Use adjacent rather than stacked bar pairs for correlation comparison

Place each bar-chart pair side by side rather than stacking the pair vertically when viewers compare similarity or correlation. For example, use left-right bar pairs instead of top-bottom bar pairs when asking which pair is more similar.

For this task, side-by-side placement makes correspondence between paired bars easier than a vertical stack.

Mechanism: Adjacent placement shortens the comparison path between corresponding bars, while stacked placement makes viewers work harder to match values across the two panels.

Evidence: In the bar-chart correlation experiment, adjacent small multiples outperformed stacked small multiples, and mirrored small multiples improved further beyond adjacent (Ondov et al., 2019).

• User Goal: Decide which bar-chart pair is more similar or more correlated.
• Task: Compare pairs of quantitative bar-chart series.
• Data: Two quantitative series per pair, with comparable means and standard deviations.
• Chart Setting: You want a conventional same-direction small-multiple layout rather than a mirrored one.
• Audience: Viewers inspect one paired comparison at a time.
• Success Criterion: Readers can identify the more correlated pair without requiring very large correlation differences.

Break it when: You can use a mirrored two-way small-multiple pair instead. Why: The same experiment showed mirrored bar pairs were more precise than standard adjacent pairs.

Sacrifice: The shared baseline alignment that stacked panels provide. Risk: A conventional adjacent pair is still weaker than a mirrored pair for this task. Mitigation: Use adjacent as the fallback when you want conventional same-direction axes.

Mistake: Stack the two bar charts vertically and assume the common baseline will outweigh the harder cross-panel matching. Why it fails: The correlation task depends more on correspondence between the paired bars than on the stacked baseline.

Failure Sign: Readers struggle more with top-bottom bar pairs than with left-right pairs on the same data. Quick Check: Show the same correlated and less-correlated bar pairs once in adjacent form and once in stacked form, then ask readers which pair is more similar. Stronger Test: Reduce the correlation gap and keep the placement that still yields correct choices.

• Move the second chart beside the first instead of above or below it.
• Keep corresponding categories aligned across the left-right gap.
• If the adjacent pair is still hard to compare, reverse one axis and mirror the pair.