Guidelines
Suggest edit

Avoid luminance encoding for group comparisons of numerosity

For relative numerosity comparison, avoid luminance encoding for group identity on point-based displays to prevent count bias and mitigate darker-looks-more errors for overview readers.

  • purpose:refine
  • basis:empirical
  • task:compare
  • quality:fidelity
  • lever:encoding
  • channel:color-lightness:avoid
  • reading-mode:overview

advice

Group coding for count comparison

Do not separate compared groups mainly by dark-versus-light coding when readers must judge which group is more numerous. For example, use hue or orientation rather than a darker and lighter set of marks when the display asks readers to compare counts.

reason

Why luminance biases count judgments

Readers do not see numerosity independently of all other mark properties.

Mechanism: Darker groups can appear more numerous than lighter groups, which biases relative count judgments even when the true counts are the same.

Evidence: The paper states that darker collections can appear more numerous and warns designers to be careful using luminance when relative numerosity matters; it also notes that relative numerosity is robust across color and orientation in visualization studies (Szafir et al., 2016).

context

Use when relative counts matter

  • User Goal: Compare how many items belong to each group.
  • Task: Make a relative numerosity judgment.
  • Chart Setting: Multiple groups of marks are distinguished visually within one display.
  • Audience: Readers are making quick overview estimates rather than exact counts.
  • Success Criterion: Equal-count groups do not look different in quantity because of the encoding.

exceptions

Do not use when count is not being read from the view

Break it when: Relative numerosity is not a relevant reading of the display. Why: The reported bias matters specifically for count judgments.

costs

Tradeoffs of avoiding luminance

Sacrifice: Lightness is no longer available as the main grouping cue for the categories being counted.
Risk: If kept, darker groups can look larger than equal lighter groups.
Mitigation: Move the compared groups to hue or orientation instead.

mistakes

Common failure mode

Mistake: Using a dark group and a light group to carry categories whose counts will be compared. Why it fails: The dark group can inherit a false numerosity advantage.

check

Check for luminance-driven count bias

Failure Sign: An equal-count darker group looks more numerous than a lighter one.
Quick Check: Make a test display with equal counts in dark and light groups and ask which group seems larger.
Stronger Test: Repeat the same comparison after switching the groups to hue or orientation and compare the error pattern.

fix

Fix the encoding

  • Replace lightness-based group coding with hue for the categories being counted.
  • Replace lightness-based group coding with orientation when that channel is available for the same comparison.
  • Remove dark-versus-light grouping from the primary view if numerosity comparison is part of the task.

References

Szafir, D. A., Haroz, S., Gleicher, M., & Franconeri, S. (2016). Four types of ensemble coding in data visualizations. Journal of Vision, 16(5), 11. https://doi.org/10.1167/16.5.11