Sample size and power planning: Prepare your study assumptions here, then use the external MedCalc resource to identify the appropriate sample size method.
MedCalc sample size resource
Open MedCalc’s sample size calculation menu and guidance
The linked page is part of MedCalc’s software manual. It organizes methods for means, proportions, paired samples, correlation, survival analysis, and other designs, alongside precision-based sample size guidance. Follow the relevant method’s instructions and check any software requirements. Resource checked: 10 October 2026.
Choose the calculation around your study
Begin with the primary question and outcome. A calculation for estimating a proportion answers a different planning question from one for comparing means, assessing a time-to-event outcome, or analyzing clustered data. Match the calculation to the proposed design and analysis.
Prepare these inputs
- Primary objective: What will the main analysis estimate or test?
- Outcome and design: Specify the measurement, comparison groups, allocation, and whether data are paired, repeated, or clustered.
- Planning target: Define the required precision or a meaningful effect that the study aims to detect.
- Assumptions: Record expected variability, event rates, or other design-specific inputs and where they came from.
- Statistical settings: If relevant, specify the significance level, power, and one-sided or two-sided approach with a rationale.
- Practical losses: Consider nonresponse, dropout, unusable measurements, and eligibility failures separately.
Worked example: allowing for dropout
Suppose a separate, appropriate sample size calculation establishes that you need 100 completed participants. If you expect 20% of enrolled participants not to complete follow-up, the recruitment target is:
Required enrollment = completed participants needed ÷ (1 − anticipated dropout proportion).
In this example: 100 ÷ (1 − 0.20) = 125 participants to enroll. Round up when the result is not a whole number. Simply adding 20% to 100 would give 120, which would leave an expected 96 completers under this assumption.
This adjustment only addresses anticipated losses. It does not establish whether 100 completers is statistically sufficient, and it does not resolve bias caused by dropout.
Report a range of scenarios
Compare plausible input values rather than relying on one optimistic assumption. Document the method, software and version, inputs, source of assumptions, result, and any recruitment adjustment.
Use the result in your proposal
Check whether the recruitment target is achievable within your time, budget, and setting. Discuss the proposed calculation with a supervisor or statistician when the design involves complex dependencies or multiple objectives.
Continue with biostatistics and research methods, or return to the hub.