People are continually faced with decisions that have uncertain outcomes. Often there are many different uncertainties within each of the options available. This research focuses on uncertainty that is continuous, such as the amount of time it will take for a person to commute home or their monthly grocery expenses. When making choices with uncertain outcomes people must (1) construct judgments about the nature of the uncertainty and then (2) form a preference about these uncertainties. Chapter 1 looks at the impacts of different elicitations metrics on uncertainty judgments. Specifically, how do different, but equivalent metrics, impact the amount and shape of the uncertainty estimated. This research relies on three empirical regularities from the cognitive psychology literature: uncertainty perceptions scale with magnitude, a tendency to neglect units, and a strong prior that uncertainty distributions are symmetric. Capitalizing on these empirical regularities, we find striking inconsistencies in judgments of uncertainty, both the amount and the symmetry, depending on the elicitation metric used. In Chapter 2, we find a robust preference for positively correlated uncertainty over negatively correlated uncertainty in the domain of gains. Based on the stimuli used, only someone with risk-seeking preferences would be expected to choose the option with positively correlated uncertainty. Two common decision heuristics, two different display formats, and three different outcomes all result in the same general preference for positively correlated uncertainty. A third display format, similar to a histogram, was the only manipulation that appeared to align single attribute risk preferences and multi-attribute choice. Both chapters offer new phenomena to be accounted for in future models of uncertainty.