Conceptual issues in sample size calculation for clinical trials are discussed. Each parameter and related issues are highlighted: type I and II errors and the balance between them, one versus two-sided testing, the smallest clinically meaningful difference versus the anticipated treatment effect, endpoint selection, multiple comparisons, missing data, unequal allocation, repeated measures, equivalence, and group sequential trials. For survival trials, conceptual issues are discussed in further detail, including the binomial versus log rank statistic, exponential versus proportional and nonproportional hazards, and adjustment for treatment lag, staggered entry, noncompliance, drop-in, loss-to-follow-up and competing risks, and duration of treatment effect.