Piloting is the process of running a small-scale test before committing to a full-scale project. While the term applies across many fields, from aviation to software development, it’s most commonly encountered in scientific and clinical research, where a pilot study serves as a trial run of an entire research plan. The core question piloting answers is simple: “Can we actually do this, and if so, how?”
What Piloting Means in Research
A pilot study is the first step in a research protocol. It mirrors all the procedures of the planned full-scale study but on a much smaller scale. Researchers use it to test whether their methods, tools, and logistics work together before investing the significant time and money a large study demands.
The goal is not to answer the main research question. A pilot study doesn’t tell you whether a new treatment works. Instead, it tells you whether your plan for testing that treatment is sound. Can you recruit enough participants? Do your measurement tools function properly? Can your team follow the procedures consistently? These are the questions piloting is designed to resolve.
Think of it like a dress rehearsal. A theater company doesn’t perform for a full audience without first running through the entire show to find problems with lighting, timing, and blocking. Piloting does the same thing for research.
What a Pilot Study Actually Tests
Pilot studies evaluate several practical dimensions at once. Recruitment is a major one: researchers check whether they can find and enroll enough people who meet the study’s criteria within a reasonable timeframe. They also test whether participants stick with the study long enough to be useful, since dropout rates can destroy a project’s results.
Beyond recruitment, piloting examines:
- Safety of treatments or interventions before exposing a larger group
- Equipment and instruments to confirm they measure what they’re supposed to measure
- Data collection procedures to see if forms, surveys, or lab protocols actually work in practice
- Randomization and blinding processes to ensure participants are properly assigned to groups without bias
- Team readiness by giving researchers and assistants hands-on experience with the study methods
Piloting also helps estimate how many participants the full study will need. Getting this number wrong is one of the most expensive mistakes in research: too few participants and the study can’t detect real effects, too many and you’ve wasted resources and potentially exposed extra people to experimental treatments unnecessarily.
How Big a Pilot Study Needs to Be
There’s no single answer, but researchers have developed several rules of thumb. Common recommendations for total pilot sample size range from 20 to 70 participants, depending on the study design and who you ask. One widely cited guideline suggests at least 30 participants to estimate basic study parameters. Another recommends 12 per treatment group as an absolute minimum, which means 24 total for a study comparing two groups.
When the expected effect of a treatment is very small and subtle, pilot studies need more participants to generate useful estimates. For a study designed to detect a very small effect, researchers may need 75 people per group in the pilot phase. For a large, obvious effect, as few as 10 per group can suffice. The logic is straightforward: the harder the effect is to spot, the more data you need just to plan for it properly.
Piloting vs. Feasibility Studies
These two terms overlap significantly, and even researchers debate the distinction. One useful framework treats pilot studies as a subset of feasibility studies. Both ask “Can this be done?” but a pilot study has a specific design feature: it runs the future study, or part of it, in miniature. A feasibility study might instead explore broader questions, like whether a particular population is available or whether an intervention can be delivered in a certain setting, without necessarily replicating the full study design.
The UK’s National Institute for Health Research once drew a sharper line, positioning feasibility studies earlier in the research process and defining pilot studies specifically as “a version of the main study that is run in miniature.” In practice, many research frameworks now group them together under a single “feasibility and piloting” phase, recognizing that the distinction matters less than the shared purpose: making sure the full study is worth doing and capable of succeeding.
How Researchers Judge Whether to Proceed
A well-designed pilot study establishes clear benchmarks before it begins. These are specific, quantitative thresholds that define success. For example, a research team might decide in advance that at least 80% of enrolled participants must complete the study for the design to be considered workable. Or they might set an adherence benchmark like “at least 70% of participants in each group will attend at least 8 of 12 scheduled sessions.”
If the pilot hits these targets, the team moves forward with the full study, often with minor adjustments. If it falls short, they either redesign key elements or abandon the approach entirely. This is one of the most valuable functions of piloting: it prevents researchers from launching expensive, time-consuming studies that were doomed from the start by flawed logistics.
Some pilot results don’t function as strict go/no-go gates. When researchers are testing methods for reaching diverse or hard-to-access populations, for instance, the pilot data is more useful for refining the approach than for making a binary proceed-or-stop decision.
What Pilot Studies Can’t Tell You
One of the most common misuses of pilot data is treating it as evidence that an intervention works. Because pilot studies are small, they lack the statistical power to reliably detect whether a treatment has a real effect. A positive-looking result in a pilot could easily be due to chance, and a negative one doesn’t mean the intervention is ineffective.
The National Institutes of Health states this plainly: in a pilot study, you are not answering “Does this intervention work?” You are answering “Can I do this?” Pilot studies should report simple counts, averages, and percentages of their feasibility outcomes, not formal statistical tests of whether the treatment outperformed a placebo. Confusing these two questions has led to misleading published results and wasted follow-up research.
Reporting Standards for Pilot Trials
Published pilot studies follow a specific extension of the CONSORT guidelines, which are the international standard for reporting clinical trials. The pilot-specific checklist includes 26 items covering how participants were identified, how consent was obtained, and what criteria were set in advance for deciding whether to proceed to the full trial. Researchers must also describe any unintended consequences and outline proposed changes for the larger study.
These standards exist because pilot studies were historically published without enough detail for other researchers to evaluate them or build on them. A pilot study that simply says “recruitment was adequate” without reporting actual numbers, timelines, or predefined benchmarks gives the scientific community very little to work with.
Piloting Beyond Medical Research
The concept of piloting extends well beyond clinical trials. Software companies run pilot programs to test new features with a small group of users before a wide release. Schools pilot new curricula in a handful of classrooms before adopting them district-wide. Businesses pilot new processes or products in a single location before scaling up. In each case, the logic is identical: test at small scale, identify problems, adjust, then commit resources to the full rollout. The research world simply formalized what is, at its core, a common-sense principle of trying something small before going big.