In this chapter we consider a different approach to estimating subjective probability distributions, focused on designing novel experimental methods for measuring those distributions directly. The key idea is to design experiments that let us sample from subjective probability distributions. In this way, we can form an estimate of the distribution from the samples.