Sample Size Calculator
Choose a scenario by study type and enter the parameters to estimate the required sample size. Covers common clinical/imaging research designs such as mean comparison, correlation, proportion estimation and reference interval studies.
① Choose study type
② Enter parameters
How to use & methodology
Why estimate sample size before a study?
Too small a sample gives insufficient power and false-negative conclusions; too large wastes resources and prolongs the study. The design stage (and ethics/grant applications) usually requires a sample-size justification, so it should be done before data collection.
What do the key parameters mean?
α (significance level) is usually 0.05 — the false-positive probability; power (1−β) is usually 0.8 or 0.9 — the probability of detecting a true difference; the expected difference δ and standard deviation σ reflect the effect size and can come from a pilot study or the literature.
How to choose the study type?
Use "two-group mean comparison" to compare two group means; "single-group mean estimation" for the precision of estimating a single mean; "correlation study" to study two variables' correlation; "single-group proportion estimation" for a positivity rate; and "reference interval study" to establish a normal reference range.
Is the computed value the final number?
No. It is the theoretical minimum and excludes loss to follow-up and dropout. Studies usually add 10–20% on top to allow for missing data. For multifactor, stratified or repeated-measures designs, consult a statistician.