Uncertainty Visualization Study Group Notes – 01/28/2013

For paper overview, see slides

General Discussion Inspired by Paper

• Expert vs General Public
  • Who benefits more from the inclusion of uncertainty information?
• Discrete vs Continuous Probability
  • Do we always mentally map the continuous case to a discrete approximation?
  • What do we do if we know people will use the wrong mental model when dealing with continuous probability?
• Well-Formed vs the Ill-Formed / Unknown
  • Perhaps uncertainty visualization is more needed for ill-formed questions, when there isn’t a right answer that we can just tell them
  • Sometimes the question is well formed and there is a “right” answer, but assessment of the level of risk deemed “acceptable” needs to remain a human decision (either from a public policy standpoint or because it’s a choice that must be left to the individual)
    • In case of hurricane, decision of “should I evacuate?” is more complex than “Is the hurricane going to hit my home?”. The potential for damage vs dealing with traffic gridlock in the ensuing panic all factor in to evacuation decisions.
      • Exactly what info needs to / should be given to the general public in this case? How does it differ from the info that needs to be provided to the policy makers who actually give the evacuation orders?
  • Response from Mike Goodchild lecture: Historically, the cost of making a single map was high, so they needed to be as general-purpose as possible. Now there is very little cost to generate a new map; creating many maps each with a distinct purpose is not prohibitive.
    • Doesn’t really address what to do in the case that you have an ill-formed / unknown question
    • Still, it’s important to recognize that the argument for the primary purpose of visualizations being general-purpose, best suited to answering ill-formed / unknown questions, is based on a premise that doesn’t apply today in the same way it did historically
• Uncertainty and Decision-Making
  • Does uncertainty visualization produce better decision-making? Is there literature to support this?
    • Joslyn (however her papers present well-formed questions)
    • Visualization of Reasoning paper –> slow thinking as producing better choices than fast thinking
      • Uncertainty visualization as “slowing someone down”
        • The introduction of doubt produces slower and better decision-making
  • Last week’s Gigerenzer paper, as a response, argues that heuristics (for fast-thinking) are adaptive and can be effective
  • [Miriah] Perhaps a series of ecologically rational decisions, with information up front, would be useful: see how much they use facts, heuristics, etc.; see how quickly people actually make decisions…