Use a pie chart instead of an unscaled stacked bar for exact part-to-whole estimation

Use a pie chart when readers must estimate the exact size of one highlighted share in a two-part whole and the alternative is an unscaled stacked bar. For example, use a baseline pie chart instead of a baseline two-segment stacked bar when the reader must enter the darker segment as a whole-number percentage.

This design choice changes the chart family, not the data. In this specific two-segment exact-reading setup, the pie chart produced smaller estimation error than the unscaled stacked bar.

Mechanism: The pie chart gives readers a more accurate basis for estimating a single highlighted share in a two-part whole than the baseline stacked bar used in the comparison.

Evidence: In the baseline chart comparison, the pie chart ranked above the stacked bar for accuracy, and the difference was reported as significant. The review records this result as an empirical chart-selection finding for part-to-whole estimation (Redmond, 2019; Zeng & Battle, 2023).

Notes: This guideline applies to the baseline bar contrast tested in the source, not to all bar-chart variants.

• User Goal: Read the value of one highlighted share as accurately as possible.
• Data: Two segments that sum to 100%.
• Chart Setting: A single part-to-whole chart with no external quantitative scale and no added internal cue marks.
• Success Criterion: Lower absolute error in estimated segment size.

Break it when: The alternative stacked bar includes an external quantitative scale or stronger internal cue marks. Why: The tested pie advantage was against the baseline unscaled bar, and the source also found more accurate refined bar variants.

Sacrifice: You give up the bar chart version that can carry an external quantitative scale. Risk: You can overgeneralize this result if you apply it to every bar variant instead of the unscaled baseline bar. Mitigation: If the bar chart must stay, add a quantitative scale or stronger internal cues rather than leaving it as the baseline bar.

Mistake: Keep an unscaled stacked bar because bar charts are assumed to be more accurate for part-to-whole estimation. Why it fails: In this tested two-segment exact-estimation setup, the baseline bar had higher estimation error than the baseline pie.

Failure Sign: Readers give noticeably inconsistent or inaccurate percentage estimates from the unscaled stacked bar. Quick Check: Build a matched pie-chart version and a matched unscaled stacked-bar version of the same two-part split, then ask reviewers to estimate the highlighted share as a whole number. Stronger Test: Compare absolute error across several share sizes rather than judging the choice from one example only.

• Replace the unscaled stacked bar with a pie chart for the same two-part whole.
• Keep the highlighted segment styling consistent across both versions when you compare them.
• If the stacked bar must remain, add an external quantitative scale or stronger internal cues instead of leaving it unscaled.