How to Read Graphs and Percentages Without Being Misled

One bar looks twice as tall as another. Does that mean its value has doubled? Check the scale before answering. A graph can contain correct numbers and still encourage a rushed conclusion. Reading the picture is only part of the job. You also need to understand what was measured and how it was shown.

This skill helps with schoolwork, exam questions and news shared on your phone. We will practise with clearly labelled fictional examples. These numbers are for learning. They do not describe real pupils or report the results of a real survey.

Start with the question, not the calculation

Imagine a graph called “Books read by two classes”. The title leaves several questions open. Does it show books per pupil or books added together? Is the period one week or one year? Do the classes contain the same number of pupils? The meaning of a difference can change when you discover the unit and the period.

Before calculating, complete this sentence: “This graph compares this quantity, measured in this unit, over this period.” If you cannot finish it, look for the missing information in the labels or nearby text. If the information is absent, say so. Do not fill the gap with a convenient guess.

The graph guide from IBGE, Brazil's national statistics institute, recommends checking the title, source, labels and scales. The guide is in Portuguese. Its reading checks are useful in any language: subject, unit, groups, period and source.

The baseline changes how bars look

Compare two fictional book counts: 50 for group A and 60 for group B. The difference is 10 books. Relative to the first value, the increase is 20%, because 10 divided by 50 is 0.2. The second value is not twice the first. To double 50, it would have to reach 100.

Now draw the same values in two ways. In the first graph, the vertical axis starts at zero. In the second, it starts at 40. An axis is a reference line for reading values. The baseline is the starting level from which the visible bars rise.

Above a baseline of 40, one bar shows the interval from 40 to 50 and the other from 40 to 60. Those visible lengths are 10 and 20 units. The second visible length is twice the first, although the actual values are still 50 and 60.

A non-zero axis does not automatically make every graph dishonest. On a line graph, it can help show a small change. Ask whether the choice is clear and whether the picture invites a misleading comparison. For bar-length comparisons, zero is a particularly important reference.

IBGE also offers a graphs-and-tables activity, in Portuguese, that addresses misleading scales and missing information. One useful check is to copy the graph's values into a table. You can then compare the numbers without depending on the picture's visual impact.

A percentage needs a whole

In a separate fictional example, 10 pupils in a class of 20 choose to read a short story. In another class of 40, 16 make that choice. The second class has more readers in absolute numbers: 16 is greater than 10. But the first class has a larger proportion: 10 out of 20 is 50%, while 16 out of 40 is 40%.

A proportion compares a part with its whole. A percentage expresses that proportion per hundred. In the fraction 10/20, the denominator, 20, is the total used for the comparison. Losing track of that total can change your answer.

Fictional example	Short-story readers	Total pupils	Percentage
Class A	10	20	50%
Class B	16	40	40%

Both comparisons are correct. They answer different questions. “How many people?” asks for a count. “What share of the group?” asks for a proportion. The statement “this class reads more” is incomplete until you know which comparison it means.

The same care applies when comparing schools, neighbourhoods or years. Always ask: a percentage of what? If the total changes, a count can rise while its share falls. That is not a contradiction. It means you need to follow both measures.

From 30% to 45%: two different changes

In another fictional example, the share taking part in an activity rises from 30% to 45%. The direct difference is 15 percentage points. A percentage point is the unit used when subtracting one percentage from another: 45 minus 30.

The relative increase answers a different question: how large is that change compared with the original value? Here, 15 is half of the original 30. The participation share has therefore increased by 50% relative to its earlier level.

Saying “it increased by 15%” would mean something else. Fifteen per cent of 30 is 4.5. Adding that increase would give 34.5%, not 45%. The words attached to the number matter just as much as the calculation.

Write the comparison base beside your working. For relative change, divide the difference by the initial value, then multiply by 100. For percentage points, subtract the two percentages directly. Include the correct measure in your answer.

Check whether the categories fit together

Suppose one graph shows average reading time and another shows library loans. These quantities might be related, but they measure different things. A library could receive more visitors, change its borrowing rules or add books. Two rising lines do not tell you which explanation is correct.

Check whether categories overlap, too. A fictional survey might let a pupil choose both “short stories” and “comics”. If several answers are allowed, the percentages need not add up to 100%. A pie chart that presents them as separate parts of one whole could mislead.

Two quantities changing together show an association, an observed relationship. Causation means that one thing brings about a change in another. Establishing causation needs more evidence than a matching pattern in two graphs.

Test the answer with a different question

Return to the fictional classes. Which has more short-story readers? Class B, if you are counting people. Which has the larger participating share? Class A. The data stayed the same; the question changed. Write both answers with their measures to make the difference clear.

In a school question, ask whether you calculated what the instruction requested or answered a similar question instead. Underline words such as total, proportion, difference and relative increase. They tell you which comparison to make.

A short exercise for home or class

Choose a graph with an identified source. Cover the headline and describe only what the graph supports. Then uncover it. Does the headline describe a count, a share, a difference or a cause? Does any word go beyond the evidence?

1. Record the title, unit, period, groups and source.
2. Copy two values into a table.
3. Check where the axis starts and whether its scale has equal steps.
4. Do the requested calculation and name the measure.
5. Write one supported conclusion and one unanswered question.

A family can use a news report; a class can compare descriptions of one graph. Ask each person to show where their conclusion came from. You do not need a contest to find the fastest answer. A clear explanation lets someone else check your reasoning.

For further examples, visit the official IBGEeduca graph guide. Then answer without looking back: why does a bar that looks twice as tall not necessarily represent twice the value?