Do You Need a Pilot or Feasibility Study Before Your Main Research?

Does Every Main Study Need a Pilot or Feasibility Study First?

You have developed the research question, selected a design, identified participants, chosen instruments, and planned data collection. Before beginning the main study, another question appears: are you supposed to conduct a pilot or feasibility study first?

The answer is not automatically yes. Preliminary research consumes time, funding, researcher effort, and sometimes participant contributions. Conducting it simply because “researchers should always pilot their studies” can create work without resolving any meaningful uncertainty.

At the same time, proceeding directly to the main study can be risky when critical assumptions remain untested. If recruitment is uncertain, procedures are unfamiliar, an intervention has not been implemented in the intended context, or several complex study processes must work together, discovering problems after full-scale data collection begins may be far more costly than investigating them beforehand.

You Need Preliminary Work When Important Uncertainty Remains

How to Decide Whether Preliminary Research Is Necessary

Do Not Ask Whether Piloting Is Mandatory; Ask What You Still Do Not Know

A pilot or feasibility study is not a ritual that automatically makes a research project rigorous. Its value comes from reducing specific uncertainties before greater resources are committed to the main study.

Start by examining the assumptions built into your protocol. Can you recruit the required participants? Will they remain in the study? Can the intervention or procedure be delivered as intended? Can participants complete the measurements? Can researchers implement the protocol consistently? Will the resulting data be usable?

If these questions are already supported convincingly by previous research and by conditions that closely match your proposed study, additional preliminary work may have limited value. If the answers are uncertain, some form of testing may be warranted.

The broader principle is to test the uncertain parts of the research design before committing to the main study, rather than automatically reproducing the entire protocol at a smaller scale.

You Are More Likely to Need Preliminary Work When the Design Contains Important Unknowns

Some studies involve substantially more operational uncertainty than others. A newly developed intervention, unfamiliar technology, novel recruitment strategy, complex sequence of assessments, difficult-to-reach population, multisite protocol, or procedure being transferred into a new context may introduce questions that cannot be answered adequately from the protocol alone.

For example, a study may require participants to complete repeated assessments over six months. The measures themselves may be well validated, yet the researchers may have little evidence about whether the intended population will remain engaged for that long. In that situation, the unresolved problem is not measurement validity but retention and study burden.

Preliminary work should be designed around that uncertainty.

Novelty Alone Does Not Determine Whether You Need a Pilot

A study does not need preliminary testing merely because its research question is new. Conversely, a familiar research question does not guarantee that its implementation is straightforward.

The relevant novelty is often procedural or contextual. You may be using a familiar survey with a population for whom its language has not been tested. You may be implementing an established intervention in a substantially different setting. You may be using validated measures but combining them into a demanding assessment schedule that has never been tried.

Ask what is genuinely new about the way the study will operate and whether that novelty creates meaningful uncertainty.

Existing Evidence Can Sometimes Replace New Preliminary Research

Before planning a pilot or feasibility study, examine what is already known. Previous studies may provide credible evidence about recruitment rates, intervention delivery, instrument performance, follow-up, participant burden, or other procedures relevant to your design.

The closer that evidence is to your intended population, setting, procedures, and implementation conditions, the more informative it may be. If several well-conducted studies have already demonstrated that essentially the same procedures work under comparable conditions, repeating them solely to say that your study was “piloted” may not be an efficient use of resources.

However, evidence should not be treated as transferable without thought. A recruitment strategy successful in one health system, university, country, age group, or delivery environment may perform differently elsewhere. The question is not whether evidence exists, but whether it adequately addresses your particular uncertainty.

Sometimes You Need Only a Focused Pretest

Not every uncertainty requires a separate pilot or feasibility study. If the main concern is whether participants understand questionnaire items, interview questions, instructions, or a particular task, focused pretesting may answer the question more efficiently.

Likewise, a researcher may rehearse a laboratory workflow, test a data-entry system, examine questionnaire completion time, or simulate an analysis pipeline without conducting a miniature version of the complete study.

Before choosing an approach, distinguish among a pretest, pilot study, and feasibility study. The smallest form of preliminary work capable of resolving the consequential uncertainty may be the most defensible choice.

