Use color encoding for mean and variance judgments in time-series views

Encode time-series values with color when the main readout is the average level or variability across an interval. For example, use a color-coded time view rather than relying only on line position when readers need to judge which period has the higher mean or greater variance.

Color lets readers treat an interval as an aggregate instead of tracing individual highs and lows one by one.

Mechanism: Color supports ensemble-style summary judgments, which helps readers estimate average value and variability across many time points.

Evidence: The paper reports time-series studies in which mean and variance were judged more accurately from color encodings than from positional encodings, while the reverse pattern held for extrema and range (Szafir et al., 2016).

Notes: This is a task tradeoff, not a claim that color is always better than position.

• User Goal: Compare average level or variability across time intervals.
• Task: Estimate mean or variance from a time-series display.
• Data: Ordered-time values shown across many points or intervals.
• Chart Setting: A time-series view can encode the same values with either color or position.
• Success Criterion: Readers can accurately tell which interval is higher on average or more variable.

Break it when: The primary question is the minimum, maximum, or range in the series. Why: Positional encodings supported those identification judgments more accurately in the reported time-series studies.

Sacrifice: Peaks, troughs, and range become less direct than in a positional display.
Risk: Readers may miss the highest point or widest spread if the chart is asked to do both summary and extrema work at once.
Mitigation: Keep this encoding only when mean or variance is the main readout.

Mistake: Keeping a positional time-series view as the primary display when the real task is average or variability judgment. Why it fails: Position helped extrema and range more than summary statistics in the reported tradeoff.

Failure Sign: Reviewers can find peaks and range but struggle to answer which interval is higher on average or more variable.
Quick Check: Show a color version and a positional version of the same series, then ask mean and variance questions; keep the version that yields cleaner answers on those tasks.
Stronger Test: Compare error rates on representative mean and variance questions across the two encodings.

• Re-encode the time-series values with color instead of relying only on y-position.
• Replace a line-first summary view with a color-coded time view for the same ordered intervals.
• Remove the color-first version from the primary view if the task shifts to maxima, minima, or range.