Use text axis labels instead of pictograph labels for brief recall tasks

Keep axis category labels in text when viewers must remember values after a brief glance. For example, place short words under the bars instead of replacing those labels with pictographs, even if later prompts or surrounding graphics use icons.

Short text labels can be recognized quickly and linked to a value with little extra decoding. Icon labels add another interpretation step and can weaken the connection between category and value during brief exposure.

Mechanism: Text axis labels make it easier to encode which category goes with which number when the chart disappears and the reader must recall the values from memory.

Evidence: Across chart types, pictographs used as x-axis labels increased recall error, while changing the response prompt between text and pictographs made no reliable difference (Haroz et al., 2015).

• User Goal: Remember which value belongs to each category.
• Task: Brief-glance recall after the chart is removed.
• Data: Quantitative values with category names that can be shown as short words.
• Chart Setting: The x-axis labels could be rendered either as text or as pictographs.
• Audience: Viewers reading short, unambiguous category labels.
• Success Criterion: Lower recall error for category-value matches.

Break it when: The category cannot be expressed as succinct, easy-to-read text. Why: The reported text advantage was observed with short readable labels, and the study did not test longer or more complex wording.

Sacrifice: You give up some visual illustration in the axis. Risk: The chart may feel less pictorial. Mitigation: Keep imagery in the data marks if you need it, but leave the axis labels as text.

Mistake: Replacing every axis label with an icon to match the subject matter. Why it fails: The icon label is harder to encode quickly than the short text label it replaces.

Failure Sign: Viewers remember the numbers but confuse which category each number belonged to. Quick Check: Compare a text-labeled version against an icon-labeled version in a brief-glance recall task. Stronger Test: Randomize the recall order of categories and compare average absolute error across the two label styles.

• Restore text words on the axis where categories are named.
• Shorten category wording so the text remains quick to read.
• Remove pictograph labels from the axis rather than duplicating the category in both icon and word form.