pilot Prove My Survey Is Valid?

No. A pilot can reveal confusing wording, missing response categories and technical faults. However, smooth completion doesn’t establish every aspect of validity or reliability. You still need to explain what each measure captures and why it suits your question. Use relevant measurement evidence alongside the pilot findings.

Does Every Dissertation Need a Separate Pilot?

No. The value depends on the design and unresolved uncertainties. Established procedures may need targeted testing rather than a complete rehearsal. Secondary-data projects can instead trial access, cleaning or coding. The decision should follow a clear assessment of what could prevent the planned research from producing usable evidence.

Must a Pilot Lead to a Randomised Controlled Trial?

No. Clinical pilot trials often prepare for full-scale randomised controlled trials, but dissertation pilots also support surveys, interviews, laboratory studies and document analysis. The next stage should match the research question. A pilot may also show that the proposed larger study needs substantial redesign or shouldn’t proceed.

Use the Pilot to Make a Defensible Decision

A useful pilot produces a clear decision about your methods: proceed, revise or stop. Its strength lies in documented learning, rather than impressive-looking findings from a small sample.

Before full data collection, retain the tested materials, revision record and reasons for your final choices. Use them to decide your next steps. These records also help you explain how your dissertation design became workable and what its results can reasonably support.