Interpreting statistics: Introduction

Statistics - it's all about patterns, compiling the information, and its interpretation. Proper statistical analysis makes large amounts of data understandable and meaningful. But incorrect statistical analysis may lead to wrong interpretations and conclusions. Here are a few examples of what could go wrong in a statistical analysis. Here are four friends.

On average, they have two rabbits each. But actually, none of them has two rabbits. One of them is crazy about rabbits and has eight of them. The others don't have any rabbits at all. In this case, the average says nothing about how many rabbits a typical person has.

When one or several values deviate this much from the others, it's often better to use the median rather than the average. Here's a diagram showing how much Maria gets as her monthly allowance. The rightmost bar is twice as high as the one on the left. It looks like her allowance has doubled in two years, but it hasn't. Look at the Y axis.

It doesn't start at zero. Instead, there's a zigzag line there. The zigzag line shows that it's a broken scale. Sometimes a broken scale is a good way to save space and zoom in on the essentials in a graph. But a broken scale can also make changes seem bigger than they really are.

Here's another diagram. It shows how many students in each class play soccer at least once a week. The height of the ball tells us how many students play. In the class 9A, there are twice as many soccer players as in the class 9B. In the diagram, the ball for 9A is twice as tall as the ball for 9B.

But it's four times larger. Area scale equals linear scale squared. It would be better to use a bar chart like this. Those three examples showed how the presentation of statistical data could be misleading. When you create a diagram, think about simplifications you make.

Perhaps, there is a clearer, or more honest way to present the same information.