Pilot studies play a vital role in the development of scientific research. By providing a preliminary sketch of the research design and probable outcomes, pilot studies allow researchers to deal with potential issues that may arise in the actual study. Pilot studies also provide funding agencies with an idea of how their resources will be used. As such, the design of a pilot study can have great implications for research funding. Though every study is unique, there are some general guidelines that should be followed to ensure a well-designed pilot study.
Determine your sample size. Using past research in your area, make estimates of means and variability to determine the minimum sample size necessary to obtain statistically significant results. Where past research is not available, a good rule of thumb is to use a minimum of 30 participants per cell for quantitative research. Much fewer participants are needed for qualitative research, and it is often possible to conduct a pilot study using a single participant (depending on your study's design).
Remove any unnecessary or repetitive elements from your research design. Focus on those aspects of your study that differ from past research and run your study using only necessary tools or variables of interest. For example, your proposed study may involve numerous drug types and doses, but your pilot study may only use the two or three drugs that you're particularly interested in. This will help both to limit resources spent on your pilot study and to determine the feasibility of your proposed study (albeit on a much smaller scale).
Test your consent forms and recruitment process. Recruit participants for the pilot study in the same way that you would for the proposed study and have these participants complete your informed consent procedures. An inability to attract the desired number of participants or obtain consent for your pilot study presents severe problems for your full study. As these issues will be much more pronounced in your full study due to the greater number of participants desired, problems with recruitment or consent may indicate a need to prematurely abandon your study.
Test any data collection forms or questionnaires by having them administered as you would in your full study. This helps you to determine if there are any issues with understanding or completion of the forms, both for forms completed by participants and data collection performed by multiple research assistants.
Use any deception or randomization procedures planned for your full study in the pilot study. As these can be difficult to plan and administer, this allows you to determine whether your proposed method of randomization or deception is feasible. It also provides you with the opportunity to ask participants whether they were aware of deception or the group to which they had been randomly assigned.
Test any elements of your study that may lead to discomfort or difficulties for the participant. For example, certain drugs may have severe side effects, or a certain interview protocol may involve questions that participants are not comfortable answering. If certain elements are found to be unacceptable to most participants, you may need to alter these aspects, remove them from your study or replace them with something more tolerable (where possible).
Select your outcome measures. Pilot studies can greatly help in the design of your study, as they allow you to test out multiple outcome measures to determine which will be used in the full study. Where the use of multiple outcome measures is incorporated into your study's design, a pilot study may help you to weed out those that are either troublesome or unimportant.
Matthew Lee has been writing professionally since 2007. Past and current research projects have explored the effect of a diagnosis of breast cancer on lifestyle and mental health and adherence to lifestyle-based (i.e. nutrition and exercise) and drug therapy treatment programs. He holds a Master of Arts in psychology from Carleton University and is working toward his doctorate in health psychology.