This document discusses the challenges of communicating scientific uncertainty to non-expert audiences. It notes that people vastly underestimate their own uncertainty and that perceptions of uncertainty are culturally mediated. Additionally, scientific literacy and numeracy varies widely among the general public. The document advocates for empirically testing risk communication messages with audiences and presenting probabilities and uncertainties in multiple consistent ways using words, numbers, and graphics. It also cautions that for some issues with high uncertainty, simpler approaches than probabilistic analysis may be better suited.