Permute colorfield cells within known intervals for average judgments

Shuffle values within each known aggregation interval of a colorfield when the task is to compare interval averages. For example, randomly permute the colored cells inside each month block of a time-series colorfield so local color mixtures represent the block average more directly than the original ordered strip.

Within a colorfield, permutation brings the colors to be averaged closer together and makes smaller local regions look more like the interval as a whole. That improves average judgments for the color display, but the same idea did not help line graphs.

Mechanism: When color samples from the same interval are spatially mixed, viewers can pool more local color evidence instead of averaging across a longer ordered run.

Evidence: In the study, permuted colorfields outperformed ordered colorfields on the interval-average task, while permuting line graphs did not produce a significant benefit, yielding a significant interaction between display type and permutation (Correll et al., 2012).

Notes: The paper treats this as a more specialized move than the base colorfield because it depends on known aggregation boundaries and removes low-level patterns within the interval.

• User Goal: Choose which predefined interval has the highest average.
• Task: Compare averages across fixed interval blocks already shown in the display.
• Data: One ordered series encoded as color blocks, where preserving within-interval order is not required for the immediate task.
• Chart Setting: A colorfield or heatmap-like strip with explicit interval boundaries.
• Audience: Viewers making rapid aggregate judgments from the full display.
• Success Criterion: Higher accuracy on the interval-average decision than the ordered colorfield version.

Break it when: The aggregation duration is not known in advance, or readers need the within-interval temporal pattern. Why: The method depends on predefined aggregation ranges and the permutation destroys the original low-level ordering.

Sacrifice: You lose the original order and low-level pattern inside each interval. Risk: The permuted block can show visual patterns that are not representative of the original sequence. Mitigation: Limit the permutation to cases where fixed-interval average judgment matters more than within-interval sequence interpretation.

Mistake: Permuting points inside each interval of a line chart to improve average judgments. Why it fails: The study found no significant performance gain from permuted line graphs.

Failure Sign: Readers still struggle to identify the best interval on an ordered colorfield, especially when top intervals are close or noisy. Quick Check: Compare ordered and permuted versions of the same colorfield on representative hard cases and measure which one yields more correct interval choices. Stronger Test: Include noisy series and close competing intervals, where the performance difference was most apparent.

• Randomly permute the colorfield cells inside each predefined interval block.
• Keep the permutation confined within each interval so interval membership stays unchanged.
• Do not apply the same permutation as a repair for a line chart; use it only on the colorfield.