Use a scatterplot instead of a positive parallel coordinates plot for correlation judgments
For relate tasks, use a scatterplot on bivariate quantitative correlation displays instead of a parallel coordinates plot to improve fidelity and mitigate imprecise positive-correlation judgments for readers distinguishing nearby association strengths.
- purpose:select
- basis:empirical
- task:relate
- chart:scatter:use
- chart:parallel:avoid
- data:quantitative
- quality:fidelity:use
- lever:chart-family
- operator:association
advice
Choose scatterplots over positive parallel coordinates
Use a scatterplot rather than a positive parallel coordinates plot when readers need to judge the strength of correlation between two quantitative variables. For example, if the same bivariate data can be shown as points in x/y position or as lines across two parallel axes, keep the scatterplot and avoid the positive parallel-coordinates version.
reason
Why the scatterplot works better here
This contrast is about precision, not stylistic preference. The better chart is the one that lets readers tell apart nearby correlation strengths more reliably.
Mechanism: The scatterplot produced smaller just-noticeable differences for positive correlations, so nearby positive association levels were easier to distinguish.
Evidence: Positive scatterplots significantly outperformed positive parallel coordinates in the correlation-judgment experiment, while negative parallel coordinates were not significantly different from scatterplots; the 2023 review collated this as an empirical relate-task finding (Harrison et al., 2014; Zeng & Battle, 2023).
context
Use when positive association is the reading task
- User Goal: Judge which of two displays shows the stronger positive correlation.
- Task: Compare nearby association strengths.
- Data: Two quantitative variables with positive correlation.
- Chart Setting: A single static view where scatterplot and parallel-coordinates versions are both feasible.
- Success Criterion: More reliable discrimination of nearby positive correlation values.
exceptions
Do not use this contrast when the parallel view can be made negative
Break it when: The parallel-coordinates version can be arranged so the same relationship appears as a negative correlation pattern. Why: Negative parallel coordinates performed comparably to scatterplots, while the positive parallel-coordinates version did not.
costs
What you give up by replacing the parallel view
Sacrifice: You give up the parallel-coordinates encoding for this judgment task. Risk: If a parallel-coordinates view is required elsewhere in the workflow, replacing it may not fit the surrounding display. Mitigation: If the parallel-coordinates form must stay, switch to a negative-correlation arrangement instead of keeping the positive one.
mistakes
Common chart-choice mistake
Mistake: Treating positive and negative parallel-coordinates layouts as equivalent for correlation judgment. Why it fails: The study found a clear asymmetry between them.
check
How to test the choice
Failure Sign: Readers struggle to tell which of two nearby positive correlations is stronger in the parallel-coordinates view. Quick Check: Render the same positive-correlation data as a scatterplot and as a positive parallel-coordinates plot at the same size, then ask reviewers which of two close-correlation pairs is more correlated. Stronger Test: Repeat that side-by-side judgment across several nearby correlation levels and keep the chart that produces more consistent answers.
fix
What to change
- Replace the positive parallel-coordinates view with a scatterplot for the correlation-reading step.
- Keep size, data, and correlation level matched when comparing the two versions.
- If the parallel-coordinates form must remain, flip or rearrange axes so the relationship appears as a negative pattern.