Use superimposed value-and-uncertainty maps for integrated lookup

Use a single superimposed map when readers must combine value and uncertainty in one judgment. For example, encode both variables in the same heatmap cell instead of asking readers to match a value heatmap to a separate uncertainty heatmap.

Superimposed maps remove a second search step. Readers can decode one location once, instead of finding the same region in two separate displays and mentally joining the results.

Mechanism: Superposition reduces cross-chart correspondence work when a task depends on both the value and the uncertainty of the same mark.

Evidence: In the paper’s identification experiment, superimposed charts were significantly more accurate than juxtaposed charts for locating a target value-uncertainty pair in a heatmap-based task (Correll et al., 2018).

• User Goal: Find or verify locations that satisfy both a value condition and an uncertainty condition.
• Task: Integrated lookup from a bivariate legend.
• Data: Quantitative values paired with uncertainty for the same spatial cells or regions.
• Chart Setting: Dense map-like or heatmap-like displays where the same position appears in every view.
• Audience: Readers who need to make one decision from both variables at once.
• Success Criterion: Higher identification accuracy with less cross-view confusion.

Break it when: The goal is to inspect value patterns and uncertainty patterns separately rather than fuse them into one immediate judgment. Why: Separate views let readers consider each variable on its own, which the paper notes can be useful for orthogonal analysis even though it hurts integrated lookup.

Sacrifice: You lose the simplicity of two independent univariate scales. Risk: Integrated color channels can be harder to decode than two separate single-variable displays. Mitigation: Keep the superimposed display for fusion tasks, and add coordinated highlighting if a separate view must remain.

Mistake: Put value and uncertainty in adjacent maps and expect readers to combine them quickly during lookup. Why it fails: Readers must perform two searches and then match the results across views, which adds error.

Failure Sign: Readers look back and forth between views or misidentify cells that satisfy only one of the two conditions. Quick Check: Build one superimposed version and one side-by-side version of the same task, then ask a reviewer to find a specific value-uncertainty pair. Stronger Test: Time and score a short A/B lookup task; keep the structure that yields fewer mismatches on integrated questions.

• Merge value and uncertainty into one co-located bivariate encoding for each mark.
• Replace separate value and uncertainty heatmaps with one heatmap that encodes both variables in each cell.
• If you must keep separate views, add interaction that highlights corresponding regions across charts.