Why charts can mislead
A chart is an argument made with shapes. Even when every number is correct, the way the numbers are drawn can lead readers to a conclusion the data does not support. Sometimes this is deliberate; more often it is accidental, caused by software defaults, haste or a wish to make a result look striking.
The aim of honest visualization is simple: a reader who looks at the chart should come away with an impression close to what the data actually shows. The points below apply whether you are making a chart or judging one in a news story, a report or a dashboard.
Start with the question
Before choosing a chart type, decide what the chart should help someone understand. Is it a comparison between categories, a change over time, a share of a whole, or a relationship between two measures? A chart with no clear question tends to become decoration.
Choose a chart type that fits
- Bar charts compare amounts across categories. They work well because people judge length accurately.
- Line charts show change over time and make trends easy to follow.
- Scatter plots show the relationship between two measures.
- Histograms show how values are spread.
- Pie charts show parts of a whole, but they are hard to read when there are many slices or when slices are similar in size. A sorted bar chart is often clearer.
- Maps suit geographic patterns but can mislead because large, sparsely populated areas draw the eye.
Three-dimensional effects, shadows and decorative icons usually make it harder to read values and should be avoided without a good reason.
Common ways charts mislead
Truncated axes
For bar charts, the baseline should normally start at zero, because the bar’s length is meant to represent the value. Consider a hypothetical comparison of two values, 98 and 100. If the axis starts at 0, the bars look almost identical, which is accurate. If the axis starts at 96, the first bar is 2 units tall and the second is 4 units tall, so the second looks twice as large even though it is only about 2 percent bigger. Line charts are more forgiving, since they show change rather than size, but a narrow range can still exaggerate small movements, so the scale should be stated clearly.
Stretched or squashed proportions
Making a chart very tall exaggerates a rise; making it very wide flattens it. There is no perfect shape, but extreme proportions should be avoided and the choice should not be made to fit a desired story.
Cherry-picked time ranges
Starting a line at a low point makes growth look steep; starting at a peak makes it look like decline. Show enough history for the pattern to be seen in context, and say why the range was chosen.
Dual axes
Charts with two vertical scales let the author make lines cross or align almost at will. They can be useful but should be labeled carefully, and readers should be cautious about apparent correlations.
Misleading areas and sizes
When a symbol is scaled in both width and height to represent a number, the area grows with the square, so a value that is twice as large can appear four times bigger. Prefer encodings that readers judge accurately, such as length.
Inconsistent intervals
If the horizontal axis skips periods or uses uneven spacing without saying so, trends look different from reality. Gaps in data should be shown as gaps, not joined by a line that implies continuity.
Color problems
Color should carry meaning. Using red and green together can be unreadable for people with common forms of color blindness. Sequential colors suit ordered values, while distinct hues suit categories. Strong colors drawn to one series can make it seem more important than it is. Always check that the chart is understandable in grayscale or with a color-blind-friendly palette.
Omitted context
A figure without a comparison, a unit or a source invites misreading. Percentages need a base: a rise from 2 to 4 is a 100 percent increase, but only a change of 2 units. Both views can be informative, and showing the raw numbers alongside helps readers judge size.
Confusing correlation with cause
Two lines moving together do not show that one causes the other. A chart can imply a link without stating it, so be careful with titles and annotations. Deeper data analytics can examine such relationships more carefully, although even rigorous analysis has limits.
Practical habits for honest design
- Give the chart a plain title that states what is shown, not a conclusion the data cannot support.
- Label axes, units and sources.
- Start bar charts at zero.
- Keep the design simple; remove anything that does not help the reader.
- Order categories meaningfully, for example by size.
- Show uncertainty or ranges where they exist.
- Test with someone unfamiliar with the data: ask what they take from the chart.
- Check the underlying records. A clean chart of flawed data is still wrong, as the article on data quality explains.
Reading charts critically
Readers can protect themselves by asking a few questions. Where does the axis start? What period is shown, and why? What is the source? Is there something missing, like a comparison or a larger group? These habits belong to a broader set of digital literacy skills and are useful when evaluating claims in the news. They can also help you notice when a chart contributes to online misinformation, whether by design or by accident, and they overlap with general media literacy and fact-checking methods.
The overall principle is to let the data speak at its real volume. A modest, honest chart tends to build more trust than an impressive one that falls apart under scrutiny.