<p>Testing occurs at nearly every stage of a complex system’s lifetime to demonstrate margin to engineering requirement. If the population of units is known and finite, a scenario common in many defense applications where production quantities of defense systems are often constrained, then factoring in population size and the proportion sampled can reduce sampling uncertainty. To account for this fact, a novel methodology for characterizing uncertainty in estimates of variance and quantiles–key parameters for assessing margin to requirements–in samples taken from finite populations is proposed in this paper. Through innovative bootstrap resampling, confidence intervals are constructed for the variance and quantile estimators under the assumption of finite populations. In extensive simulation studies on synthetic data these intervals are shown to be shorter in length than their infinite population counterparts. This increases the feasibility and likelihood of asserting positive margin when it exists. The method developed is then demonstrated via an additive manufacturing exemplar, where a known population of piece-parts was produced and evaluated against dimensional performance requirements.</p>