Data visualizations play a powerful role in shaping public understanding. Charts and graphics are often interpreted as objective evidence, even when they contain errors or are intentionally misleading. Because of this perceived authority, visualizations often become—intentionally or not—vehicles for both misinformation and disinformation. Avoiding these outcomes requires careful design choices, ethical intent, and a clear understanding of how visual communication can mislead.
Information, Misinformation, and Disinformation: Key Distinctions
Before examining best practices, it is important to clarify the differences between information, misinformation, and disinformation.
Information refers to data or claims that are accurate, well-contextualized, and communicated in a way that reflects reality as faithfully as possible.
Misinformation is inaccurate or misleading information shared without the intent to deceive. In data visualization, this often results from poor design choices, misunderstanding the data, or lack of visual literacy.
Disinformation is deliberately false or misleading information created and shared with the intent to deceive, manipulate, or persuade audiences unfairly.
An ethical data visualization practice seeks to communicate information while actively guarding against the accidental spread of misinformation and resisting the intentional tactics associated with disinformation.
Recognize How Visual Design Can Mislead
Visualizations can mislead even when the underlying data is technically correct. Design decisions often influence interpretation more than the numbers themselves.
What to do:
- Choose visual forms that align with how the data should be interpreted.
- Test visuals for potential misreadings by non-expert audiences.
- Ask whether the design encourages accurate comparisons.
Example (responsible):
A time-series line chart uses consistent intervals and clearly labeled axes, making trends easy to interpret without exaggeration.
What not to do:
- Use chart types that obscure scale or relationships.
- Rely on visual complexity that distracts from meaning.
- Assume viewers will notice subtle design caveats.
Example (misleading):
A stacked area chart is used to compare individual categories over time, making it difficult to see that one category is actually declining.
Avoid Unintentional Misinformation
Many misleading visuals are the result of mistakes rather than malice. These errors can spread widely if left unchecked.
What to do:
- Double-check data accuracy, calculations, and labels.
- Match chart types to data structure and purpose.
- Seek peer review or feedback before publishing.
Example (responsible):
A bar chart comparing survey responses clearly states the sample size and labels percentages accurately.
What not to do:
- Mislabel axes or units.
- Confuse percentages with raw counts.
- Combine incompatible datasets without explanation.
Example (misinforming):
A chart compares raw population counts from one source to percentages from another, leading viewers to draw incorrect conclusions.
Prevent Deceptive Visual Framing
Disinformation often relies on framing rather than outright falsehoods. Visual emphasis can guide viewers toward a predetermined conclusion.
What to do:
- Use neutral titles and captions that describe rather than argue.
- Present multiple relevant data points when possible.
- Separate factual display from interpretive commentary.
Example (responsible):
A chart titled “Reported Cases by Year” presents data without implying causes or blame.
What not to do:
- Use emotionally charged or sensational titles.
- Highlight only the most extreme data points.
- Design visuals to suggest causation where none is established.
Example (disinforming):
A chart titled “Why Policy X Failed” uses selectively chosen data to support a political claim without acknowledging alternative explanations.
GuardAgainst Scale and Proportion Manipulation
Manipulating scale is one of the most common techniques used to mislead visually.
What to do:
- Use consistent, proportional axes.
- Start bar charts at zero unless a clear justification is provided.
- Clearly mark breaks in scale when they are unavoidable.
Example (responsible):
A bar chart with a truncated axis includes a visible break and explanatory note explaining the reason.
What not to do:
- Truncate axes to exaggerate differences.
- Use inconsistent scales across related charts.
- Hide scale manipulation through tight cropping.
Example (disinforming):
Two side-by-side charts use different y-axis ranges, making similar values appear dramatically different.
Address Uncertainty to Avoid False Certainty
Misinformation and disinformation both thrive on false precision and overstated confidence.
What to do:
- Display uncertainty using error bars, ranges, or annotations.
- Clarify whether values are estimates, projections, or measurements.
- Explain the reliability of the data source.
Example (responsible):
A forecast visualization shows a range of possible outcomes rather than a single definitive line.
What not to do:
- Present projections as guaranteed outcomes.
- Use overly precise numbers without justification.
- Omit margins of error to simplify the message.
Example (misleading):
A predictive chart shows exact values years into the future with no indication of variability or risk.
Avoid Selective Data Presentation
Selective inclusion of data is a common tactic in disinformation campaigns and a frequent source of unintentional misinformation.
What to do:
- Show full time ranges when trends are central to interpretation.
- Disclose reasons for excluding data.
- Avoid visual narratives that depend on narrow windows of time.
Example (responsible):
A trend line includes periods of increase and decline, allowing viewers to see the broader pattern.
What not to do:
- Cherry-pick dates or categories to support a claim.
- Hide contradictory data.
- Present partial datasets as comprehensive.
Example (disinforming):
A chart showing a dramatic rise in incidents begins immediately after a historical low, exaggerating the perceived trend.
Cite Sources and EnableVerification
Disinformation relies on opacity. Transparency makes misleading visuals easier to detect and correct.
What to do:
- Cite data sources clearly and consistently.
- Provide links or references when possible.
- Document data transformations and assumptions.
Example (responsible):
A visualization includes a source note, publication date, and brief methodological explanation.
What not to do:
- Use vague attributions such as “studies show.”
- Remove source information in social media adaptations.
- Present unverifiable claims as established facts.
Example (misleading):
A graphic circulates widely online with no source, making it difficult for audiences to assess its credibility.
Design for Responsible Sharing
In digital environments, visualizations often travel far beyond their original context.
What to do:
- Design visuals that remain accurate when shared independently.
- Include essential context within the graphic itself.
- Anticipate how a visualization might be cropped or reposted.
Example (responsible):
A social media–ready chart includes a concise title, clear labels, and a visible source line.
What not to do:
- Rely on surrounding text to prevent misinterpretation.
- Create visuals that become misleading when detached from an article.
- Optimize solely for engagement metrics.
Example (misinforming):
A chart shared without its explanatory caption appears to support a claim the original article explicitly rejected.
Conclusion
Avoiding misinformation and disinformation in data visualization requires more than technical skill—it demands ethical awareness and intentional design. By understanding how visual choices shape interpretation, clearly distinguishing honest error from deliberate deception, and committing to transparency and context, data communicators can reduce the spread of misleading visuals. In an era where charts and graphics circulate rapidly and widely, responsible visualization is essential to preserving public understanding and trust.
*Content on this page was curated and edited by expert humans with the creative assistance of AI.