論文アブストラクト：We often base our decisions on uncertain data - for instance, when consulting the weather forecast before deciding what to wear. Due to their uncertainty, such forecasts can differ by provider. To make an informed decision, many people compare several forecasts, which is a time-consuming and cumbersome task. To facilitate comparison, we identified three aggregation mechanisms for forecasts: manual comparison and two mechanisms of computational aggregation. In a survey, we compared the mechanisms using different representations. We then developed a weather application to evaluate the most promising candidates in a real-world study. Our results show that aggregation increases users' confidence in uncertain data, independent of the type of representation. Further, we find that for daily events, users prefer to use computationally aggregated forecasts. However, for high-stakes events, they prefer manual comparison. We discuss how our findings inform the design of improved interfaces for comparison of uncertain data, including non-weather purposes.