Keep the reference bar visible during exact comparison

Keep the reference bar visible when readers need an exact bar-value comparison. For example, during a bar-height adjustment or cross-view comparison, show the original bar concurrently—such as a centrally overlaid translucent reference—instead of hiding it and forcing recall.

The main wide-versus-tall bias arose when readers had to remember the bar after it disappeared. When the original bar remained onscreen during tracing, the aspect-ratio bias pattern largely vanished.

Mechanism: Simultaneous perception removes the memory step that pulls wide bars upward and tall bars downward. A visible reference lets readers match position directly instead of reconstructing it from memory.

Evidence: In the experiment that compared memory-based reproduction with direct tracing, the original wide-overestimate and tall-underestimate pattern appeared in memory but not when a concurrently visible reference bar stayed on screen during the response (Ceja et al., 2021).

Notes: Direct tracing still showed a small general overestimation, which the paper attributes to motor overshoot rather than the aspect-ratio memory bias.

• User Goal: Make an exact comparison or reproduce a bar’s value.
• Task: Compare a current bar to one from a previous display or previous step.
• Data: Quantitative values encoded by bar-top position.
• Chart Setting: The current workflow hides the earlier bar before the next judgment, and the design can keep a reference onscreen.
• Audience: Viewers making precision judgments rather than rough impressions.
• Success Criterion: Reduce memory-driven overestimation of wide bars and underestimation of tall bars.

Break it when: The task is a deliberate memory test or the earlier value cannot remain onscreen during the judgment. Why: This rule works by removing the memory step that caused the tested bias.

Sacrifice: This change reduces the role of recall, so it does not test what viewers remember without support.
Risk: Direct tracing can still produce a small positive overshoot from motor response.
Mitigation: Treat a small uniform overshoot differently from the wide-versus-tall bias pattern.

Mistake: Keeping a reference available but placing it away from the response bar so the viewer must shift attention. Why it fails: The paper specifically used centrally co-located bars to avoid the memory transfer that eye movements or attention shifts would reintroduce.

Failure Sign: The earlier bar disappears before the exact judgment is made.
Quick Check: Step through the comparison workflow; if the reader must answer after the reference vanishes, memory bias can enter.
Stronger Test: Compare a sequential version with a concurrent-reference version; the wide-versus-tall signed-error pattern should shrink when the reference stays visible.

• Leave the earlier bar onscreen during the exact comparison.
• Overlay the reference bar directly with the adjustable or compared bar instead of separating them.
• Avoid interactions that remove the compared bar before the reader responds.
• Keep the reference and response bars centrally aligned when possible so the comparison stays perceptual rather than memory-based.