Misleading graphs

When interpreting graphs, there are three things you should always look for first:

What are the labels on the axes, what do the data points represent?
What units are these quantities measured in?
What is the pattern? Is it a straight or curved line?

We also have to be on the lookout for misleading graphs. These graphs are often used to convince us of a certain viewpoint or to sell a product.

Imagine this graph was displayed during a news story about house prices rising:

A graph showing how the average house price has increased from £80,000 in 1998 to £82,000 in 1999

The graph is designed to make you think that there has been a major change in price over the last year. Look closely at the y-axis. Did you notice that the first number on the axis was £80,000?

If the graph was drawn correctly it would look like this:

The same graph showing how the average house price has increased from £80,000 in 1998 to £82,000 in 1999, however, the y-axis begins at zero

This graph is not as shocking however and so is far less likely to be used in a news story.

Sometimes misleading graphs don’t have any axis measurements at all! Look at the following example:

A three dimensional graph whose y-axis is labelled ‘number of singles sold’. The y-axis is not numbered

This graph is completely meaningless without a label on the y-axis – it is impossible to interpret the graph without it. The way the graph is drawn could also fool you into thinking that the number of CD singles sold in 1995 was greater than in 1997 when actually they are the same! A better version is drawn below.