A Feasibility Study Is Useful When You Need to Know Whether and How the Study Can Be Done

Feasibility work is appropriate when important questions remain about the practicability of the proposed research. These might concern recruitment, retention, acceptability, intervention delivery, staffing, resources, site participation, assessment completion, or other features that determine whether the future study can proceed successfully.

The methods should follow the question. If the uncertainty concerns participant acceptability, qualitative interviews may be informative. If it concerns recruitment, researchers may need evidence about eligibility, approach rates, consent, and recruitment over time. If it concerns multiple interacting processes, implementing those processes together may be necessary.

In other words, a feasibility study should produce information that changes your confidence about whether and how to proceed. Before undertaking one, specify what the feasibility study needs to tell you.

A Pilot Is Useful When You Need to See the Intended Study in Operation

A pilot is particularly useful when the uncertainty cannot be resolved by examining individual components separately. The recruitment method may appear reasonable, the intervention may appear acceptable, and the instruments may be validated, but researchers may still need to see whether the complete sequence operates as intended.

A pilot can therefore test part or all of the future study on a smaller scale. The exact objectives might concern whether recruitment and randomization can be completed, whether intervention delivery is sufficiently consistent, whether assessments occur on schedule, whether data are recorded correctly, or whether follow-up procedures are workable.

This is also why a pilot should have explicit objectives. Researchers should know what the pilot is supposed to test before recruiting its first participant.

The More Consequential Failure Would Be, the Stronger the Case for Testing

The decision is partly about risk. Imagine two proposed studies.

The first is a modest interview study using a familiar recruitment route and an established interview procedure. The second is a large multisite intervention study requiring substantial funding, specialized training, repeated participant assessments, and coordination among several institutions.

Even if both contain uncertainties, the consequences of discovering a fundamental procedural problem after launch are very different. Preliminary testing is particularly valuable when failure would waste substantial resources, expose many participants to research procedures without producing useful evidence, compromise data quality, or make the main study impossible to complete.

This does not mean that larger studies always require pilots and smaller studies never do. It means that the extent of preliminary work should be proportionate to both uncertainty and consequence.

Do Not Conduct a Pilot Merely to Look for a Promising Effect

A common temptation is to conduct a small pilot, examine whether the substantive outcome appears favorable, and proceed to the main study only if the result looks promising. That can distort the purpose of pilot research.

Small preliminary samples often provide imprecise estimates of effects. A large observed effect may be an overestimate; a small effect may underestimate what the definitive study would find. Statistical significance in a pilot may therefore provide a poor basis for deciding whether the substantive research question deserves a properly designed main study.

Plan What the Preliminary Findings Will Make You Do

Preliminary research becomes much more useful when researchers specify how its findings will affect the next decision. In some contexts, this is formalized through progression criteria.

For example, researchers might specify acceptable recruitment or retention levels, requirements for successful intervention delivery, or conditions under which procedures would need modification. The result may support proceeding unchanged, proceeding with modifications, conducting further preliminary work, or reconsidering the proposed main study.

The thresholds should be justified for the particular study rather than copied mechanically from unrelated research. Their purpose is to connect evidence with action.

When a Pilot Is Useful and When It May Add Little

The difference is not that one research design belongs to a category that “requires a pilot.” The difference is the amount and consequence of unresolved uncertainty.

Common Mistakes When Deciding Whether to Conduct Preliminary Research

Use Uncertainty and Consequence to Make the Decision

Before adding a pilot or feasibility study to your research plan, identify the assumptions that must hold for the main study to succeed. Then assess how confidently each assumption is already supported.

Also consider whether your proposed preliminary study is large and informative enough for its objective. “Small” is not synonymous with “pilot.” The sample should be justified by the feasibility parameter, process, or uncertainty being investigated rather than chosen simply as an arbitrary fraction of the intended main sample.

Finally, leave enough time to act on what you learn. A pilot conducted immediately before main-study recruitment, with no opportunity to revise procedures or obtain any necessary amended approvals, defeats much of its purpose.

Before Deciding Whether You Need a Pilot or Feasibility Study

Frequently Asked Questions About Whether Preliminary Research Is Needed

Do Preliminary Research Because You Need an Answer, Not Because You Need a Pilot

How to Cite This Guide

This guide is intended to be read, shared, and used in research, teaching, and academic work. If you draw on its ideas, explanations, or other content, please acknowledge the source by citing the guide. Doing so gives appropriate credit and helps your readers locate the original resource.