Estimating the Mean and Variance of Subjective Distributions in PERT and Decision Analysis

Abstract

The original PERT formulae [5] employ the mode and endpoints to estimate the mean and standard deviation of subjective probability distributions. Though widely-used, they have met with criticism for their inaccuracy, and for their being limited to the beta distribution when there is no a priori reason why the distribution should be beta ([4], [6]). This note presents alternative, accurate and relatively distribution-free formulae for estimating the means and standard deviations of subjective probability distributions.