Stability analysis of slopes commonly deals with limited data relating to the characterisation of the material strength properties. This generates an uncertainty level that depends on the number of measurements used to determine them, and that is not related to the intrinsic variability of the material known as epistemic uncertainty. Considering a Mohr–Coulomb material characterised by its cohesion and friction angle, we describe a method to use interval estimation methods (confidence intervals) to estimate the level of epistemic uncertainty as a function of the number of available measurements (sample size), obtaining ranges of plausible values for the mean of the strength parameters. The calculated uncertainty in strength parameters is later propagated through a deterministic slope stability analysis, obtaining a relation between test sample size and the uncertainty in Factor of Safety (FoS) calculation, which can be compared with the typical safety margins used in slope designs.
