Interpreting and questioning graphs
Drawing a graph is half the battle. Reading one , really reading it , is the other half. Every day you will see graphs in newspapers, on TV, in textbooks, and online. A careful reader does not just glance; they question.
Concept
To interpret a graph means to look at it and pull out the information it carries. Three questions are always worth asking:
- What is being shown? What are the categories, the axes, the scale, the title?
- What does the graph tell us? Which category is biggest, smallest? Are there trends (rising, falling, steady)?
- What does the graph not tell us? What is missing? Could the data be biased? Could the scale mislead?
A few common tricks that misleading graphs use:
- Cut axis. If the vertical axis starts at instead of , small differences look huge. A bar at looks ten times taller than a bar at , but the real difference is tiny.
- Unequal scale. If the spacing between numbers on an axis is uneven, the graph distorts the data.
- Cherry-picked categories. Choosing only "good" months or only "loud" customers gives a one-sided picture.
- Pictures of different sizes. In a pictograph, if the pictures themselves grow in size (not just in count), the bigger pictures lie about the count.
When you read a chart in a newspaper, do this little checklist:
- Does it start at zero?
- Are the units and the scale clearly written?
- Is the source mentioned (who collected the data)?
- Are the categories fair, or is something missing?
Asking these questions does not mean every chart is wrong. It just means we read graphs the way a careful person reads a newspaper article: alert and curious.
A good graph honestly conveys what the data says. A bad graph hides, exaggerates, or tricks. Mathematics gives you the tools to spot the difference.
Worked examples
Example 1. A bar graph shows daily sales of a shop. The vertical axis starts at and goes to . Bar heights: Mon , Tue , Wed . Why is this graph misleading?
- The differences (a few rupees) are tiny compared to the absolute sales.
- Because the axis starts at , the bar for Wednesday looks times taller than for Monday , visually exaggerating a trivial difference.
- Honest graph: start at .
Example 2. From a bar graph of monthly book sales, with months Jan–Jun shown but Jul–Dec missing, what is one question a careful reader should ask?
- "Why are only the first six months shown?" Maybe the rest of the year had poor sales, deliberately hidden.
Example 3. A pictograph claims to compare ice cream sales. For shop A, small cone = ice creams. For shop B, the picture is a big cone that is times the size of the small one. The label says "Shop B sells more!" Why is this misleading?
- If both cones each count as one unit, the bigger cone is just a fancier picture. But to a casual eye it suggests Shop B sells times more.
- In pictographs, every picture must be the same size.
Example 4. A graph shows that "people who eat ice cream are more likely to drown". The reader sees a positive bar relationship. What's the explanation?
- Both increase during hot summer months. They are correlated but neither causes the other. Don't confuse correlation with causation.
Try it yourself
- A bar graph shows test scores: A , B , C . The axis starts at . Is this graph trustworthy? Why or why not?
- A pictograph compares the number of children in schools. School A has pictures of a small child; School B has pictures of a giant child. What's wrong?
- From a bar graph of rainfall: Jan mm, Feb mm, Mar mm, Apr mm, May mm. What's the trend?
- A bar graph in a magazine claims petrol prices went "shockingly up". The actual numbers: over three months. Is "shockingly" honest?
- What three things must always appear on a clean bar graph?
- If a graph has no scale on its vertical axis, what should a reader assume?
- Critical thinking: A graph compares school exam results of two schools, only for the top students in each. What is being hidden?
Activity
Graph hunt. Over a single day, find different bar graphs or pictographs , in a newspaper, on a TV news channel, or in your textbook. For each, write down: (1) what data is shown, (2) one true thing the graph tells you, (3) one thing the graph does not tell you or could mislead about. Share with a classmate and discuss whether the graphs were fair.