Sign in to save

Bookmark this page so you can find it later.

Sign in to save

Bookmark this page so you can find it later.

Misleading graphs can make real data look more dramatic, less important, or more certain than it actually is. This cheat sheet helps students check axes, scales, labels, sample information, and visual proportions before trusting a graph. These skills are useful in statistics, science, news, advertising, and social media.

A quick graph check can prevent a wrong conclusion from a visually convincing display.

The most important ideas are to compare the picture with the numbers and to ask whether the graph uses a fair scale. Bar graphs should usually start at 00, line graphs should show consistent intervals, and pie charts should total 100%100\%. Percent change should be computed with newoldold×100%\frac{\text{new} - \text{old}}{\text{old}} \times 100\%, not guessed from the image.

Any graph should also be checked for missing context, unclear units, cherry-picked time ranges, and sample size.

Key Facts

  • For a bar graph, the vertical axis should usually start at 00 because bar length represents the size of the value.
  • Percent change is calculated by newoldold×100%\frac{\text{new} - \text{old}}{\text{old}} \times 100\%.
  • A relative comparison can be measured with the ratio larger valuesmaller value\frac{\text{larger value}}{\text{smaller value}}, which helps check whether the visual size matches the data.
  • A pie chart should represent parts of one whole, so all slice percentages should add to 100%100\%.
  • Equal spacing on an axis must represent equal numerical intervals, such as 0,10,20,300, 10, 20, 30 rather than 0,10,15,400, 10, 15, 40 with the same spacing.
  • Area-based pictures can exaggerate differences because doubling both height and width makes the area 2×2=42 \times 2 = 4 times as large.
  • A small sample can be unstable, so a result from n=20n = 20 people is usually less reliable than a similar result from n=2000n = 2000 people.
  • A graph with no units, no source, or no sample description is incomplete because the viewer cannot judge what the numbers actually mean.

Vocabulary

Truncated axis
A truncated axis is an axis that does not start at the expected baseline, often 00, which can make differences look larger.
Scale
The scale is the pattern of values marked on an axis, such as counting by 55s, 1010s, or 100100s.
Cherry-picking
Cherry-picking means selecting only the data points or time period that support a desired conclusion while leaving out important context.
Sample size
Sample size is the number of observations or people in a data set, often written as nn.
Percent change
Percent change compares an old value to a new value using newoldold×100%\frac{\text{new} - \text{old}}{\text{old}} \times 100\%.
Misleading graph
A misleading graph is a data display that uses design choices, missing information, or unfair comparisons to encourage an inaccurate interpretation.

Common Mistakes to Avoid

  • Ignoring a truncated bar graph axis is wrong because a bar starting above 00 can make a small difference look very large.
  • Comparing picture heights without checking areas is wrong because icons and bubbles may change in both height and width, so the visual area can grow much faster than the data.
  • Trusting a percent without asking for the original value is wrong because an increase of 50%50\% from 22 is only 33, while an increase of 50%50\% from 200200 is 300300.
  • Assuming equal spacing means equal time or equal numbers is wrong because some graphs place unequal intervals at the same visual distance.
  • Accepting a graph without a source, units, or sample size is wrong because the data may be incomplete, biased, or impossible to verify.

Practice Questions

  1. 1 A bar graph compares sales of 4848 and 5252 items, but the vertical axis starts at 4545. What is the actual difference, and why might the graph look exaggerated?
  2. 2 A value increases from 8080 to 100100. Use newoldold×100%\frac{\text{new} - \text{old}}{\text{old}} \times 100\% to find the percent change.
  3. 3 A pie chart has slices labeled 35%35\%, 30%30\%, 20%20\%, and 25%25\%. What problem should you notice?
  4. 4 A news graph shows only the last 33 months of a 1010 year trend and claims the pattern is permanent. Explain why this graph may be misleading.

Understanding Misleading Graphs and How to Spot Them

Graphs influence judgment because people notice shape and color before they read values. A steep rising line can feel like a sudden crisis even when the actual increase is small. The apparent steepness depends on the width and height chosen for the graph.

A designer can make the same data look calm by stretching it wide or urgent by making it narrow. This does not always mean someone intended to deceive.

It does mean the picture needs to be separated from the claim. Read several data points and describe the change in actual units before deciding how large it is.

Time is another source of confusion. A graph may begin just before a rise or end just after a fall. This is called choosing a time window, and it can hide a longer pattern.

For example, a one month chart of temperatures may suggest unusual warming while a ten year chart shows normal seasonal variation. In school science, this matters when comparing experiments done on different days. In news reports, it matters for prices, test results, crime rates, or attendance.

Look for the starting date, ending date, and any missing periods. A trend needs enough observations to distinguish a real pattern from ordinary ups and downs.

The data collection process matters as much as the graph design. A survey of students leaving one club cannot represent every student in a school. People who choose to answer an online poll may have stronger opinions than people who ignore it.

Even a large group can give a biased result if the group was selected in an unfair way. Results can differ because of the wording of a question too. A question that suggests a preferred answer can push people toward it.

When studying a graph, identify the population the claim is about, then compare it with the people or objects actually measured. Notice whether the source explains when, where, and how the data was collected.

Percentages need a reference value. An increase from two customers to four customers is an increase of one hundred percent, yet it represents only two additional customers. A decrease from one hundred to fifty is a decrease of fifty percent.

Going from fifty back to one hundred then requires an increase of one hundred percent, not fifty percent. This happens because each percentage is based on a different starting amount. Use the original value as the reference and state the actual difference beside the percentage.

In your own graphs, give a clear title, label units, include the source, and choose a display that fits the data. A careful graph makes comparison easier without forcing the viewer toward one conclusion.