Guidelines
Suggest edit

Mirror bar small multiples to compare overall similarity

For association comparison between two paired quantitative series, use mirrored alignment on bar-chart small multiples to improve fidelity and mitigate missed overall-similarity judgments for brief visual comparison.

  • purpose:refine
  • basis:empirical
  • task:relate
  • chart:bar
  • structure:small-multiples
  • quality:fidelity:use
  • lever:layout-structure
  • operator:association
  • group-cardinality:binary

advice

Mirror the bar pair for similarity judgment

Reverse one bar panel so matching bars face across a center line when the job is judging which pair of series is more similar overall. For example, use a mirrored left-right bar pair instead of stacking the two bar panels vertically.

reason

Why mirroring helps similarity reading

A mirrored pair puts corresponding values near each other and supports comparison of overall pattern similarity without the larger separation of stacked panels.

Mechanism: Mirroring lets readers compare the two series across a shared center, which helps them judge overall similarity more efficiently than a vertically split layout.

Evidence: In the collated record, mirrored bar small multiples ranked first for the correlation task and significantly outperformed stacked bars; the original experiment framed this task as choosing which pair of bar-chart series was more similar overall (Zeng & Battle, 2023; Ondov et al., 2019).

context

Use when judging overall similarity

  • User Goal: Decide which paired series is more similar overall.
  • Task: Compare overall correlation or similarity between two series.
  • Data: Two paired quantitative series per comparison.
  • Chart Setting: Bar-chart small multiples used for brief visual comparison.
  • Success Criterion: Readers can choose the more similar pair with smaller required correlation differences.

exceptions

Do not use when the comparison stops being a two-way pattern judgment

  • Break it when: More than two datasets must be compared at once. Why: The paper notes that mirroring is unlikely to scale beyond two datasets.
  • Break it when: The main goal shifts to finding the single biggest change and a shared overlaid bar view is feasible. Why: A different arrangement led for that task.

costs

Tradeoffs of mirrored similarity comparison

Sacrifice: A conventional same-direction axis across both panels. Risk: The mirrored layout is less suitable when many datasets must be shown together. Mitigation: Use it for binary comparisons only.

mistakes

Common correlation-comparison mistake

Mistake: Use a vertically stacked bar pair when asking readers to judge which series pair is more similar overall. Why it fails: The stacked layout required the strongest signal in the study.

check

Check the similarity layout

Failure Sign: Readers need almost identical patterns before they can confidently pick the more similar pair. Quick Check: Compare a mirrored pair against a stacked pair on the same two-way similarity judgment. Stronger Test: Keep the version that supports correct judgments at smaller correlation differences.

fix

Repair the similarity layout

  • Reverse one panel’s x-axis direction.
  • Place the two bar panels around a shared center line.
  • Restrict the view to one two-way comparison at a time.

References

Ondov, B., Jardine, N., Elmqvist, N., & Franconeri, S. (2019). Face to Face: Evaluating Visual Comparison. IEEE Transactions on Visualization and Computer Graphics, 25(1), 861–871. https://doi.org/10.1109/TVCG.2018.2864884
Zeng, Z., & Battle, L. (2023). A Review and Collation of Graphical Perception Knowledge for Visualization Recommendation. Proceedings of the 2023 CHI Conference on Human Factors in Computing Systems, 1–16. https://doi.org/10.1145/3544548.3581349